<?xml version='1.0' encoding='UTF-8'?><?xml-stylesheet href="http://www.blogger.com/styles/atom.css" type="text/css"?><feed xmlns='http://www.w3.org/2005/Atom' xmlns:openSearch='http://a9.com/-/spec/opensearchrss/1.0/' xmlns:blogger='http://schemas.google.com/blogger/2008' xmlns:georss='http://www.georss.org/georss' xmlns:gd="http://schemas.google.com/g/2005" xmlns:thr='http://purl.org/syndication/thread/1.0'><id>tag:blogger.com,1999:blog-1407002244064487228</id><updated>2024-10-07T00:50:25.393-04:00</updated><category term="Solaris"/><category term="Solaris 10"/><category term="Disaster Recovery"/><category term="Oracle"/><category term="Sun"/><category term="Linux"/><category term="ZFS"/><category term="File System"/><category term="HP-UX"/><category term="Network"/><category term="Red Hat"/><category term="Restore"/><category term="SPARC"/><category term="Solaris 11"/><category term="Solaris 9"/><category term="Zones"/><category term="Backup"/><category term="Bugs"/><category term="Containers"/><category term="File Systems"/><category term="IP Address"/><category term="Logical Volume Management"/><category term="Netbackup"/><category term="Solaris 8"/><category term="Static"/><category term="Virtualization"/><category term="Vxfs"/><category term="patches"/><category term="recover"/><category term="Avamar"/><category term="Branded Zones"/><category term="Business Continuity Planning"/><category term="CentOS"/><category term="DHCP"/><category term="Database"/><category term="EMC"/><category term="IBM"/><category term="Install"/><category term="Jumpstart"/><category term="Kernel Tuning"/><category term="LDOMs"/><category term="Logical Domains"/><category term="Maverick Meerkat"/><category term="Miniroot"/><category term="Networker"/><category term="Package"/><category term="Performance Tuning"/><category term="RSC"/><category term="RTO"/><category term="Recovery Time Objective"/><category term="Repository"/><category term="SC"/><category term="SLA"/><category term="Server"/><category term="Service Level Agreement"/><category term="Shared Memory"/><category term="Storage"/><category term="Symantec"/><category term="TSM"/><category term="Tapeless Backup"/><category term="Ubuntu"/><category term="Ubuntu 10.10"/><category term="Unix"/><category term="Windows 2003 Enterprise"/><category term="attach"/><category term="detach"/><category term="dladm"/><category term="firewall"/><category term="hardware"/><category term="ifconfig"/><category term="ignite"/><category term="ipadm"/><category term="iptables"/><category term="mminfo"/><category term="repo"/><category term="reset"/><category term="rscadm"/><category term="save set"/><category term="scadm"/><category term="ssid"/><category term="sunsolve"/><category term="system state"/><category term="v240"/><category term="v440"/><category term="v480"/><category term="winworkr"/><category term="yum"/><category term="zoneadm"/><category term="zonecfg"/><category term="zpool"/><title type='text'>/var/adm</title><subtitle type='html'>Observations and Technical Tips from a Disaster Recovery Professional</subtitle><link rel='http://schemas.google.com/g/2005#feed' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/posts/default'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default?redirect=false'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/'/><link rel='hub' href='http://pubsubhubbub.appspot.com/'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><generator version='7.00' uri='http://www.blogger.com'>Blogger</generator><openSearch:totalResults>23</openSearch:totalResults><openSearch:startIndex>1</openSearch:startIndex><openSearch:itemsPerPage>25</openSearch:itemsPerPage><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-8881298629091782482</id><published>2013-11-08T14:10:00.001-05:00</published><updated>2013-11-08T14:10:47.668-05:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Linux"/><category scheme="http://www.blogger.com/atom/ns#" term="Red Hat"/><category scheme="http://www.blogger.com/atom/ns#" term="repo"/><category scheme="http://www.blogger.com/atom/ns#" term="Repository"/><category scheme="http://www.blogger.com/atom/ns#" term="yum"/><title type='text'>Use a Red Hat DVD as a local software repository</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhGGgG7CCkmWOUnEwg0iX1LOnqBrvLWEnsJZ8Z4zRfuGK2DEIt_JJBHlIEruKQ9ofY1e7NQzyobfpeStmnQF-eoAR7iW8yabSHJebMdiusFSIf3RnmAm2P2JzEvY65d7p-hjfHR1oxyAUY/s1600/redhat-logo.jpeg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;200&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhGGgG7CCkmWOUnEwg0iX1LOnqBrvLWEnsJZ8Z4zRfuGK2DEIt_JJBHlIEruKQ9ofY1e7NQzyobfpeStmnQF-eoAR7iW8yabSHJebMdiusFSIf3RnmAm2P2JzEvY65d7p-hjfHR1oxyAUY/s200/redhat-logo.jpeg&quot; width=&quot;181&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
Invariably, when I deploy a new linux server from a RHEL 6 dvd (or iso), I always forget to install a component, such as ftp, or mt-st, or some other utiltiy I may need. If you do not have a support subscription, a local repository, or a reliable connection to the general internet, your options are limited. You can, however, use your RHEL install dvd as a software repo very easily with a few minimal tricks.&lt;br /&gt;
&lt;br /&gt;
The first step is to have a RHEL installation DVD, or (iso file if you are using a VM). Insert the DVD into your drive (or mount the iso as a DVD in your vSphere client for you VMware folks), and mount it- my old Solaris habits mean I always prefer to mount removable media to the /cdrom directory for the sake of nostaligia...&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mkdir /cdrom&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mount /dev/cdrom /cdrom&lt;br /&gt;mount: block device /dev/sr0 is write-protected, mounting read-only&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
We can now specify the dvd as a repository under the /etc/yum.repos.d directory. I usually call this repo by the imaginitive name of &quot;dvd&quot;&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /etc/yum.repos.d&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# vi dvd.repo&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
The first line of this file should be your repo name (dvd) in square brackets. The second linecan have any string you want after the &quot;name=&quot;. The &quot;baseurl&quot; value should point to the &quot;Server&quot; directory on your dvd. The &quot;enabled&quot; setting allows rpm to search the repo for sources, and the &quot;gpgcheck&quot; setting determines if yum should check GPG signatures during package installs- which is not usually necessary in most cases I&#39;vce dealt with (I&#39;d be more concerned if we were using a public repo across open internet). When all is said and done, your dvd.repo file should look like the example below:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# more dvd.repo&lt;br /&gt;[dvd]&lt;br /&gt;name=RHEL 6 dvd repo&lt;br /&gt;baseurl=file:///cdrom/Server&lt;br /&gt;enabled=1&lt;br /&gt;gpgcheck=0&lt;/span&gt;&lt;br /&gt;
Now if you do a &quot;yum repolist&quot; you&#39;ll see your DVD-based repo listed:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# yum repolist&lt;br /&gt;Loaded plugins: product-id, refresh-packagekit, security, subscription-manager&lt;br /&gt;This system is not registered to Red Hat Subscription Management. You can use subscription-manager to register.&lt;br /&gt;repo id&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; repo name&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; status&lt;br /&gt;dvd&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; RHEL 6 dvd repo&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 3,648&lt;br /&gt;ol6_UEKR3_latest Latest Unbreakable Enterprise Kernel for Oracle Linux 6S&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 29&lt;br /&gt;ol6_UEK_latest&amp;nbsp;&amp;nbsp; Latest Unbreakable Enterprise Kernel for Oracle Linux 6S&amp;nbsp;&amp;nbsp;&amp;nbsp; 239&lt;br /&gt;ol6_latest&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Oracle Linux 6Server Latest (x86_64)&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 22,558&lt;br /&gt;repolist: 26,474&lt;/span&gt;&lt;br /&gt;
You can now install any of the packages that are available on the DVD, such as ftp (for file transfers) or mt-st (for scsi tape device control/testing):&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# yum install ftp&lt;br /&gt;Loaded plugins: product-id, refresh-packagekit, security, subscription-manager&lt;br /&gt;This system is not registered to Red Hat Subscription Management. You can use subscription-manager to register.&lt;br /&gt;Setting up Install Process&lt;br /&gt;Resolving Dependencies&lt;br /&gt;--&amp;gt; Running transaction check&lt;br /&gt;---&amp;gt; Package &lt;/span&gt;&lt;a href=&quot;ftp://ftp.x86_64/&quot;&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;ftp.x86_64&lt;/span&gt;&lt;/a&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt; 0:0.17-53.el6 will be installed&lt;br /&gt;--&amp;gt; Finished Dependency Resolution&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Dependencies Resolved&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;================================================================================&lt;br /&gt;&amp;nbsp;Package&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Arch&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Version&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Repository&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Size&lt;br /&gt;================================================================================&lt;br /&gt;Installing:&lt;br /&gt;&amp;nbsp;ftp&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; x86_64&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0.17-53.el6&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; dvd&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 58 k&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Transaction Summary&lt;br /&gt;================================================================================&lt;br /&gt;Install&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 1 Package(s)&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Total download size: 58 k&lt;br /&gt;Installed size: 95 k&lt;br /&gt;Is this ok [y/N]: y&lt;br /&gt;Downloading Packages:&lt;br /&gt;Running rpm_check_debug&lt;br /&gt;Running Transaction Test&lt;br /&gt;Transaction Test Succeeded&lt;br /&gt;Running Transaction&lt;br /&gt;&amp;nbsp; Installing : ftp-0.17-53.el6.x86_64&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 1/1&lt;br /&gt;dvd/productid&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; | 1.7 kB&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 00:00 ...&lt;br /&gt;&amp;nbsp; Verifying&amp;nbsp; : ftp-0.17-53.el6.x86_64&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 1/1&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Installed:&lt;br /&gt;&amp;nbsp; &lt;/span&gt;&lt;a href=&quot;ftp://ftp.x86_64/&quot;&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;ftp.x86_64&lt;/span&gt;&lt;/a&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt; 0:0.17-53.el6&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Complete!&lt;/span&gt;&lt;br /&gt;
When you are done, you can disable the repo by removing (or moving) the /etc/yum.repos.d/dvd.repo file, and umounting the DVD.</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/8881298629091782482/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/11/use-red-hat-dvd-as-local-software.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/8881298629091782482'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/8881298629091782482'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/11/use-red-hat-dvd-as-local-software.html' title='Use a Red Hat DVD as a local software repository'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhGGgG7CCkmWOUnEwg0iX1LOnqBrvLWEnsJZ8Z4zRfuGK2DEIt_JJBHlIEruKQ9ofY1e7NQzyobfpeStmnQF-eoAR7iW8yabSHJebMdiusFSIf3RnmAm2P2JzEvY65d7p-hjfHR1oxyAUY/s72-c/redhat-logo.jpeg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-8558847629434585743</id><published>2013-10-24T14:18:00.002-04:00</published><updated>2013-10-24T14:18:33.061-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Backup"/><category scheme="http://www.blogger.com/atom/ns#" term="Business Continuity Planning"/><category scheme="http://www.blogger.com/atom/ns#" term="Disaster Recovery"/><category scheme="http://www.blogger.com/atom/ns#" term="IBM"/><category scheme="http://www.blogger.com/atom/ns#" term="Recovery Time Objective"/><category scheme="http://www.blogger.com/atom/ns#" term="Restore"/><category scheme="http://www.blogger.com/atom/ns#" term="RTO"/><category scheme="http://www.blogger.com/atom/ns#" term="Service Level Agreement"/><category scheme="http://www.blogger.com/atom/ns#" term="SLA"/><category scheme="http://www.blogger.com/atom/ns#" term="TSM"/><title type='text'>Disable password authentication for TSM (and a good Disaster Recovery tip)</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjJaZRlwhyphenhyphenJ1wImImsk6R6qVXSpXTDPjjU39q0dsq-MJYzE_8-as_XLLVnV_su1tpoTuv5VB5Hw1UzjndKfWMIW0lg59pToJh5kpyWRpmjJg94vLOPJnjPSojov-zVKriPgzqGSlN8_Rvc/s1600/TSM_logo.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;100&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjJaZRlwhyphenhyphenJ1wImImsk6R6qVXSpXTDPjjU39q0dsq-MJYzE_8-as_XLLVnV_su1tpoTuv5VB5Hw1UzjndKfWMIW0lg59pToJh5kpyWRpmjJg94vLOPJnjPSojov-zVKriPgzqGSlN8_Rvc/s200/TSM_logo.jpg&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
When you are attempting to perform a disaster recovery, the last thing you ever want to do is be stuck looking for a password- or several passwords. Such a simple task can bring a recovery operation to a screeching halt, and have an adverse impact on your ability to meet RTO or other SLA obligations.&lt;br /&gt;
&lt;br /&gt;
This can be a particularly bad problem if you are using IBM&#39;s Tivoli Storage Manager (TSM) as your backup/recovery software. TSM can require an indiviadual password to authenticate each backup client. This password can be different for each host/client, and may or may not have any relation to any domain credentials or more universal user authentication. The next to last thing you want to do when planning for a recovery is to collect a multitude of individual access credentials, that may change long before you ever need them. It can easily become a recordkeeping nightmare- the bane of any disaster recovery plan.&lt;br /&gt;
&lt;br /&gt;
Luckily, there is one saving grace. The good folks at IBM thought of this, and provides a workaround that only requires knowing a single access credential. As long as you know the administrative login for the TSM web-management tool (aka the Tivoli Integrated Portal), you can easily bypass the need for authentication.&lt;br /&gt;
&lt;br /&gt;
If you have TSM authentication turned on, any call to the dsmc program (the TSM command line client) will prompt you for a password. If you do not have the correct password, or enter a blank password, you will have an authentication failure. This may also hamper any attempts at automation since you would need to pass a password option to dsmc to invoke it in the context of a larger script.&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgDU2bDvdjzbDCFQEJq6m9REng3hUhB4ppgB3LGO1vUya4JS7IZbzeDRinT2o_wpSWyeRp97CvrDr1jsufMxekrzPP08gbiLmg_3hYmt8Mh3zy-nhTOj7odIFoSzp9-PqS0vwYRr36cTdA/s1600/TSM3.PNG&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;180&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgDU2bDvdjzbDCFQEJq6m9REng3hUhB4ppgB3LGO1vUya4JS7IZbzeDRinT2o_wpSWyeRp97CvrDr1jsufMxekrzPP08gbiLmg_3hYmt8Mh3zy-nhTOj7odIFoSzp9-PqS0vwYRr36cTdA/s640/TSM3.PNG&quot; width=&quot;640&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
Again, as long as you have administrative access to the TSM Web UI, you can change this behavior easily- and you will only need a single login /password pair to do it (the TSM UI administrative login). From the TSm UI, click on &quot;Manage Servers&quot; then the name of the TSM Server.&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgtqLVqVhpAaXtHs6vengrtTRws8YoB1iHfMFg9DsCN7_y_vV3JWpJ2uOeTWld5a3_5O9zqCwCo7zHGLKUPA3DxE4_7sy0Wm4c-KLuVDBzPHs3Wi2o8kSWrGArQRWcR0_zhNsnZtXaXOOU/s1600/TSM1.JPG&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;198&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgtqLVqVhpAaXtHs6vengrtTRws8YoB1iHfMFg9DsCN7_y_vV3JWpJ2uOeTWld5a3_5O9zqCwCo7zHGLKUPA3DxE4_7sy0Wm4c-KLuVDBzPHs3Wi2o8kSWrGArQRWcR0_zhNsnZtXaXOOU/s640/TSM1.JPG&quot; width=&quot;640&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
After that, click on the &quot;Security&quot; tab, then make sure the selection titled &quot;Require Logon Password&quot; is unchecked. Click on &quot;OK&quot; or &quot;Apply&quot; at the bottom of the window, and you are set, no reboots, or service restarts are needed.&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg0NCJggsCp9Gc-QvYtgJoYoeOUU-gDCfc0kLNa9UbNXAGsR23RpB9Fp0rfUglnK016xQgorsyuYAfUR7AnR3QgbTomQI1yTXut-dDrbz7PzJuLbEWgykbpL576kpwV-l0HwWhmqrVgtDE/s1600/TSM2.JPG&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;258&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg0NCJggsCp9Gc-QvYtgJoYoeOUU-gDCfc0kLNa9UbNXAGsR23RpB9Fp0rfUglnK016xQgorsyuYAfUR7AnR3QgbTomQI1yTXut-dDrbz7PzJuLbEWgykbpL576kpwV-l0HwWhmqrVgtDE/s640/TSM2.JPG&quot; width=&quot;640&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
Now when you invoke the dsmc program, you can enter any nonsense string when prompted for a password- or us any string you desire as the password option if you are passing the login credentials via command line (if you are invoking the command as part of a larger script).&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhqh519H9eb6vuDfPpVmzgZe4UbYjQaXdHLWSq4V9tGxx-OFIHMp6HLP5Nuus3rG9kSJgSvXQZ3Sn_mXyq3F_7X3Ll1c7Q5Iounk_9uOPA4fS1VItR8F5yJzz9cziHfAHfFxtvo2XJ_mbY/s1600/TSM4.JPG&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;208&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhqh519H9eb6vuDfPpVmzgZe4UbYjQaXdHLWSq4V9tGxx-OFIHMp6HLP5Nuus3rG9kSJgSvXQZ3Sn_mXyq3F_7X3Ll1c7Q5Iounk_9uOPA4fS1VItR8F5yJzz9cziHfAHfFxtvo2XJ_mbY/s640/TSM4.JPG&quot; width=&quot;640&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
The key to recovery planning is to have good, and up-to-date information. If you set yourself up to require a multitude of login / password pairs, you are setting yourself up for failure- always look for ways to reduce the amount of information you need to perform your recovery. This simple trick can greatly reduce the number of credentials you need to track for recoveries where TSM passowrd authentication is used in the production environment.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/8558847629434585743/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/10/disable-password-authentication-for-tsm.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/8558847629434585743'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/8558847629434585743'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/10/disable-password-authentication-for-tsm.html' title='Disable password authentication for TSM (and a good Disaster Recovery tip)'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjJaZRlwhyphenhyphenJ1wImImsk6R6qVXSpXTDPjjU39q0dsq-MJYzE_8-as_XLLVnV_su1tpoTuv5VB5Hw1UzjndKfWMIW0lg59pToJh5kpyWRpmjJg94vLOPJnjPSojov-zVKriPgzqGSlN8_Rvc/s72-c/TSM_logo.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-4333419953005759159</id><published>2013-09-26T16:56:00.002-04:00</published><updated>2013-09-26T16:56:44.009-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="CentOS"/><category scheme="http://www.blogger.com/atom/ns#" term="firewall"/><category scheme="http://www.blogger.com/atom/ns#" term="iptables"/><category scheme="http://www.blogger.com/atom/ns#" term="Linux"/><category scheme="http://www.blogger.com/atom/ns#" term="Network"/><category scheme="http://www.blogger.com/atom/ns#" term="Red Hat"/><title type='text'>Disable iptables Firewall in CentOS/RHEL</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiLiZoX8wghvMGTQ9759H4dqehLBFmhdZ-WlsqMbvBmwUAbIl7ZIEcppI9RyUjWDeWtJ8NBNBxjaUFVvo-fLVO1J9HbqsnYPaet7kjcCQyQylq7z9cgE2cxyaRBeS9CZ9EpU9utMXHKxZY/s1600/Centos-Logo.png&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;181&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiLiZoX8wghvMGTQ9759H4dqehLBFmhdZ-WlsqMbvBmwUAbIl7ZIEcppI9RyUjWDeWtJ8NBNBxjaUFVvo-fLVO1J9HbqsnYPaet7kjcCQyQylq7z9cgE2cxyaRBeS9CZ9EpU9utMXHKxZY/s320/Centos-Logo.png&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
&lt;br /&gt;
I ran across an interesting problem while trying to install/configure the NetBackup client software on a newly deployed system running CentOS 6. I could install the client software just fine, however, despite my best attempts, I could not connect to the client from my NetBackup master server and run backups- anything I tried resulted in a &quot;cannot connect on socket&quot; error.&lt;br /&gt;
&lt;br /&gt;
I check all the usual obvious problems- on both the CentOS client and the Solaris master server, I did forward and reverse name lookups for both names, and everything appeared ok. I had ping connectivity both ways, and could ssh between the systems without issue. &lt;br /&gt;
&lt;br /&gt;
In short, I was stumped.&lt;br /&gt;
&lt;br /&gt;
Out of desperation, I logged into the master server and ran a snoop -d on the primary network interface and let is sit there on my desktop while I hacked away some more. That&#39;s when I noticed a conspicuous message fly by in the snoop output whenever I tried anything that made the master server try to connect to the client:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;ICMP Destination unreachable (Host administratively prohibited)&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
The message would fly by consistantly, any time I tried to initiate a backup/restore, or even have the NetBackup UI probe the client for it&#39;s software version. This was the clue I needed. I knew there were no firewalls between the two systems (they were on the same physical neetwork), and there were no crazy ACL settings on the switch either, so something had to be intentionally blocking traffic on the very ports NetBackup needed to do what it does.&lt;br /&gt;
&lt;br /&gt;
After a few google runs, and some more digging, I came across references to iptables- the firewall software built into many RHEL-like linux builds. Apparently the default settings for iptables are a little agressive and can wreak havoc on those who are not &quot;in the know&quot;... &lt;br /&gt;
&lt;br /&gt;
Luckily, turning this little monster off is a simple procedure- first, make sure the iptables config is saved in case you need to turn it back on at some point:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /etc/init.d/iptables save&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Stop the current running instance of iptables:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /etc/init.d/iptables stop&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Finally, make sure iptables does not restart if you boot the system:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# chkconfig iptables off&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Like magic, my NetBackup client began to work. Problem solved. Of course- this is fine if you are inside a network that has decent perimeter defenses against the untamed wilds of the internet. If you plan on having your system facing the internet directly, you may want to do a little more research into the &quot;dark art&quot; of iptables configuration to make sure you&#39;re protected, but are able to pass communications that you want (like backups) at the same time. The &lt;a href=&quot;http://wiki.centos.org/HowTos/Network/IPTables&quot; target=&quot;_blank&quot;&gt;CentOS Wiki&lt;/a&gt; site is a good place to start delving deeper if you want to do more than turn the service on and off.</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/4333419953005759159/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/09/disable-iptables-firewall-in-centosrhel.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4333419953005759159'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4333419953005759159'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/09/disable-iptables-firewall-in-centosrhel.html' title='Disable iptables Firewall in CentOS/RHEL'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiLiZoX8wghvMGTQ9759H4dqehLBFmhdZ-WlsqMbvBmwUAbIl7ZIEcppI9RyUjWDeWtJ8NBNBxjaUFVvo-fLVO1J9HbqsnYPaet7kjcCQyQylq7z9cgE2cxyaRBeS9CZ9EpU9utMXHKxZY/s72-c/Centos-Logo.png" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-6333354328982976261</id><published>2013-09-05T17:58:00.000-04:00</published><updated>2013-09-05T17:58:04.280-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Branded Zones"/><category scheme="http://www.blogger.com/atom/ns#" term="Oracle"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 8"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 9"/><category scheme="http://www.blogger.com/atom/ns#" term="ZFS"/><category scheme="http://www.blogger.com/atom/ns#" term="zoneadm"/><category scheme="http://www.blogger.com/atom/ns#" term="zonecfg"/><category scheme="http://www.blogger.com/atom/ns#" term="Zones"/><title type='text'>Use a ZFS file system on a Solaris 8 (or 9) host (Sort of)</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s1600/zfs_feature_2.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;160&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s200/zfs_feature_2.jpg&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
Yes, it is actually possible to use ZFS file systems on legacy versions of Solaris. The trick is that you cannot do it on a physical server. If you are using branded zones to run virtual instances of Solaris 8 or Solaris 9 hosts, you can attach zfs filesystems to them. The only limitation is that the zfs/zpool management needs to be done from the Solaris 10 global zone (physical server) itself rather than from within the virtualized OS instance. In order to directly manage the zfs/zpool properties and configuration (also called &quot;delegated&quot; file systems) you would need a virtual OS instance that contains the zfs and zpool packages- which is not the case for pre-10 Solaris. Mount access to zfs file systems is, however, very possible since non-delegated mode basically makes the file system (zfs or otherwise) a virtual object akin to a UNIX file handle that is accessed the same way regardless of underlying technology.&lt;br /&gt;
&lt;br /&gt;
In order to begin, you&#39;ll need a Solaris 10 system to host your zone on. You&#39;ll need at least one empty disk to host the zone root, and/or the zfs file systems you intend to add. You can create a Solaris 8 or Solaris 9 zone from just about any flash archive once you have a few prerequisites in place.&lt;br /&gt;
&lt;br /&gt;
You will note that base installations of Solaris 10 have two packages pre-installed to support each flavor of branded zones available (see the Solaris 8 packeage in the example below- a similar set of packages exist for Solaris 9 zones)&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# pkginfo | grep SUNWs8&lt;br /&gt;system&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; SUNWs8brandr&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;Solaris 8 Containers: solaris8 brand support (Root)&lt;br /&gt;system&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; SUNWs8brandu&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;Solaris 8 Containers: solaris8 brand support (Usr)&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
What you&#39;ll rapidly discover is that there is a third package for each type of brandeed zone, the &quot;brandk&quot; package, also known as the &quot;right-to-use&quot; enabler. This package must be downloaded from Oracle&#39;s download site- the distribution usually includes both the Solaris 8 and Solaris 9 packages. See Oracle&#39;s notes and installation instructions&amp;nbsp;&lt;a href=&quot;http://docs.oracle.com/cd/E22645_01/html/820-2914/gftrg.html&quot; target=&quot;_blank&quot;&gt;here&lt;/a&gt; or go directly to the download site &lt;a href=&quot;https://edelivery.oracle.com/&quot; target=&quot;_blank&quot;&gt;here&lt;/a&gt;. There is an ominous statement about requiring a signed&amp;nbsp; contract- and it&#39;s true. You will not be able to access the download site without an Oracle account that is linked to a valid service contract. Items such as this used to be pblically available, but those days are long gone.&lt;br /&gt;
&lt;br /&gt;
Once your enabler package is installed, you should see a &quot;brandk&quot; package listed as well.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# pkginfo | grep SUNWs8&lt;br /&gt;system&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; SUNWs8brandk&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Solaris 8 Containers: solaris8 brand support RTU&lt;br /&gt;system&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; SUNWs8brandr&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Solaris 8 Containers: solaris8 brand support (Root)&lt;br /&gt;system&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; SUNWs8brandu&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;Solaris 8 Containers: solaris8 brand support (Usr)&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;/span&gt;&lt;br /&gt;
You can now go ahead and begin to create your Solaris 8 zone. First you&#39;ll need a directory for your zone root. The easiest way to do this is to define a zpool, then name a zfs dataset under it when you define your zonepath variable. The zonecfg command will create the underlying dataset for you without intervention. Alternatively, you can create a directory anywhere on your system that has enough available space, give that directory rwx------ permissions (700), then use that as your zonepath value.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zpool create zones c1t1d0s2&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Note that when we create out zone with zonecfg, we need to define a &quot;brand&quot; variable, and set an attribute to track the architecture type of the flash archive we&#39;ll be virtualizing (sun4u).&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zonecfg -z zone8&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zone8: No such zone configured&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Use &#39;create&#39; to begin configuring a new zone.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8&amp;gt; create -b&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8&amp;gt; set zonepath=/zones/zone8&lt;br /&gt;zonecfg:zone8&amp;gt; set brand=solaris8&lt;br /&gt;zonecfg:zone8&amp;gt; set autoboot=true&lt;br /&gt;zonecfg:zone8&amp;gt; set ip-type=shared&lt;br /&gt;zonecfg:zone8&amp;gt; add net&lt;br /&gt;zonecfg:zone8:net&amp;gt; set address=192.168.167.252&lt;br /&gt;zonecfg:zone8:net&amp;gt; set physical=igb0&lt;br /&gt;zonecfg:zone8:net&amp;gt; end&lt;br /&gt;zonecfg:zone8&amp;gt; add attr&lt;br /&gt;zonecfg:zone8:attr&amp;gt; set name=machine&lt;br /&gt;zonecfg:zone8:attr&amp;gt; set type=string&lt;br /&gt;zonecfg:zone8:attr&amp;gt; set value=sun4u&lt;br /&gt;zonecfg:zone8:attr&amp;gt; end&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8&amp;gt; commit&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8&amp;gt; exit&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
Note that we configured the basics here, the bare minium needed to define a zone, but did not add any file systems yet- we&#39;ll add those separately for demonstration purposes.&lt;br /&gt;
&lt;br /&gt;
Next, we can create a new zfs file system to add to our new zone. Just for the sake of simplicity we&#39;ll create it under the same zpool as our zonepath. Note that we will create the file system with the mountpoint=legacy option - this means that the zfs file system will not be mounted in the global zone unless we specifically force it to mount in the /etc/vfstab. This ensures that the file system is free to be used by our zone.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zfs create -o mountpoint=legacy zones/zonefs1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Now comes the importrant part- attaching the file system to the zone. There are actually three ways to attach a file system to a zone. We can attach a file system or directory that is actually mounted on the global zone as a loopback file system, which gives it a presence both within the global, and the non-global zone. We can also give the non-global zone more direct control over the file system by attaching the underlying raw device (delegated mode), or we can attach a filesystem&amp;nbsp;but&amp;nbsp;have the global zone retain&amp;nbsp;direct control over file system properties and volume mangement.&lt;br /&gt;
&lt;br /&gt;
Loopback file systems (lofs) is&amp;nbsp;the most common file system attachment method, so we won&#39;t go into details. Theoretically, we could attach a zfs filesystem to literally anything as an lofs, but lofs introduces other complications, and usually forces backups to be managed at the global zone level.&lt;br /&gt;
Delegated mode for a zfs file system will only work if the zone can also support zfs natively. Since legacy versions of Solaris did not support zfs directly, you&#39;ll be greeted with a friendly error message if you try to use delegated mode (add a file systems with the &quot;add dataset&quot; subcommand):&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg -z zone8&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8&amp;gt; add dataset&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8:dataset&amp;gt; set name=zones/zonefs1&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8:dataset&amp;gt; end&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8&amp;gt; commit&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;solaris8_support error: solaris8 zones do not support ZFS datasets&lt;br /&gt;zone8: Brand-specific error&lt;br /&gt;zonecfg:zone8&amp;gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;span style=&quot;font-family: inherit;&quot;&gt;You&#39;ll want to exit zonecfg without saving, then try again, only using the &quot;add fs&quot; metehod instead:&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg -z zone8&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8&amp;gt; add fs&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8:fs&amp;gt; set type=zfs&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;zonecfg:zone8:fs&amp;gt; set special=zones/zonefs1&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;zonecfg:zone8:fs&amp;gt; set dir=/fs01&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8:fs&amp;gt; end&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:zone8&amp;gt; commit&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;This says to add a zfs file system zones/zonefs1, and mount it in the zone as /fs01. This is very similar to how we would add a loopback file system (for that we would use type=lofs).&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Now that we&#39;ve added our file system, we can install a solaris 8 OS in the zone from a flash archive. The zoneadm command will perform a physical to virtual conversion once the flash archive is extracted. Note that you will need to specify the absolute path to the flash archive, even if you are already in the directory containing the archive when you run the zoneadm install.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;br /&gt;
&lt;/span&gt;&lt;/span&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;
# zoneadm&amp;nbsp;-z zone8 install -u -a /net/server/sol8_image.flar&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&lt;br /&gt;
&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &amp;nbsp; Log File: /var/tmp/zone8.install.21207.log
&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &amp;nbsp;&amp;nbsp; Source: /net/machinename/sol8_image.flar
&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &amp;nbsp;Installing: This may take several minutes...
&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;Postprocessing: This may take a minute...

&amp;nbsp;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp; &amp;nbsp;Result: Installation completed successfully.
&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; &lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp; Log File: /export/home/zone8/root/var/log/zone8.install.21207.log&lt;/span&gt;
&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;br /&gt;
The &quot;-u&quot; option &amp;nbsp;forces a sys-unconfig the first time the zone boots, so you&#39;ll need to zlogin -C to the console and answer the usual solaris questions the first time you boot. If you want to keep the sysidcfg information contained in the original flash archive, replace the &quot;-u&quot; with a &quot;-p&quot;. Once the install completes, you should be ready to boot and check your work&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zoneadm -z zone8 boot&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
If you used the -u option, connect to the zone&#39;s console session and go through the sysidcfg questions.&lt;br /&gt;
&lt;br /&gt;
Once you&#39;ve done that, you should be able to zlogin to the zone and see your file system attached:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zlogin zone8&lt;br /&gt;[Connected to zone &#39;zone8&#39; pts/2]&lt;br /&gt;Last login: Thu Sep&amp;nbsp; 5 17:52:36 on pts/2&lt;br /&gt;Sun Microsystems Inc.&amp;nbsp;&amp;nbsp; SunOS 5.8&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Generic Patch&amp;nbsp;&amp;nbsp; February 2004&lt;br /&gt;zone8#&lt;br /&gt;zone8# uname -a&lt;br /&gt;SunOS zone8 5.8 Generic_Virtual sun4u sparc sun4v&lt;br /&gt;zone8#&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zone8# df -k /fs01&lt;br /&gt;Filesystem&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; kbytes&amp;nbsp;&amp;nbsp;&amp;nbsp; used&amp;nbsp;&amp;nbsp; avail capacity&amp;nbsp; Mounted on&lt;br /&gt;zones/zonefs1&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 282221243&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 31 282221212&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 1%&amp;nbsp;&amp;nbsp;&amp;nbsp; /fs01&lt;br /&gt;zone8#&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/6333354328982976261/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/09/use-zfs-file-system-on-solaris-8-or-9.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/6333354328982976261'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/6333354328982976261'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/09/use-zfs-file-system-on-solaris-8-or-9.html' title='Use a ZFS file system on a Solaris 8 (or 9) host (Sort of)'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s72-c/zfs_feature_2.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-5261411513041539831</id><published>2013-08-27T17:17:00.002-04:00</published><updated>2014-04-21T13:34:57.089-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="attach"/><category scheme="http://www.blogger.com/atom/ns#" term="Containers"/><category scheme="http://www.blogger.com/atom/ns#" term="detach"/><category scheme="http://www.blogger.com/atom/ns#" term="Disaster Recovery"/><category scheme="http://www.blogger.com/atom/ns#" term="recover"/><category scheme="http://www.blogger.com/atom/ns#" term="Restore"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><category scheme="http://www.blogger.com/atom/ns#" term="Virtualization"/><category scheme="http://www.blogger.com/atom/ns#" term="Zones"/><title type='text'>Recover / Migrate Solaris 10 zones to different physical hosts without performing a detach or export</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi9SOj4c7KuHDIX4SRU_mw4spfH0NFi66Nw9kBja2si_Ic4knrcDuVaiIDPlsLsLr0FUuMVCl6D8u7WRHZrDy8vEemjq2o0uo4ogfTWW6jJEtsn-s7D__IMx9H4GfPIcGl3vhYB6RTZ6lQ/s1600/oralogo-small.gif&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi9SOj4c7KuHDIX4SRU_mw4spfH0NFi66Nw9kBja2si_Ic4knrcDuVaiIDPlsLsLr0FUuMVCl6D8u7WRHZrDy8vEemjq2o0uo4ogfTWW6jJEtsn-s7D__IMx9H4GfPIcGl3vhYB6RTZ6lQ/s1600/oralogo-small.gif&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Sun intended that if you ever had to move a zone from one physical host to another, you would have time to do an orderly shutdown, perform a zoneadm detach, capture the config file that this generates, and move it to the new system. This is very impractical in a disaster recovery situation, and is also unnecessary since all the data you need to define a zone is contained in it&#39;s zone root or zonepath directory, and the configuration data stored in the zone.xml file under /etc/zones.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Sun, however, did not give us a simple way to&amp;nbsp;move a zone between systems when the original system is unavailable, unless you happened to capture an export of your zone configuration (or performa a formal &quot;detatch&quot;). However, as long as you have the system backed up properly, you can still recover. If you plan on recovering your zone to the original global zone host, no problem- you can get all the pieces you need with minimal fuss simply be recovering the global zone.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
In a disaster recovery scenario, however, you may need to improvise. Hardware constraints may force you to mix and match zones on different globaal zone hosts. Perhaps you have a global zone host at home that hosts a alrge number of zones, and has an accordingly high amount of memory. At your DR site, you may not have a single system with that quantity of memory, so you may need to plan on splitting the big hosts&#39;s zones between two or three smaller systems to keep the load balanced. You may decide to pslit up your zones into more servers to balance the network throughput for your restores if you&#39;re a shop that recovers from tape / VTL / or similar traditional backup products. Either way, you&#39;ll have a very real need to quickly port your zones from one system to another with only the file system data you have in your backups.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Luckily, the only pieces you need, should be covered by those backups. The most critical pieces of data is the zone&#39;s xml file from /etc/zones, and the file system data- particularly for he zonepath directory. You&#39;ll also need the backups of any additional data file systems required, but that is the relatively easy part. All these items would normally be included in any full backup of a system, so there is certainly a way to do it, even if you need to break warranties to get there.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Logically, you would think that all you have to do is recreate the zonepath directory, and restore data to it on the new host, then restore the &amp;lt;zone name&amp;gt;.xml file found in /etc/zones on the original system to /etc/zones on the new system and run an attach operation. You would, of course, also need to change the network devices assigned to the zone to match your recovery environment, and you&#39;d need to recreate and restore any additional file systems assigned to the zone, but for now lete&#39;s take a simple case with no network changes, and no additional file systems to deal with.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zoneadm -z zone3 attach&lt;br /&gt;zoneadm: zone3: could not get state: No such zone configured&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;/span&gt;&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;You didn&#39;t think it would be that simple did you? &lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
The problem here is that the zoneadmd process tracks the state of all known zones and their unique zone ID numbers (which is usually an under the hood item that you wouldn&#39;t see in day-to-day operations). The state and basic details for all valid zones&amp;nbsp;are tracked in the /etc/zones/index file- if there is no entry there, you will never be able to attach and boot your zone, since zoneadmd won&#39;t know that it exists.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
If you look in an existing index file, you will see an entry for each zone that includes the zone name, the state of the zone, the zonepath directory, and a system-generated zone ID number.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
The most simple trick to make your zone attachable, is to add a dummy entry to the index file for the zone you are attempting to port onto the new host, but make sure you put it in the &quot;configured&quot; state, and leave off the zone ID number (it will be assigned automatically later). Simply follow the example below where I&#39;ve added a dummy entry for&amp;nbsp;a zone called &quot;zone3&quot; with a zonepath directory (which I&#39;ve already created and restored) called /zones/zone3.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# vi /etc/zones/index&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&lt;/span&gt;&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;#&lt;br /&gt;# DO NOT EDIT: this file is automatically generated by zoneadm(1M)&lt;br /&gt;# and zonecfg(1M).&amp;nbsp; Any manual changes will be lost.&lt;br /&gt;#&lt;br /&gt;global:installed:/&lt;br /&gt;zone1:installed:/zones/zone1:e391de3c-7e93-eb1e-b9a9-fcef2bf3ad52&lt;br /&gt;zone2:installed:/zones/zone2:d2adac51-1dc3-4c6a-a5b0-f2ae4623a7a9&lt;br /&gt;&lt;span style=&quot;background-color: yellow;&quot;&gt;zone3:configured:/zones/zone3:&lt;/span&gt;&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
The first thing you&#39;ll notice is the warning message that you should not edit this file manually. It&#39;s OK, Sun/Oracle is just trying to scare you into doing things their way, rather than the faster way... as long as you do not stop/start or otherwise configure or change a zone before you do your attach, you have nothing to worry about. You can get away with a&amp;nbsp;warranty-breaking edit if you do things in the right order.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
If you have not already done so, copy or restore the zone&#39;s original xml file to /etc/zones&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# pwd&lt;br /&gt;/etc/zones&lt;br /&gt;# ls -l zone3.xml&lt;br /&gt;-rw-r--r--&amp;nbsp;&amp;nbsp; 1 root&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; root&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 337 Aug 27 16:16 zone3.xml&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Again, you&#39;re not officially supposed to edit the xml files directly, but if you needed to adjust the network interface assignments for the zone, you could do so now by editing the xml file before you attempt to boot or attach the zone.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# vi /etc/zones/zone3.xml&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&lt;/span&gt;&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;lt;?xml version=&quot;1.0&quot; encoding=&quot;UTF-8&quot;?&amp;gt;&lt;br /&gt;&amp;lt;!DOCTYPE zone PUBLIC &quot;-//Sun Microsystems Inc//DTD Zones//EN&quot; &lt;/span&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;file:///usr/shar&lt;/span&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;e/lib/xml/dtd/zonecfg.dtd.1&amp;gt;&lt;br /&gt;&amp;lt;!--&lt;br /&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; DO NOT EDIT THIS FILE.&amp;nbsp; Use zonecfg(1M) instead.&lt;br /&gt;--&amp;gt;&lt;br /&gt;&amp;lt;zone name=&quot;zone3&quot; zonepath=&quot;/zones/zone3&quot; autoboot=&quot;true&quot;&amp;gt;&lt;br /&gt;&amp;nbsp; &amp;lt;network address=&quot;192.168.167.112/24&quot; physical=&quot;&lt;span style=&quot;background-color: yellow;&quot;&gt;igb0&lt;/span&gt;&quot;/&amp;gt;&lt;br /&gt;&amp;lt;/zone&amp;gt;&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Typically for a shared IP zone (which is what you will find in most cases) you may have to change the physical device name when recovering to match whichever network interface you intend to use on the recovered system. Again, ignore the warnings, as long as you make your edits while the zone is not booted, and as long as you don&#39;t make any horrible typos, you can ignore the DO NOT EDIT warning. the zonecfg subcommands are very powerful, but for a simple operation like changing a network interface name from &quot;e1000g0&quot; to &quot;igb0&quot; - you&#39;ll find it easier to just edit the file while the zone is halted/down/etc.&amp;nbsp; If your zone is an exclusive IP zone, you will need to make a similar edit, although you will also need to adjust your hostname.&amp;lt;interface&amp;gt; files within the zone&#39;s /etc directory to match once your zone is booted.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Now, with these few simple intermediate steps, you can actually perform your attach operation, whcih will change the state in the index file from &quot;configured&quot; to &quot;installed&quot;, and assign a zone ID number as well. This will put your zone into a state where it can be booted.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zoneadm -z zone3 attach&lt;br /&gt;#&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Notice the lack of error messages this time. Note that if your new host system is a newer Solaris release or patch level, in some cases you may need to use the -u or -U options to update patches during the attach, in most cases, especially for whole root zones, this is not necessary.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
How can you tell if it worked without actually booting your zone? Simply take another look at the index file and note the differences in the entry for your zone:&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cat /etc/zones/index&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# Copyright 2004 Sun Microsystems, Inc.&amp;nbsp; All rights reserved.&lt;br /&gt;# Use is subject to license terms.&lt;br /&gt;#&lt;br /&gt;# ident &quot;@(#)zones-index&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 1.2&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 04/04/01 SMI&quot;&lt;br /&gt;#&lt;br /&gt;# DO NOT EDIT: this file is automatically generated by zoneadm(1M)&lt;br /&gt;# and zonecfg(1M).&amp;nbsp; Any manual changes will be lost.&lt;br /&gt;#&lt;br /&gt;global:installed:/&lt;br /&gt;zone1:installed:/zones/zone1:e391de3c-7e93-eb1e-b9a9-fcef2bf3ad52&lt;br /&gt;zone2:installed:/zones/zone2:d2adac51-1dc3-4c6a-a5b0-f2ae4623a7a9&lt;br /&gt;&lt;span style=&quot;background-color: yellow;&quot;&gt;zone3:installed:/zones/zone3:4dc8a587-d7ec-cc25-8f77-ca6d69dc5cd1&lt;/span&gt;&lt;br /&gt;#&lt;/span&gt;&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Notice that the entry for zone3 has changed to the &quot;installed&quot; state, and now has a zone ID string associated with it. It will now also show up in a zoneadm list as &quot;installed&quot;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zoneadm list -iv&lt;br /&gt;&amp;nbsp; ID NAME&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; STATUS&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; PATH&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; BRAND&amp;nbsp;&amp;nbsp;&amp;nbsp; IP&lt;br /&gt;&amp;nbsp;&amp;nbsp; 0 global&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; running&amp;nbsp;&amp;nbsp;&amp;nbsp; /&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; native&amp;nbsp;&amp;nbsp; shared&lt;br /&gt;&amp;nbsp;&amp;nbsp; 6 zone1&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; running&amp;nbsp;&amp;nbsp;&amp;nbsp; /zones/zone1&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; native&amp;nbsp;&amp;nbsp; shared&lt;br /&gt;&amp;nbsp;&amp;nbsp; 8 zone2&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; running&amp;nbsp;&amp;nbsp;&amp;nbsp; /zones/zone2&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; native&amp;nbsp;&amp;nbsp; shared&lt;br /&gt;&amp;nbsp;&amp;nbsp; &lt;span style=&quot;background-color: yellow;&quot;&gt;- zone3&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; installed&amp;nbsp; /zones/zone3&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; native&amp;nbsp;&amp;nbsp; shared&lt;/span&gt;&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
If your zone has any additional file systems, now would be a good time to make sure you have recreated them. You can usually tell what the requirements are by examining any file system definintions in the zone&#39;s xml file, and comparing them to a df -h taken from the zone. Depending on how much your zone uses lofs file systems will determine if you can restore the additional file systems now, or if you need to wait until your zone is booted and running. I personally prefer zones built using all lofs file systems, since you&#39;ll be able to recreate them on your global zone, and recover their data prior to booting your zone, which allows a cleaner boot and less troubleshooting, but your milage may vary depending on the quirks of the environment you are attempting to recover or migrate.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Asusming we have no additional file systems to tend to, we can go ahead and boot our zone. Note that if the zone is a &quot;branded&quot; zone running Solaris 8, or Solaris 9, you will get an initial error stating that you need to run the appropriate p2v script against the zone first. Luckily, the error message will tell you pretty much exactly what to run. In most cases, the only actions the p2v script will do is update internal records to say that the branded zone was properly converted and imported. Normally it only needs to update a zone if it was created from a flash archive of a physical system directly- something you wouldn&#39;t typically run into in most recovery scenarios (although it can be done, and can be useful in certain situations). If your zone is a native Soalris 10 zone, it should boot up without any issues, unless you skipped something or there is some unusual quirk.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zoneadm -z zone3 boot&lt;br /&gt;#&lt;/span&gt;&lt;/div&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;# zoneadm list -iv&lt;br /&gt;&amp;nbsp; ID NAME&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; STATUS&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; PATH&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; BRAND&amp;nbsp;&amp;nbsp;&amp;nbsp; IP&lt;br /&gt;&amp;nbsp;&amp;nbsp; 0 global&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; running&amp;nbsp;&amp;nbsp;&amp;nbsp; /&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; native&amp;nbsp;&amp;nbsp; shared&lt;br /&gt;&amp;nbsp;&amp;nbsp; 6 zone1&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; running&amp;nbsp;&amp;nbsp;&amp;nbsp; /zones/zone1&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; native&amp;nbsp;&amp;nbsp; shared&lt;br /&gt;&amp;nbsp;&amp;nbsp; 8 zone2&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; running&amp;nbsp;&amp;nbsp;&amp;nbsp; /zones/zone2&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; native&amp;nbsp;&amp;nbsp; shared&lt;br /&gt;&amp;nbsp; &lt;span style=&quot;background-color: yellow;&quot;&gt;16 zone3&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; running&amp;nbsp;&amp;nbsp;&amp;nbsp; /zones/zone3&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; native&amp;nbsp;&amp;nbsp; shared&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;#&lt;br /&gt;# zlogin zone3&lt;br /&gt;[Connected to zone &#39;zone3&#39; pts/2]&lt;br /&gt;Last login: Tue Aug 27 16:01:26 on pts/2&lt;br /&gt;Oracle Corporation&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; SunOS 5.10&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Generic Patch&amp;nbsp;&amp;nbsp; January 2005&lt;br /&gt;#&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;span style=&quot;font-family: inherit;&quot;&gt;&lt;/span&gt;&amp;nbsp;&lt;/span&gt;&lt;/div&gt;
&lt;span style=&quot;font-family: Courier New;&quot;&gt;&lt;/span&gt;&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/5261411513041539831/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/08/recover-migrate-solaris-10-zones-to.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5261411513041539831'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5261411513041539831'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/08/recover-migrate-solaris-10-zones-to.html' title='Recover / Migrate Solaris 10 zones to different physical hosts without performing a detach or export'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi9SOj4c7KuHDIX4SRU_mw4spfH0NFi66Nw9kBja2si_Ic4knrcDuVaiIDPlsLsLr0FUuMVCl6D8u7WRHZrDy8vEemjq2o0uo4ogfTWW6jJEtsn-s7D__IMx9H4GfPIcGl3vhYB6RTZ6lQ/s72-c/oralogo-small.gif" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-5331861692331139095</id><published>2013-07-17T15:26:00.000-04:00</published><updated>2013-08-21T11:59:26.218-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="ifconfig"/><category scheme="http://www.blogger.com/atom/ns#" term="IP Address"/><category scheme="http://www.blogger.com/atom/ns#" term="Linux"/><category scheme="http://www.blogger.com/atom/ns#" term="Network"/><category scheme="http://www.blogger.com/atom/ns#" term="Red Hat"/><category scheme="http://www.blogger.com/atom/ns#" term="Static"/><title type='text'>Red Hat Enterprise Linux Static IP Configuration</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhGGgG7CCkmWOUnEwg0iX1LOnqBrvLWEnsJZ8Z4zRfuGK2DEIt_JJBHlIEruKQ9ofY1e7NQzyobfpeStmnQF-eoAR7iW8yabSHJebMdiusFSIf3RnmAm2P2JzEvY65d7p-hjfHR1oxyAUY/s1600/redhat-logo.jpeg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;200&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhGGgG7CCkmWOUnEwg0iX1LOnqBrvLWEnsJZ8Z4zRfuGK2DEIt_JJBHlIEruKQ9ofY1e7NQzyobfpeStmnQF-eoAR7iW8yabSHJebMdiusFSIf3RnmAm2P2JzEvY65d7p-hjfHR1oxyAUY/s200/redhat-logo.jpeg&quot; width=&quot;181&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
This is a simple task that any UNIX/LINUX admin should be able to do with no assistance. I, however, have reached a point in my career where I touch too many different things to remember the little nuances particular to each operating system or technology that I deal with on a regular basis- so this article, while covering something very basic, is something that I need as my personal note pad so I don&#39;t have to go searching around the next time I have to build a Red Hat system.&lt;br /&gt;
&lt;br /&gt;
The instructions here are meant for Linux servers running Red Hat 6, although they most likely apply to other flavors of linux as well (i.e. fedora, and CentOS). I have trouble remembering every little detail since it&#39;s similar to the other UNIXes I work with, but still different enough that there are details that evade my attempts to store permanently in my brain. Conceptually, the task is the same for any UNIX or LINUX, but again, the devil is in the details.&lt;br /&gt;
&lt;br /&gt;
Most servers and/or VMs are deployed with DHCP addressing turned on by default. Maybe it&#39;s my Solaris/AIX/HP-UX background, but I just can&#39;t stand DHCP in general, at least for a server environment (it&#39;s great for a workstation/end user environment, but that&#39;s a different animal). What you&#39;ll need to do is modify the proper network configuration files to disable dhcp for your specific network interface, and configure the correct static IP address, netmask, and/or gateway addresses (and DNS if needed as well). In some UNIXes, this means making smaller edits in several different configuration files. In RedHat, however, most of the work is done in a single place, with one or two minor exceptions.&lt;br /&gt;
&lt;br /&gt;
The first thing you need to check is to make sure networking services are turned on. Look at the /etc/sysconfig/network file- it should indicate &quot;yes&quot; for networking, and list the server&#39;s hostname (usually fully qualified), and optionally, a default gateway address- such as seen in the example below.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; # cat /etc/sysconfig/network&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; NETWORKING=yes&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; HOSTNAME=myhost.mydomain.local&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; GATEWAY=10.10.10.254&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
If the &quot;NETWORKING&quot; parameter is not set to yes, change it to match the above or networking services will not be started during boot. While here, verify that your hostname is listed correctly, and validate your default gateway address. You may or may not have a gateway address defined here- it can also be defined in the configuration file for the specific network interface you are using as well.&lt;br /&gt;
&lt;br /&gt;
Most of the work occurs in the configuration file for your actual network interface, found under /etc/sysconfig/network-scripts. You will find a number of files in this directory, including a file names in the form &quot;ifcfg-&amp;lt;interface name&amp;gt;&quot; - one for each physical interface present on the system. Identify which network interface you need to work with, and edit the corresponding file. In my examples, I will use the interface &quot;eth0&quot;, so we&#39;ll be editing ifcfg-eth0.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; # cd /etc/sysconfig/network-scripts&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; # ls ifcfg-*&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; ifcfg-eth0&amp;nbsp; ifcfg-lo&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; # vi ifcfg-eth0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
You will probably find that the file is already populated with several entries similar to the below:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; DEVICE=eth0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; HWADDR=00:50:56:B3:00:15&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; TYPE=Ethernet&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; UUID=ee5d8111-9368-486b-96a6-39295790cd44&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; ONBOOT=yes&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; NM_CONTROLLED=yes&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; BOOTPROTO=dhcp&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Basically, you have a record of the interface name, hadware mac address, UUID number, and other control settings (i.e. dhcp). In order to change this interface to use a static IP address, we need to change the &quot;BOOTPROTO&quot; variable to &quot;static&quot;, add our IP address, netmask, and (optionally) gateway address, as shown in the example below:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; DEVICE=eth0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; HWADDR=00:50:56:B3:00:15&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; TYPE=Ethernet&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; UUID=ee5d8111-9368-486b-96a6-39295790cd44&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; ONBOOT=yes&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; NM_CONTROLLED=yes&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; BOOTPROTO=static&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; IPADDR=10.10.10.27&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; NETMASK=255.255.255.0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; GATEWAY=10.10.10.254&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Finally, if you plan on using DNS, and it is not already configured, you&#39;ll need to set up your /etc/resolv.conf file as you normally would- for example:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; domain mydomain.local&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; search mydomain.local mydomain.net myotherdomain.net&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; nameserver 10.10.10.138&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; nameserver 10.10.10.139&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; nameserver 8.8.8.8&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Normally, your /etc/nsswitch.conf file will typically already resolve hosts to local files first, and dns second, you may want to check that as well- you should see the &quot;hosts&quot; entry in /etc/nsswitch.conf appear as the below:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; hosts:&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; files&amp;nbsp;&amp;nbsp; dns&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Once you have your static IP information,, and (optional) DNS information set, you simply need to restart network services, and validate that your network settings came up as expected:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; # /etc/init.d/network restart&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; # ifconfig eth0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; eth0&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; Link encap:Ethernet&amp;nbsp; HWaddr 00:50:56:B3:00:15&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; inet addr:10.10.10.27&amp;nbsp; Bcast:10.10.10.255&amp;nbsp; Mask:255.255.255.0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; inet6 addr: fe80::250:56ff:feb3:15/64 Scope:Link&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; UP BROADCAST RUNNING MULTICAST&amp;nbsp; MTU:1500&amp;nbsp; Metric:1&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; RX packets:2036 errors:0 dropped:0 overruns:0 frame:0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; TX packets:252 errors:0 dropped:0 overruns:0 carrier:0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; collisions:0 txqueuelen:1000&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; RX bytes:205045 (200.2 KiB)&amp;nbsp; TX bytes:40710 (39.7 KiB)&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; # ping www.google.com&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; PING www.google.com (173.194.75.106) 56(84) bytes of data.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 64 bytes from&amp;nbsp;mydomain.net (173.194.75.106): icmp_seq=1 ttl=45 time=20.1 ms&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 64 bytes from mydomain.net (173.194.75.106): icmp_seq=2 ttl=45 time=20.1 ms&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; ^C&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; --- www.google.com ping statistics ---&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 2 packets transmitted, 2 received, 0% packet loss, time 1868ms&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; rtt min/avg/max/mdev = 20.170/20.170/20.170/0.000 ms&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Pinging a remote server by name is a great test- if it is successful, you know you have base connectivity, name resolution, and routing functioning. If this type of test fails, then you can worry about more detailed troubleshooting.&lt;br /&gt;
&lt;br /&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/5331861692331139095/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/07/red-hat-enterprise-linux-static-ip.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5331861692331139095'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5331861692331139095'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/07/red-hat-enterprise-linux-static-ip.html' title='Red Hat Enterprise Linux Static IP Configuration'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhGGgG7CCkmWOUnEwg0iX1LOnqBrvLWEnsJZ8Z4zRfuGK2DEIt_JJBHlIEruKQ9ofY1e7NQzyobfpeStmnQF-eoAR7iW8yabSHJebMdiusFSIf3RnmAm2P2JzEvY65d7p-hjfHR1oxyAUY/s72-c/redhat-logo.jpeg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-9135060893344652565</id><published>2013-05-01T17:58:00.000-04:00</published><updated>2013-05-01T17:58:41.101-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Install"/><category scheme="http://www.blogger.com/atom/ns#" term="Linux"/><category scheme="http://www.blogger.com/atom/ns#" term="Maverick Meerkat"/><category scheme="http://www.blogger.com/atom/ns#" term="Package"/><category scheme="http://www.blogger.com/atom/ns#" term="Server"/><category scheme="http://www.blogger.com/atom/ns#" term="Ubuntu"/><category scheme="http://www.blogger.com/atom/ns#" term="Ubuntu 10.10"/><category scheme="http://www.blogger.com/atom/ns#" term="Unix"/><title type='text'>Linux (Ubuntu) package installation without apt-get</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgMBV5XDLiDT14EPsOo-52fUeDeuTMbON2rl6qsIN_r5BknjGopfMj5fh0AWraR5AGBxWrZJCotEXOy9Uo6U62sBCJ_CxnXX4SQJE1n9I0Y9koa8_ZtNqV15x9gm7O10o7uRKyxgDKOuus/s1600/600px-Former_Ubuntu_logo_svg.png&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;85&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgMBV5XDLiDT14EPsOo-52fUeDeuTMbON2rl6qsIN_r5BknjGopfMj5fh0AWraR5AGBxWrZJCotEXOy9Uo6U62sBCJ_CxnXX4SQJE1n9I0Y9koa8_ZtNqV15x9gm7O10o7uRKyxgDKOuus/s320/600px-Former_Ubuntu_logo_svg.png&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
One of my recent projects involved deploying a home-grown piece of software that required a fairly old, outdated linux build to use as a base- Ubuntu 10.10 (Maverick Meerkat). The problem I faced was that the hardware in question was pre-existing equipment that lives in a somewhat secure environment with limited direct access to the internet, aside from administrative access via VPN. The problem I found was that the legacy ubuntu build I used did not have an ssh server installed by default, so once I stood up the ubuntu build, I had no way to communicate with the official repositories (if one even still exists for 10.10) in order to install the openssh-server package.&lt;br /&gt;
&lt;br /&gt;
A few google searches quickly turned up answers- Ubuntu is a fairly well documented linux build that is more popular amongst independent enthusiasts than enterprise users (except for the developers I&#39;ve been working with!)- so there are an enormous number of web resources for help. The part that seemed daunting was the fact that most of them claimed that manual installation of packages was a tedious process. After going through it, and getting a feel for some of the potential issues, I actually found it relatively easy. The largest problem is determining package dependencies- which turns out to not be particularly difficult- especially for someone with a Solaris background- where understanding package and patch dependencies is almost considered one of the Dark Arts that requires the initiated to sign over rights to their very souls to Scott McNealy &amp;nbsp;(or Larry Ellison for the younger Solaris gurus out there).&lt;br /&gt;
&lt;br /&gt;
Ubuntu is derived from Debian, so still uses Debian style packages under the hood. Ubuntu software packages are stored as *.deb files that apt users never see. In practice, the *.deb files seem to be very similar to the Solaris *.pkg framework, minus the part where you sign over your soul to the current Sun/Oracle CEO.&lt;br /&gt;
&lt;br /&gt;
I was able to find packages for legacy Ubuntu builds on the internet fairly easily. A quick search on the &lt;a href=&quot;https://launchpad.net/ubuntu/maverick/i386/openssh-server/1:5.3p1-3ubuntu3&quot; target=&quot;_blank&quot;&gt;launchpad.net&lt;/a&gt; website led me to find the *.deb file for the openssh-server package.&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjbm9Znv9tdUrLukK76Tmdm7uIH5uWouOl1lNF1eUD79B-XXfZ3FBX55csp8hy3qOiE7kkLshAWziKW8vpM5tNjQSOAjRUGzc6hQziGCnaNMsIV6iCSV7OpU9JPGOO7WWtpNcFAhI8qKwU/s1600/openssh-server.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;118&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjbm9Znv9tdUrLukK76Tmdm7uIH5uWouOl1lNF1eUD79B-XXfZ3FBX55csp8hy3qOiE7kkLshAWziKW8vpM5tNjQSOAjRUGzc6hQziGCnaNMsIV6iCSV7OpU9JPGOO7WWtpNcFAhI8qKwU/s640/openssh-server.jpg&quot; width=&quot;640&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
The same page also listed the package dependencies- again, nothing overly arcane that puts ownership of my soul in peril.&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjC5-penWNnu3S_8aQ6rzqi8BWHwwhvr-06xxBxcyauT1ihLhSE7IAk6Xls39gdwuHonkER_HAnYQGDEoZD60-Q4lNiBjQjqeCIdQKBDYOwE3D3jF77fhhdRDcfsIb3bttNPlLaVDTIUbk/s1600/openssh-server2.JPG&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;444&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjC5-penWNnu3S_8aQ6rzqi8BWHwwhvr-06xxBxcyauT1ihLhSE7IAk6Xls39gdwuHonkER_HAnYQGDEoZD60-Q4lNiBjQjqeCIdQKBDYOwE3D3jF77fhhdRDcfsIb3bttNPlLaVDTIUbk/s640/openssh-server2.JPG&quot; width=&quot;640&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div align=&quot;left&quot; class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
﻿&lt;/div&gt;
So I downloaded the .deb file for the openssh-server package. I wasn&#39;t certain if I had version conflicts, or missing dependencies yet, so I kept the page open while I dropped the *.deb file on a USB stick and ran it over to the secure server.&lt;br /&gt;
&lt;br /&gt;
I copied the file over to /var/tmp for safekeeping, and ran the evil linux incantation to install the package (the Debian package management command- dpkg):&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$ cd /var/tmp&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$ ls&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;openssh-server_5.3p1-3ubuntu3_i386.deb&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$ sudo dpkg -i openssh-server_5.3p1-3ubuntu3_i386.deb&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Selecting previously deselected package openssh-server.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;(Reading database ... 118265 files and directories currently installed.)&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Unpacking openssh-server (from openssh-server_5.3p1-3ubuntu3_i386.deb) ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;background-color: yellow; font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&lt;strong&gt;dpkg: dependency problems prevent configuration of openssh-server:&lt;/strong&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;background-color: yellow; font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&lt;strong&gt;&amp;nbsp;openssh-server depends on openssh-client (= 1:5.3pl-3ubuntu3); however:&lt;/strong&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;background-color: yellow; font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&lt;strong&gt;&amp;nbsp;&amp;nbsp; Version of openssh-client on system is 1:5.5pl-4ubuntu4.&lt;/strong&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;dpkg: error processing openssh-server (--install):&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;dependency problems - leaving unconfigured&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Processing triggers for ufw ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Processing triggers for man-db ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Errors were encountered while processing:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;openssh-server&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Hmm, a failure, but somehow my soul was still firmly in my own hands. Perhaps this process is not so difficult after all. The error message actually seemed quite polite when compared to my Solaris experiences where a package can be left in a partially installed state, requiring manual surgery to back out of far enough to allow another attempt- even if you manage to figure out why the install failed. The highlighted portion of the output seemed to explain it all- the package was complaining that the version of the openssh-client currently on my system was not compatible with the version of openssh I grabbed. In fact, If I, fairly new to Ubuntu, am reading it right, it looks like the openssh-client I currently have is actually too new, and needs to be downgraded.&lt;br /&gt;
&lt;br /&gt;
So I wnet back to my browser window where I had my &lt;a href=&quot;https://launchpad.net/ubuntu/maverick/i386/openssh-server/1:5.3p1-3ubuntu3&quot; target=&quot;_blank&quot;&gt;package dependency page for openssh-server&lt;/a&gt; up, &amp;nbsp;and took a closer look- &lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjNZPByG-R7lXP6EgE-c2vRBAJcmXbg4XXYd0JZKrRhOEQ-VRR9F4EiUfzZDqGUJPRp6yv79HYt6-9-VZlx8BOMgF5g6JasVj9xh5rpvThJZH6pwMUv-PNzK_XC9FEXop0jo6FFEL17h2s/s1600/ssh3.JPG&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjNZPByG-R7lXP6EgE-c2vRBAJcmXbg4XXYd0JZKrRhOEQ-VRR9F4EiUfzZDqGUJPRp6yv79HYt6-9-VZlx8BOMgF5g6JasVj9xh5rpvThJZH6pwMUv-PNzK_XC9FEXop0jo6FFEL17h2s/s1600/ssh3.JPG&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
Sure enough, the dependency list shows that the required version of openssh-client differed from my current version. I followed the links provided, and after minimal fuss, found myself on the page containing &lt;a href=&quot;https://launchpad.net/ubuntu/lucid/i386/openssh-client/1:5.3p1-3ubuntu3&quot; target=&quot;_blank&quot;&gt;the version of openssh-client I was looking for&lt;/a&gt;:&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgOgymVb_f7ZSR7zz70Ed37lhVkczeSpoGlkiwc8VAcW3U2ryn3YMlUQmzQIgCeDZB06fkRUYnT0dfk67lMhX4g4DYx1ro9FjUWNlmZMsGyL74ezp_-k_sdXCYMD3Q4hyphenhyphenHbVZeSuPtqRms/s1600/ssh4.JPG&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgOgymVb_f7ZSR7zz70Ed37lhVkczeSpoGlkiwc8VAcW3U2ryn3YMlUQmzQIgCeDZB06fkRUYnT0dfk67lMhX4g4DYx1ro9FjUWNlmZMsGyL74ezp_-k_sdXCYMD3Q4hyphenhyphenHbVZeSuPtqRms/s1600/ssh4.JPG&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;br /&gt;
I grabbed the file, put it on my usb stick, and, soul still intact, walked it over to my secured server, and dropped the file in /var/tmp. I would try again, this time installing the openssh-client first, and see what happened. My soul was still at risk though- downgrading the existing package could prove to be as evil as I feared. (Cue music- &lt;a href=&quot;http://www.joeydevilla.com/2012/11/21/satan-laughing-spreads-his-wings-or-black-sabbath-transcribed-for-the-ukelele/&quot; target=&quot;_blank&quot;&gt;War Pigs, by&amp;nbsp;Black Sabbath- &quot;Larry Ellison laughing spreads his wings&quot;&lt;/a&gt;)&lt;br /&gt;
&lt;br /&gt;
I once again, called upon dpkg to do my bidding...&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$ cd /var/tmp&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$ ls &lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;openssh-client_5.3p1-3ubuntu3_i386.deb&amp;nbsp;&amp;nbsp; openssh-server_5.3p1-3ubuntu3_i386.deb&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$ sudo dpkg -i openssh-client_5.3p1-3ubuntu3_i386.deb&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;dpkg: warning: downgrading openssh-client from 1:5.5p1-4ubuntu4 to 1:5.2p1-3ubuntu3.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;(reading database ... 118283 files and directories currently installed.)&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Preparing to replace openssh client 1:5.5p1-4ubuntu4 (using openssh-client_5.3p1-3ubuntu3_i386.deb) ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Unpacking replacement openssh-client ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Installing new version of config file /etc/ssh/ssh_config ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Processing triggers for man-db ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Hmm, that wasn&#39;t evil at all. The dpkg command very politely informed me that it would replace the original openssh-client package with the version I specified- and it clearly did so with no problems.&lt;br /&gt;
&lt;br /&gt;
I forged ahead, and attempted to install the openssh-server package again:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$ sudo dpkg -i openssh-server_5.3p1-3ubuntu3_i386.deb&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;(Reading database ... 118282 files and directories currently installed.)&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Preparing to replace openssh-server 1:5.3p1-3ubuntu3 (using openssh-server_5.3p1-3ubuntu3_i386.deb) ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Unpacking replacement openssh-server ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Setting up openssh-server (1:5.3p1-3ubuntu3) ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Creating SSH2 RSA key; this ay take some time ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Creating SSH2 DSA key; this may take some time ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp;* Stopping OpenBSD Secure Shell server sshd&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;ssh start/running, process 2385&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Processing triggers for ureadahead ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Processing triggers for ufw ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Processing triggers for man-db ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;$&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New; font-size: x-small;&quot;&gt;&lt;/span&gt;&lt;br /&gt;
...and was successful. (Cue music- &lt;a href=&quot;http://www.dailymotion.com/video/x76e0x_freddy-mercury-we-are-the-champions_music&quot; target=&quot;_blank&quot;&gt;We are the Champions, by Queen- &quot;No time for losers, &#39;cause we are the Champions...&quot;&lt;/a&gt;)&lt;br /&gt;
&lt;br /&gt;
The dpkg command was even polite enough to let me know that it was replacing the failed install attempt (and did so cleanly). All was well in the datacenter, and I survived the crisis without selling my soul to the Dark Lord. Clearly, the ones who posted warnings that manual package installation was a difficult, tedious task have never had to sell their soul to install a Solaris recommended patch cluster! This seemed&amp;nbsp;&lt;a href=&quot;http://www.youtube.com/watch?v=X8PyTo6NyXA&quot; target=&quot;_blank&quot;&gt;all too easy&lt;/a&gt;. Perhaps there will be other&amp;nbsp;Ubuntu pitfalls&amp;nbsp;that will force me down the Dark Path, but not today.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/9135060893344652565/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/05/linux-ubuntu-package-installation.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/9135060893344652565'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/9135060893344652565'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/05/linux-ubuntu-package-installation.html' title='Linux (Ubuntu) package installation without apt-get'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgMBV5XDLiDT14EPsOo-52fUeDeuTMbON2rl6qsIN_r5BknjGopfMj5fh0AWraR5AGBxWrZJCotEXOy9Uo6U62sBCJ_CxnXX4SQJE1n9I0Y9koa8_ZtNqV15x9gm7O10o7uRKyxgDKOuus/s72-c/600px-Former_Ubuntu_logo_svg.png" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-5699550989661560683</id><published>2013-04-05T23:50:00.002-04:00</published><updated>2013-04-05T23:50:28.039-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Database"/><category scheme="http://www.blogger.com/atom/ns#" term="Kernel Tuning"/><category scheme="http://www.blogger.com/atom/ns#" term="Oracle"/><category scheme="http://www.blogger.com/atom/ns#" term="Performance Tuning"/><category scheme="http://www.blogger.com/atom/ns#" term="Shared Memory"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><title type='text'>Adjust Solaris 10 shared memory to prevent Oracle &quot;Out of Memory&quot; errors</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s1600/oralogo-small.gif&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;25&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s200/oralogo-small.gif&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
If you&#39;ve ever built, recovered, or administered a Solaris system hosting an Oracle database server, you&#39;ve probably seen this. In fact you&#39;ve probably seen this when your local DBA first attempted to install Oracle itself.&lt;br /&gt;
&lt;br /&gt;
While either performing your Oracle installation, or attempting to start a database, you may see the mysterious &quot;Out of Memory&quot; error:&lt;br /&gt;
&lt;br /&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;background-color: white; color: #555555; font-size: 12px; line-height: 18px; margin-bottom: 10px; margin-top: 10px;&quot;&gt;SQL&amp;gt; startup
ORA-27102: out of memory
SVR4 Error: 22: Invalid argument&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;background-color: white; margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;You can check your memory statistics as much as you like, you are not running out of physical memory- this error is generated because oracle is attempting to create a larger shared memory segment than is allowed. There are nearly no clues that this error is actually talking about shared memory, unless you look in the oracle alert log:&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;background-color: white; margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;pre class=&quot;programlisting&quot; style=&quot;color: #555555; font-size: 12px; line-height: 18px; margin-bottom: 10px; margin-top: 10px;&quot;&gt;% tail -2 alert.log
WARNING: EINVAL creating segment of size 0x000000028a006000
fix shm parameters in /etc/system or equivalent&lt;/pre&gt;
&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;background-color: white; margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;The recommended fix shown in the alert log does tell you to adjust shared memory settings, however, it is very misleading if you are running Solaris 10. Shared memory settings used to be controlled by kernel parameters in the /etc/system file in Solaris 9 and older, in Solaris 10, this is no longer the case. We actually set the shared memory settings on a per &quot;project&quot; basis in /etc/project - and tie those settings to specific user and/or group ids.&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;Before we go about altering our kernel settings, we need to figure out what values to use. The default values for Solaris 10 are supposed to be 1/4 of physical memory- assuming we are on a system running 32gb of physical memory, 1/4 of that is 8gb - if we look at our error from the alert log, we see a warning that we are creating a segment of size &quot;0x000000028a006000&quot; - this value is in hexadecimal- if we convert it to decimal, we get &quot;10905214976&quot; or 10gb&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;Now the error message makes sense- we have a default maximum shared memory segment of 8gb (25% of 32gb physical memory) - and oracle is trying to create what we now know to be a 10gb shared memory segment. This tells us that we need to move up our shared memory maximum to at least 10gb.&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;The other piece of information we need is the easy part- we need to know the user id that Oracle will be running as. Usually, by default the user id is &quot;oracle&quot; and the associated group id is &quot;dba&quot; - however in some implementations, the dba building the database server may opt to use different users and groups than the default. This is a simple matter of asking your dba what user ids apply to the oracle instance you are troubleshooting.&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;We have two options for implementing our fix, we can either modify an existing project entry in the /etc/project &amp;nbsp;file to add a maximum shared memory segment size for oracle, which could be problematic as we may accidentally impact other tuning settings built our for other workloads. A better solution is to create a new project id specifically for oracle. User definable project id numbers are generally 100 or higher. To find out what project id numbers are currently in use, you can run a simple awk filter against the contents of the /etc/project file:&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;# cat /etc/project | awk -F: &#39;{print $2}&#39;&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;This command will print out a list of all defined project IDs, to create a new project definition simply pick any project id number greater than, or equal to 100 that does not appear on the list generated above.&amp;nbsp;&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;Creating the project with a maximum shared memory segment of 10gb , and associating it with oracle is a simple task or running 2 commands using the information we&#39;ve collected so far.&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;First, create the project definition (I am using project ID number 100 in my example, please make sure the project ID you actually use is greater than 100 and is not already in use!):&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;&lt;span style=&quot;font-family: Courier New, Courier, monospace; font-size: x-small;&quot;&gt;# projadd -p 100 -c &#39;Oracle Shared Memory&#39; -U oracle -G dba -K &#39;project.max-shm-memory=(privileged,10G,deny)&#39; ORA&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;This creates a project named &quot;ORA&quot; using project ID number &quot;100&quot; associated with the user &quot;oracle, and the group &quot;dba&quot;. We define the shared memory maximum tunable parameter for this project to be &quot;10gb&quot;.&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;Once that we&#39;ve created this project, we need to modify the user definition for the oracle ID to be associated with our new project:&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;# usermod -K project=ORA oracle&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;This indicates that the user id &quot;oracle&quot; is now associated with the project &quot;ORA&quot; that we created earlier.&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;Once you&#39;ve done this, you should be able to start your Oracle database as normal. I typically do not trust Solaris 10 to always instantiate kernel changes to a running system in all cases, so if you have the luxury of being able to do so, I would also recommend rebooting the system. When your dba starts oracle again, watch the alert log in case you made a mistake and set you shared memory maximum too low. If you need to bump up your values after the fact, simply edit the value in /etc/project in the line &amp;nbsp;associated with the new project you created- keep in mind that the shared memory value (project.max-shm-memory) will be converted from gigabytes to bytes- so make sure you convert your numbers correctly if you need to make a change. The /etc/project entry will look like the lone below:&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;ORA:100:Oracle Shared Memory:oracle:dba:project.max-shm-memory=(privileged,&lt;/span&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;10905214976&lt;/span&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;,deny)&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;pre class=&quot;programlisting&quot; style=&quot;margin-bottom: 10px; margin-top: 10px;&quot;&gt;&lt;span style=&quot;font-family: Times New Roman;&quot;&gt;&lt;span style=&quot;white-space: normal;&quot;&gt;
&lt;/span&gt;&lt;/span&gt;&lt;/pre&gt;
&lt;/pre&gt;
</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/5699550989661560683/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/04/adjust-solaris-10-shared-memory-to.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5699550989661560683'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5699550989661560683'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/04/adjust-solaris-10-shared-memory-to.html' title='Adjust Solaris 10 shared memory to prevent Oracle &quot;Out of Memory&quot; errors'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s72-c/oralogo-small.gif" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-3340216833697775030</id><published>2013-03-06T22:41:00.000-05:00</published><updated>2013-03-06T22:44:04.196-05:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Jumpstart"/><category scheme="http://www.blogger.com/atom/ns#" term="Miniroot"/><category scheme="http://www.blogger.com/atom/ns#" term="patches"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><title type='text'>Patching a Solaris 10 Jumpstart Miniroot</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s1600/oralogo-small.gif&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;26&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s200/oralogo-small.gif&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
I recently ran into an issue that required an update to our Solaris 10 jumpstart miniroot. Apparently we hit upon a bug where&amp;nbsp;attempting to install a flash archive of a zfs root system that was patched up to kernel patch 147440-19 or higher would produce an unusual error if you attempted to install it using a miniroot that was below patch level 147440-19. This meant that our current jumpstart environment (running a base, install of Solaris 10 update 10) could not install the flash image. This also meant that the officially released DVD media could not install the flash archive either.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Oracle chose to fix this issue with a patch, but never went back and offered to release updated DVD media- so basically, every existing boot DVD out there is now broken. This is far from an optimal solution, but it&#39;s the solution we have to work with.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
The long term solution is going to involve building a bootable DVD that includes the kernel patch, but there was an immediate&amp;nbsp;need to get the install done, so we opted for the temporary fix (which we would need to do anyway). That fix wqas to add the required patch to the network bootable miniroot on our jumpstart environment. I&#39;ve done this in the past with older versions of Solaris, but never with Solaris 10, and not recently, so I needed to look up the process, because something didn&#39;t seem to look familiar to me..&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
In the past, the miniroot was simply laid out in the network install directories, just the way you would see it on a live root file system- you could go in and alter startup scripts, and make a few tweaks here and there at will- and patching the miniroot was aas simple as giving the patchadd command the &quot;-C&quot; option and pointing it at the network install image.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Under Solaris 9, we could find this directory under the install directory under Solaris_9/Tools/Boot- as shown below:&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;#&amp;nbsp;ls /jumpstart/9/Solaris_9/Tools/Boot&lt;/span&gt;&lt;/div&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;a&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; devices&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; mnt&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; proc&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; usr&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;bin&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; etc&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; netmask&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; reconfigure&amp;nbsp; var&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;cdrom&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; kernel&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; opt&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; sbin&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; webstart&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;dev&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; lib&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; platform&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; tmp&lt;/span&gt;&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
This hierarchy basically looks exactly like a standard root file system- and when your target system boots from the jumpstart server, this will basically be nfs mounted as if it were the root file system during a jumpstart installation.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
If we look in the corresponding directory for a Solaris 10 install image, we see a number of bz2 archives rather than :&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;# ls /jumpstart/10/Solaris_10/Tools/Boot&lt;/span&gt;&lt;/div&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;X.cpio.bz2&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; netmask&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;X_small.cpio.bz2&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; pkg_db.cpio.bz2&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;lu.cpio.bz2&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; platform&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;lu.platforms&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; usr_sbin_install_d.cpio.bz2&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
If you try applying a kernel patch to the miniroot using the old simple, one command method- it will fail. The root you need to patch just isn&#39;t laid out the way it needs to be in order for patchadd to work.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
To get around this, you need to dig around Oracle&#39;s official documents- you&#39;ll find a few how-to docs on their web site, that are actually more complicated than they need to be- those instructions will show you how to unpack the miniroot so it can be updated, and how to repack it, but also has a number of odd-seeming steps that include moving aside and restoring the original rc2 script and sudoers file, and performing a few svccfg commands to update the service management facility within the miniroot. These steps may be important if you expect to use your network install image as a entwork boot image- that may have had these files and settings altered, but for a miniroot supporting jumpstart- all these files should always be at their original, out-of-the-box settings, so I would advise skipping them unless you have a need.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
My trial and error showed that tinkering around with the rc scripts , sudoers file and svccfg settings in the miniroot, can actually break your ability to jumpstart. Leaving those steps out produced a working jumpstart miniroot that was able to install the problem&amp;nbsp; flash archive, and could still install flash archives that are below the patch level that introduced the problem in the first place.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Here are the steps I followed-&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
First- obtain the appropriate patch from roacle.com. A little research there shwoed that I needed patch 147440-19 or higher to solve my issue, and 147440-19 had been superseded by later patches- the most recent version I could find at the time was 147440-27, so that&#39;s the one I used. The patch comes from Oracle as a zip file, so I simply staged it in my root directory and unzipped it.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Next, you&#39;ll need to unpack the miniroot you intend to patch. I&#39;d strongly reccommend making a backup of your jumpstart server now, before you start, or if you, like me, have built your jumpstart environment on a zfs file system, I&#39;d advise taking a snapshot of the file system now so you can roll back to it if you make a mistake that you can&#39;t figure out how to undo. It happens to the best of us- so swallow your pride and make the backup!&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
You&#39;ll need to make a directory to stage your unpacked miniroot. In my example, I&#39;ll use /jumpstart/10_temp as a temporary staging area:&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;# mkdir /jumpstart/10_tmp&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Then you&#39;ll need to unpack the existing miniroot. The command to perfrom the unpack operation is buried in an odd place, and you probably won&#39;t be able to find it on your own without looking up the instructions on Oracle.com.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace; font-size: x-small;&quot;&gt;# /boot/solaris/bin/root_archive unpackmedia /jumpstart/10 /jumpstart/10_temp&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Note that /jumpstart/10 is where my original Solaris 10 Update 10 jumpstart install image lives, and&amp;nbsp;/jumpstart/10_temp is the directory I created to&amp;nbsp;temporarily hold the miniroot.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;If you look in your staging directory, you will now see what looks like a typical root file system exploded out.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;# ls /jumpstart/10_temp&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;a&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; dev&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; lib&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; proc&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; tmp&lt;/span&gt;&lt;/div&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;bin&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; devices&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; mnt&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; reconfigure&amp;nbsp; usr&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;boot&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; etc&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; opt&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; sbin&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; var&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;cdrom&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; kernel&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; platform&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; system&lt;/span&gt;&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Now you can go ahead and use your patchadd -C to add the desired kernel patch to your miniroot.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;# patchadd -C /jumpstart/10_temp /147440-27&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
The first argument is your temporary staging directory, the second is the path you unzipped your patch to. Most admins prefer to stage their patched somewhere other than / - I did this simply for brevity.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Once the patchadd completes, you will need to repack the modified miniroot. This command will convert the exploded miniroot in our staging directory, and build all the bz2 archives needed, and put them back in our original network boot image location.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace; font-size: x-small;&quot;&gt;# /boot/solaris/bin/root_archive packmedia /jumpstart/10_temp /jumpstart_10&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Note that we put the directory of the staging directory first, then the original install image location second. Oddly enough, the original &lt;a href=&quot;http://www.oracle.com/technetwork/systems/articles/patchmini-jsp-137348.html&quot; target=&quot;_blank&quot;&gt;oracle page&lt;/a&gt;&amp;nbsp; where I found this command has a really bad typo- it lists the cmmand path as starting with &quot;/Boot/boot&quot; - just use the command as I wrote it instead. You&#39;ll aslo find all the extraneous steps I mentioned on the same page. You&#39;ll also note that this procedure changes depending on what release of Solaris 10 you are using. The 10/08 release marks the point where the miniroot first needed to be unpacked/re-packed. Earlier versions use the same method as you would for Solaris 9- simply do a patchadd -C pointed at your &quot;Tools/Boot&quot; directory- i.e.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;# patchadd -C /jumpstart/10/Solaris_10/Tools/Boot /147440-27&lt;/span&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
and avoid the pack and unpack commands.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj7yWWxzfFreZTkj-n7ehp8ldBIf16FpxOBnjjizeHs2nxMXIiOipBQ3MVEwdrbkqy79W-wuTccmeiKT1-oaesXy3MGaDZY_Vs1mjFbraRi0pgXnbDxdH50KSnCVUjFuNQ-tfshRwYq-3A/s1600/sun-logo.png&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;198&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj7yWWxzfFreZTkj-n7ehp8ldBIf16FpxOBnjjizeHs2nxMXIiOipBQ3MVEwdrbkqy79W-wuTccmeiKT1-oaesXy3MGaDZY_Vs1mjFbraRi0pgXnbDxdH50KSnCVUjFuNQ-tfshRwYq-3A/s200/sun-logo.png&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&lt;br /&gt;&lt;/div&gt;
</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/3340216833697775030/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2013/03/patching-solaris-10-jumpstart-miniroot.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/3340216833697775030'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/3340216833697775030'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2013/03/patching-solaris-10-jumpstart-miniroot.html' title='Patching a Solaris 10 Jumpstart Miniroot'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s72-c/oralogo-small.gif" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-4294318174164892799</id><published>2012-12-23T11:42:00.002-05:00</published><updated>2013-08-21T11:58:30.945-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Disaster Recovery"/><category scheme="http://www.blogger.com/atom/ns#" term="File System"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 11"/><category scheme="http://www.blogger.com/atom/ns#" term="Sun"/><category scheme="http://www.blogger.com/atom/ns#" term="ZFS"/><category scheme="http://www.blogger.com/atom/ns#" term="zpool"/><title type='text'>Shortcut for recreating zfs file systems</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s1600/zfs_feature_2.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;160&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s200/zfs_feature_2.jpg&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
One of the more tedious tasks in recovering a Solaris system manually is rebuilding the original file system structure. We have a native tool for imaging the operating system itself with flash archives (Solaris 10 and lower) or a similar but not quite as simple method involving &lt;a href=&quot;http://var-adm.blogspot.com/#!/2012/07/clonedr-solaris-10-system-using-zfs.html&quot; target=&quot;_blank&quot;&gt;zfs snapshots&lt;/a&gt; for zfs root system (Solaris 10 and Solaris 11). After imaging the operating system, you would typically need to recreate the original file system structure before beign able to restore application data.&lt;br /&gt;
&lt;br /&gt;
This task is complicated due to the feature limitations of legacy ufs file systems, and the inherent complexities involved in building volumes with Veritas Volume Manager or the semi-native Solaris Volume Manager- neither of which were designed with rapid recovery in mind. ZFS file systems, however, give us other options.&lt;br /&gt;
&lt;br /&gt;
IBM&#39;s AIX operating system has had the advantage in this area for a long time. Since AIX has had a robust native logical volume manager built into the operating system, a rapid method for recovery was a simple addition. AIX admins can use the &quot;savevg&quot; command to basically take a snapshot of a system&#39;s file system structure (just the structure- file system sizes, mount points, etc... not the actual data). If the ouput is saved in such a way that it is accessible at time-of-recovery, it can be used to recreate the entire file system structure in one fell swoop using the &quot;restvg&quot; command.&lt;br /&gt;
&lt;br /&gt;
The method available to systems using zfs file systems is very similar to this idea. On any Solaris system running zfs file systems, the &quot;zpool history&quot; command may be used to dump a history of all zpool and zfs commands run against each zpool on the system, in chronological order. With a minimal amount of adjustment, this output can be used to basically rebuild the entire history of every zpool on the system.&lt;br /&gt;
&lt;br /&gt;
See the example below:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;# zpool history&lt;br /&gt;History for &#39;shared&#39;:&lt;br /&gt;2010-04-09.15:06:28 zpool create -f shared c0t1d0s0 c0t2d0s0 c0t3d0s0&lt;br /&gt;2012-02-12.20:38:02 zfs snapshot &lt;/span&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;shared@02122012&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;2012-03-12.20:38:02 zfs create mountpoint=/export/temp shared/temp&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;2012-11-05.11:04:40 zpool scrub -s shared&lt;br /&gt;2012-11-05.11:21:07 zfs set mountpoint=/a shared&lt;br /&gt;2012-11-05.11:22:03 zfs set mountpoint=/shared shared&lt;br /&gt;2012-11-05.12:38:36 zfs snapshot &lt;/span&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;shared@11512&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New; font-size: x-small;&quot;&gt;&lt;/span&gt;&lt;br /&gt;
In our sample output, we can see that the system has one zpool names &quot;shared&quot; - this pool was created in April 2010 using three disks- c0t1d0, c0t2d0 and c0t3d0. Next, in february of 2012, someone created a zfs snapshot of the system, then created a new dataset called &quot;shared/temp&quot; mounted on &quot;/export/temp&quot;. Later in 2012, someone performed a manual zfs scrub on the pool, then changed a mount point and subsequently changed it back. Finally another zfs snapshot was taken.&lt;br /&gt;
&lt;br /&gt;
Note that this shows how many disks compose each pool, and what their device names are- but it does not show the capacity of those disks. We can also see from the zpool create command, that this is a simple striped zpool, with no mirroring or raidz protection. We have most of, but not all of, the information we would need to receate this pool in a disaster recovery scenario. &lt;br /&gt;
&lt;br /&gt;
To truly recover this system we would also need to know the size of the zpool. The output of a &quot;df -h&quot; would help, but with zfs file systems- especially once you consider that you may have a mix of file systems with and without quotas- the df output can be somewhat misleading. A better way to capture total zpool capacity is with the output of the &quot;zpool list&quot; command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zpool list&lt;br /&gt;NAME&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; SIZE&amp;nbsp;&amp;nbsp; USED&amp;nbsp; AVAIL&amp;nbsp;&amp;nbsp;&amp;nbsp; CAP&amp;nbsp; HEALTH&amp;nbsp; ALTROOT&lt;br /&gt;shared&amp;nbsp;&amp;nbsp; 204G&amp;nbsp; 13.0G&amp;nbsp;&amp;nbsp; 191G&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 6%&amp;nbsp;&amp;nbsp;ONLINE&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
I would also reccomend saving this output to use as reference. Here we can see that the zpool named &quot;shared&quot; is 204 GB in size- we can also see that only 13 GB are currently in use. This tells us that we really need 204 GB to recreate the pool exactly as-is, but we may be able to get away with as little as 13 GB if we do not have enough disk to support the full allocation.&lt;br /&gt;
&lt;br /&gt;
So how do we use this information for a recovery- first we need to save it somewhere- if you are using a flash archive or similar method, you can simply dump the output of these commands into a directory normally captured in the falsh archive, such as /var - &lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zpool history &amp;gt; /var/zpool.history&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zpool list &amp;gt; /var/zpool.list&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
If you are not imaging your operating system, you will need to put these text files somewhere where you can get them at time-of-recovery, such as on a shared disk that gets replicated to your recovery site, or on a tape/cd/usb drive that is sent off-site as part of your &quot;crash kit&quot; on a regualr basis.&lt;br /&gt;
&lt;br /&gt;
Either way, let&#39;s assume that you&#39;ve already recovered your operating system, and placed these two files back in the /var directory for reference. You cna refer to the zpool.list file to determine how much additional disk space you need to attach to your system to support your file systems.&lt;br /&gt;
&lt;br /&gt;
Keep in mind that if you a recovering at a vendor site- there is no guaruntee that you will have the same exact hardware you have in production. You cannot assume that you will have the same number and size disks in recovery that you have in production, so simply replaying every command in the history file as-is may not produce the results you want. Your best bet is to manually recreate your zpools using the &quot;zpool create&quot; command with enough disks to cover the size(s) listed in the &quot;zpool list&quot; output. This is not really a hardship, since it only takes a single command per pool. &lt;br /&gt;
&lt;br /&gt;
In our simple example, this is actually the bulk of the work, however, remember that in most production implementations, you may have a large number of file systems under your main pools, with any number of non-default settings, such as mountpoints, quotas, and reservations. There may be other non-obvious settings created using &quot;zfs set&quot; commands that may need to be replayed as well- these more teduious steps are where this method saves a large amount of time.&lt;br /&gt;
&lt;br /&gt;
First, we will need to take our zpool history file, and strip out the parts we do not need- headers, and the time stamps:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;# egrep &quot;zfs|zpool&quot; /var/zpool.history | cut -d &quot; &quot; -f2-&amp;nbsp;&amp;gt; /tmp/zprecov&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
The &quot;zprecov&quot; file will contain a list of zpool and zfs commands, without the extra recordkeeping items:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;# cat /tmp/zprecov&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;zpool create -f shared c0t1d0s0 c0t2d0s0 c0t3d0s0&lt;br /&gt;zfs snapshot &lt;/span&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;shared@02122012&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;zfs create mountpoint=/export/temp shared/temp&lt;br /&gt;zpool scrub -s shared&lt;br /&gt;zpool clear shared&lt;br /&gt;zfs set mountpoint=/a shared&lt;br /&gt;zfs set mountpoint=/shared shared&lt;br /&gt;zfs snapshot &lt;/span&gt;&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;shared@11512&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
We can now edit this file and remove any steps we do not need. Since it is better practice to crate the top level zpool manually, we can remove the &quot;zpool create&quot; step, once we have done so. Items such as the &quot;zpool scrub&quot; and the &quot;zpool clear&quot; are administrative functions- a consistency check and acommand to clear wrnings from the zpool status- these will not cause any harm if we run them, but they are also not going to perform any tasks needed to recreate file system strcuture and settings, so we can remove them if we wish. The snapshot commands are also going to be extraneous- since we are building an empty zpool, we will basically be taking empty snapshots. We will, however, need the zfs create, and zfs set entries- in real-life cases, the set commands may capture things such as changing mountpoints, and changing disk quotas, sharing settings, etc... - so we will want to capture this. After out recommended edits, our file should look like:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;# cat /tmp/zprecov&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;zfs create mountpoint=/export/temp shared/temp&lt;br /&gt;zfs set mountpoint=/a shared&lt;br /&gt;zfs set mountpoint=/shared shared&lt;/span&gt;&lt;br /&gt;
Now we can give this file execution permissions, and if we have already created a pool called &quot;shared&quot; with the right amount of disk space, we can execute this file to re-create our pool:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# chmod +x /tmp/zprecov&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /tmp/zprecov&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
We can check our work with a &quot;zfs list&quot; command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zfs list&lt;br /&gt;NAME&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; USED&amp;nbsp; AVAIL&amp;nbsp; REFER&amp;nbsp; MOUNTPOINT&lt;br /&gt;shared&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0G&amp;nbsp;&amp;nbsp; 204G&amp;nbsp; 0G&amp;nbsp; /shared&lt;br /&gt;shared/temp&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0G&amp;nbsp;&amp;nbsp; 204G&amp;nbsp;&amp;nbsp;0G&amp;nbsp; /export/temp&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
The used values are &quot;0&quot; since we have not populated the file system with data yet, and the &quot;available&quot; column shows the total size of the pool - 204G since we have not specified any disk quotas.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/4294318174164892799/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/12/shortcut-for-recreating-zfs-file-systems.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4294318174164892799'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4294318174164892799'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/12/shortcut-for-recreating-zfs-file-systems.html' title='Shortcut for recreating zfs file systems'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s72-c/zfs_feature_2.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-847980104020344151</id><published>2012-11-22T09:07:00.001-05:00</published><updated>2012-11-22T09:07:34.818-05:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Backup"/><category scheme="http://www.blogger.com/atom/ns#" term="Disaster Recovery"/><category scheme="http://www.blogger.com/atom/ns#" term="Netbackup"/><category scheme="http://www.blogger.com/atom/ns#" term="Restore"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><category scheme="http://www.blogger.com/atom/ns#" term="Symantec"/><category scheme="http://www.blogger.com/atom/ns#" term="Tapeless Backup"/><title type='text'>Set up Backups to Disk for NetBackup 7.x</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEincfvdx1Q5bQNgT6KRohUFABETA1UQlVIDqGffZ1NXfEpt6iJB9t_Rjw0Qrnc9ECuN4gh58ytqN7L4XdM_xnGdwGO81jbfP-rstkmP0NTHs474M5mwpeueH21XAOdj9J0EXBe73_3cdTo/s1600/NetBackup_GUI_jnbSA.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;150&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEincfvdx1Q5bQNgT6KRohUFABETA1UQlVIDqGffZ1NXfEpt6iJB9t_Rjw0Qrnc9ECuN4gh58ytqN7L4XdM_xnGdwGO81jbfP-rstkmP0NTHs474M5mwpeueH21XAOdj9J0EXBe73_3cdTo/s200/NetBackup_GUI_jnbSA.jpg&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&quot;Tape is dead&quot; has been the mantra for disaster recovery planning for many years now. Somehow, tape still lingers on, but these days, every backup administrator or disaster recovery engineer should be aware that there are alternatives that work. The most basic alternative available&amp;nbsp;on most enterprise level backup software is disk-based backups. When combined with disk replication strategies (SRDF, NetApp SnapMirror, DataDomain, or even home grown jury-rigged solutions using tools such as zfs send/receive)- disk based backup solutions are a powerful tool for reducing RTO times, and ensuring against data loss.&lt;br /&gt;
&lt;br /&gt;
Symantec&#39;s NetBackup is one such product that has embraced disk based backups. Coming from a tape-based background, I always assumed that configuring and manageing disk based backups would somehow be more difficult than tape- once I started using disk in place of tape, I learned that I was very wrong. Disk backups are simple to set up in NetBackup, surprisingly simple. Below I will walk through the process of setting up a disk based storage pool, and building a backup policy to use it. &lt;br /&gt;
&lt;br /&gt;
I am starting with the assumption that you already have NetBackup (version 7.0 or better) installed, and that you have built a file system on the disks you would like to put your backups on. We will use the &quot;Basic Disk&quot; option here, so as long as your disk is available to your master server, and is mounted as whatever type of file system you prefer, you are ready to go. I am personally partial to zfs file systems, but you may use UFS, vxfs, ext3, or whatever volume management/file system schema you prefer- that part does not matter.&lt;br /&gt;
&lt;br /&gt;
Before you attempt to do any of the usual device configuration steps you would do for a tape-based NetBackup solution, stop. You do not need to muck about with pass-through drivers, or wait for the device configuration wizard to attempt discovery. All you need is an empty directory to store your backups on, and the NetBackup UI. The rest is simply a matter of pointing and clicking.&lt;br /&gt;
&lt;br /&gt;
First we will need to configure a new storage unit- in the NetBackup UI, under, NetBackup Management, expand&amp;nbsp; the &quot;Storage&quot; item, and right click on &quot;Storage Units&quot; and select &quot;New Storage Unit&quot; from the dropdown.&lt;br /&gt;
&lt;br /&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjfHrfWCurKZM0UuVgTnE3GTEYWJRLE62FC6WBMNz7oHD2b9RM9Y0MGntNUY7DUjuvOQ0J0t3LOr0o8VZVuY3Jd3SSF-o3LPzj3dR9si9FOQtuXuA9ysHfrGnAaOunV7k8jVekgkAa5Hus/s1600/disk_backup1.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;228&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjfHrfWCurKZM0UuVgTnE3GTEYWJRLE62FC6WBMNz7oHD2b9RM9Y0MGntNUY7DUjuvOQ0J0t3LOr0o8VZVuY3Jd3SSF-o3LPzj3dR9si9FOQtuXuA9ysHfrGnAaOunV7k8jVekgkAa5Hus/s320/disk_backup1.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
The next screen is where you perform the bulk of your work. Here you will give your storage unit a name. Make sure the &quot;Storage unit type&quot; selection reads &quot;Disk&quot;, and that the &quot;Disk Type&quot; is &quot;BasicDisk.&quot; Further down you will see a text box named &quot;Absolute pathname to directory&quot; - here is where you will enter the directory where you intend your backups to live. Note that you MUST use the absolute path to this directory, not a relative path, and not a path that takes you through a symbolic link. To make sure, you should cd into the directory and run the pwd command- if the result is different than what you expect, then you crossed a symbolic link somewhere.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Most of the other settings here can be left at defaults. What they do and how they effect performance are outside the scope of this guide. You should, however, take not of the &quot;Maximum concurrent jobs&quot; setting. This is set to 1 by default- meaning that only one active job can use this storage unit at any given time. In most production shops, you&#39;ll need the ability to do more than one thing at a time, so I strongly suggest bumping this number up to something higher if you intend to do anything more than running a few test backups and/or restores.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhWzcJ8Qm4nN2SjSx4Khf9kcKAFCsfG3gypNpkIBBWFh9n8z43zbT6jUm0QU0SNBKOq2kJ-iWs3eOWcA9NskansZkXaWciF2xgc2wmDwhlOLb-hJr8WlyXJG2-ql-uotfVBLYKdQU45AvU/s1600/disk_backup2.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhWzcJ8Qm4nN2SjSx4Khf9kcKAFCsfG3gypNpkIBBWFh9n8z43zbT6jUm0QU0SNBKOq2kJ-iWs3eOWcA9NskansZkXaWciF2xgc2wmDwhlOLb-hJr8WlyXJG2-ql-uotfVBLYKdQU45AvU/s320/disk_backup2.jpg&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Click &quot;OK&quot; when done. There, you&#39;re finished. &lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Really that&#39;s it. You can now go ahead and write backups directly to disk.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
OK, so you still have to configure a backup policy that uses this storage unit, or change an existing one to point to it, but that&#39;s fairly easy if you&#39;ve done it before.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
To set up a new backup policy, go to the NetBackup UI, and select &quot;Policies,&quot; then right click on the name of your master server, and select &quot;New Policy&quot; from the drop-down.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhRKZcMhQ3qWfzXN3OpwK16jBO9mVMYs2mJM_EcBoKvBtD3Vf7nBcf1S2X0Hho3wnpFflf_Bv8g03G_gyhoimYD4WtgbZzh4wyu5q0kGsA3Wx9FCqpOTMJSxIy2NJeNIAYnqlnbf2f7zq0/s1600/disk_backup3.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;266&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhRKZcMhQ3qWfzXN3OpwK16jBO9mVMYs2mJM_EcBoKvBtD3Vf7nBcf1S2X0Hho3wnpFflf_Bv8g03G_gyhoimYD4WtgbZzh4wyu5q0kGsA3Wx9FCqpOTMJSxIy2NJeNIAYnqlnbf2f7zq0/s320/disk_backup3.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;Enter the name of the policy you want to create, and if you&#39;re new to policy creation, check off &quot;use backup policy creation wizard&quot; and let NetBackup walk you through the process:&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhrO8Rdt1UiU4O2tM_H1jUh219z2oWaQKPb9ukCnhMy3TyC5mL8C-OlRz1HZT-b22z_N8oHGpsoKINqq0c8BV96KR0IuJJl8GKn5jYcvHeNKfKl8ysxAmfDJllixS8H8bp5iwfdiETORs8/s1600/disk_backup4.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhrO8Rdt1UiU4O2tM_H1jUh219z2oWaQKPb9ukCnhMy3TyC5mL8C-OlRz1HZT-b22z_N8oHGpsoKINqq0c8BV96KR0IuJJl8GKn5jYcvHeNKfKl8ysxAmfDJllixS8H8bp5iwfdiETORs8/s1600/disk_backup4.jpg&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
The wizard will give you a screen of explanations, but if you don&#39;t bother to read it, I don&#39;t blame you...&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjL0ORC-3JA7IaqEq4maJs5UEWLQugY53iaCAla1A7eqEzg_05LTwHVTwOj9XY-LrPcJil9Ih9D9ZrnVUFYbiStqyGLRooMRGQtblf-sPLaZ6LJ65pFN_VPbfRpK2t78KO3Y0kNXfulWcE/s1600/disk_backup5.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjL0ORC-3JA7IaqEq4maJs5UEWLQugY53iaCAla1A7eqEzg_05LTwHVTwOj9XY-LrPcJil9Ih9D9ZrnVUFYbiStqyGLRooMRGQtblf-sPLaZ6LJ65pFN_VPbfRpK2t78KO3Y0kNXfulWcE/s320/disk_backup5.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
The next screen will let you select the policy type (most basic file-level backups use &quot;standard&quot;)&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEijYl-oXgJrI3SfGg95XwuDDboVJddJzr6MEZWYq2t6t56g_-cw2e-ScJGfYTi6sZArFLwVGfi2owZUo3l1wmFHpENFxQdA08GqZ2edjIpFIijE4jglKqhsPuhziCZr0vu7FYuJjv8vq-s/s1600/disk_backup6.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEijYl-oXgJrI3SfGg95XwuDDboVJddJzr6MEZWYq2t6t56g_-cw2e-ScJGfYTi6sZArFLwVGfi2owZUo3l1wmFHpENFxQdA08GqZ2edjIpFIijE4jglKqhsPuhziCZr0vu7FYuJjv8vq-s/s320/disk_backup6.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
On the next screen, we will define which clients we are going to back up. Click &quot;add..&quot; then enter the name and operating system of the client. You can keep clicking &quot;add&quot; to add as many clients as you like, then hit close when done.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhH7ZpTmkYfhrn64yfAIJXznczKxBGzAsIu64ZufLaJT_uO8fYmWlEWxucqli333zCfpQ7PqimC3pMP8-fXx-xW6XvuaEvIf3TAG0DjgOoeV7eM9D1bSh3Mqnp8bnKY_df2ofU1NbuFyi4/s1600/disk_backup7.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhH7ZpTmkYfhrn64yfAIJXznczKxBGzAsIu64ZufLaJT_uO8fYmWlEWxucqli333zCfpQ7PqimC3pMP8-fXx-xW6XvuaEvIf3TAG0DjgOoeV7eM9D1bSh3Mqnp8bnKY_df2ofU1NbuFyi4/s320/disk_backup7.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
You can confirm your list of clients before hitting &quot;next&quot; and moving on...&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjPF9VgG2qLIXHKF_oLaNcFQbK4QIn0MBltJRr3Ezi1r75dr5S2qsDLifO-1ErTbYIP-mi0-cb8aTSD3ieAGSrXSnhQCFgeprTAVKWUpxUiomOkuDRVLaOfDMJGnKnTtYYi0R-v-5mec6c/s1600/disk_backup8.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjPF9VgG2qLIXHKF_oLaNcFQbK4QIn0MBltJRr3Ezi1r75dr5S2qsDLifO-1ErTbYIP-mi0-cb8aTSD3ieAGSrXSnhQCFgeprTAVKWUpxUiomOkuDRVLaOfDMJGnKnTtYYi0R-v-5mec6c/s320/disk_backup8.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
On the next screen, you may enter the directories you want this policy to back up. If you want to back up the entire system, click on &quot;back up all local drives&quot;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhFaY-1Rjt8m4H9YM_7QbXsbv6P1B8Yuu5j8EtwuxrM31aRIh7ZYDiSqLf5-_49yYUQjDAfC4mUFAbVIewBj4hVD6rz7jjYcUW7nqsOTUCfYQ5N0hl4inbg-QR6lPrxXe879tx6GaHh00U/s1600/disk_backup9.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhFaY-1Rjt8m4H9YM_7QbXsbv6P1B8Yuu5j8EtwuxrM31aRIh7ZYDiSqLf5-_49yYUQjDAfC4mUFAbVIewBj4hVD6rz7jjYcUW7nqsOTUCfYQ5N0hl4inbg-QR6lPrxXe879tx6GaHh00U/s320/disk_backup9.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Once you are happy with your backup selection, click next...&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhCMcwglhvF0mT6xAQiB-krTAhpl-sl4oQK7giOM7BNV8-HSykRmIpTVTAUkB1AuQ8eDnqYpf2FM3GN9d3gcw5m9DV51L5bJrFh5t2bwzx4iESyJpqg6K7h9MOJnlGyLmFTKwCQ2ZG2Qoo/s1600/disk_backup10.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhCMcwglhvF0mT6xAQiB-krTAhpl-sl4oQK7giOM7BNV8-HSykRmIpTVTAUkB1AuQ8eDnqYpf2FM3GN9d3gcw5m9DV51L5bJrFh5t2bwzx4iESyJpqg6K7h9MOJnlGyLmFTKwCQ2ZG2Qoo/s320/disk_backup10.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Here you can select the types of backups to perform- full backups, incremental backups, and/or user backups. I usually click on the &quot;user backup&quot; selection, so you have the option of manually starting a backup whenever you need to. When done, click next.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgad0QwFuxSZKEVO30L3HFbK_9FWcrnmEdTgGJrlRncJFJYYYc-VmA_UyCmN2lHqVLtpiFDe71L0-LSALn72JlGTiYZL8T8f76Wf6RqWWqsoWqL2EPspbyFbsZDeDzyykyIqBw-XWoXlkU/s1600/disk_backup11.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgad0QwFuxSZKEVO30L3HFbK_9FWcrnmEdTgGJrlRncJFJYYYc-VmA_UyCmN2lHqVLtpiFDe71L0-LSALn72JlGTiYZL8T8f76Wf6RqWWqsoWqL2EPspbyFbsZDeDzyykyIqBw-XWoXlkU/s320/disk_backup11.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Here we set our backup schedule- how often do we want regularly scheduled backups to run.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiP0klYGmT57y0urDO_FWY0jTbdLU2RP4I0XTZLJIEiY0FBCoB0ii45U7ivDrMHFl_PmAI0-RB0llSSdCUp2B6VYFsM4eXWRcMXsTJNVr7v5Qacyj724RiSvxh2VLslXwTTNuWnsIo0oAk/s1600/disk_backup12.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiP0klYGmT57y0urDO_FWY0jTbdLU2RP4I0XTZLJIEiY0FBCoB0ii45U7ivDrMHFl_PmAI0-RB0llSSdCUp2B6VYFsM4eXWRcMXsTJNVr7v5Qacyj724RiSvxh2VLslXwTTNuWnsIo0oAk/s320/disk_backup12.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Next pick the approximate windows of time when you want your backups to run. This window is not very granular- you can only choose &quot;business hours,&quot; &quot;off hours,&quot; or &quot;all day&quot; - for more granularity, you need to change the backup schedule manually- this wizard only gives you the most commonly used settings (which are usually good enough for most purposes).&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj1ir7Fwwd88dMLPeKng0qdE_2TgKnDPj2pv0gRd1acFBw4YVl05dkpTo23uFJn0GljBp34uIj0jWzfn3XqalPwSUU9hfmFQ3s2x7_f_HOUqSclfbsZ-BQoHhvFR2IllBd9XhW-uKkYnxM/s1600/disk_backup13.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;197&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj1ir7Fwwd88dMLPeKng0qdE_2TgKnDPj2pv0gRd1acFBw4YVl05dkpTo23uFJn0GljBp34uIj0jWzfn3XqalPwSUU9hfmFQ3s2x7_f_HOUqSclfbsZ-BQoHhvFR2IllBd9XhW-uKkYnxM/s320/disk_backup13.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Now you get to see a reminder that this wizard can only create new policies,and if you want to make a chaneg you have to modify the policy... blah blah blah...&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Speaking of modifying existing policies, you may have noticed that the wizard never gave you an opportunity to pick what storage pool to use for storing backups! To do that, we need to go back to our policies list, select our policy and change it&#39;s attributes as shown below.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiE1vSBp2yyTDOKOuompTS3LXzUbJCcUqkYgajKTshBGZ4w4gqVvio-cKrH0dTeNMOoFy2bvSEMYZUhoF6cFEaF8qY7W1H6g0QkmATbv7ImivSPlFsFzqwLCMBCX8X97z6AHEwF0txSmMw/s1600/disk_backup14.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;216&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiE1vSBp2yyTDOKOuompTS3LXzUbJCcUqkYgajKTshBGZ4w4gqVvio-cKrH0dTeNMOoFy2bvSEMYZUhoF6cFEaF8qY7W1H6g0QkmATbv7ImivSPlFsFzqwLCMBCX8X97z6AHEwF0txSmMw/s320/disk_backup14.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
We want to make sure our policy is pointing at the storage unit we created earlier, not somewhere else- I believe it defaults to something like &quot;any available&quot; which generally means &quot;whichever storage unit comes first when listed alphabetically&quot; - not necessarily what you&#39;re looking for.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg9bjVimicW0glDWM6-h00tu4frBCJys__F8YSKYpVvkl_nyXbLpft6pFEmzNnXVPcbo_M0UoYKfeatqof19F8ogh3fG7C-MhKdrOfMDMW2HR0TA0WsSbn12tG3wZUyGzGYK7QudhmDgGw/s1600/disk_backup15.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;313&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg9bjVimicW0glDWM6-h00tu4frBCJys__F8YSKYpVvkl_nyXbLpft6pFEmzNnXVPcbo_M0UoYKfeatqof19F8ogh3fG7C-MhKdrOfMDMW2HR0TA0WsSbn12tG3wZUyGzGYK7QudhmDgGw/s320/disk_backup15.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
So take a look at the &quot;Policy Storage&quot; selector in the upper left side of the window- if you have more than one storage unit defined, it may very well not be pointing to the one you want- change this selection to the storage unit you defined earlier, and click &quot;OK&quot;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Now we can either wait for the job scheduler to start running backups on it&#39;s own, or we can co back to our policy definition and kick off a backup manually.&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhxxg2Wx67rdpTCQsOzYPDPM7OE4o1swsMyFFHN-5KcR_ODX3I25i0QkE2h3WJkXRIXr2IbYUxvn-rcpR3QmxDt0EfgBz7XOZNB9avkyHDbI3dRYYn-s9jxgo82veTrTMyBmmuBpiyx6w0/s1600/disk_backup16.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;265&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhxxg2Wx67rdpTCQsOzYPDPM7OE4o1swsMyFFHN-5KcR_ODX3I25i0QkE2h3WJkXRIXr2IbYUxvn-rcpR3QmxDt0EfgBz7XOZNB9avkyHDbI3dRYYn-s9jxgo82veTrTMyBmmuBpiyx6w0/s320/disk_backup16.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Right click on the policy and select &quot;Manual Backup&quot;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi3_QZzZpBH2RgqoyXopZe90w5mG7S6uXDLh1mQTJ7I5019lCLGTEcqe-BhknoGj_tNV_zfHWIIMmS4-I16LkAxUlU4ugzUJPEiUx3gL0fnEEg_eWOINeTbznK1H_z7w0Ot4SzIvHPRRPM/s1600/disk_backup17.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;320&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi3_QZzZpBH2RgqoyXopZe90w5mG7S6uXDLh1mQTJ7I5019lCLGTEcqe-BhknoGj_tNV_zfHWIIMmS4-I16LkAxUlU4ugzUJPEiUx3gL0fnEEg_eWOINeTbznK1H_z7w0Ot4SzIvHPRRPM/s320/disk_backup17.jpg&quot; width=&quot;276&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
Here you can select the schedule you want to run (full or incremental), and select the client (or clients) you want to run. Remember- if you have more than one client defined in a schedule, and you do not select a specific client, then the backup will run for *all* clients defined in the policy when you hit ok. If you only want to backup a single client, make sure you select it!&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjmjcn6Q7AHWVpYiZZbItg4lUX47-p1B4Z50KIzPSOwtfU2Ki6uYADDN2JsxovSBimgyGpPmYrvqzy9klvBe9yyhGRWHhMqzVs3CJ87Pt4pYt9UUVqLYboWXhMjxUdxTpfLSUQ_B1N87f4/s1600/disk_backup18.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;320&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjmjcn6Q7AHWVpYiZZbItg4lUX47-p1B4Z50KIzPSOwtfU2Ki6uYADDN2JsxovSBimgyGpPmYrvqzy9klvBe9yyhGRWHhMqzVs3CJ87Pt4pYt9UUVqLYboWXhMjxUdxTpfLSUQ_B1N87f4/s320/disk_backup18.jpg&quot; width=&quot;318&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
As you can see from the job details screen, the backup will run fairly quickly - in our example I only defined a very small set of data to back up, and since disk backups eliminate overhead such as searching for available tapes, moutning, and positioning- our test backup basically ran in a matter of seconds. &lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiU1Uplz0DvPf07_FUffNrOOE9pUa935TbIs-YwJ_TQMZ3BX-HlbhVIlhDTZlZk6tlK9e5clqjyNuu265Quv69mMAuBluBxtGXA53aCp-rDbBvVmjQoqEH98hD9YRMpEuomXIO7oYj6_NY/s1600/disk_backup19.jpg&quot; imageanchor=&quot;1&quot; style=&quot;margin-left: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;208&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiU1Uplz0DvPf07_FUffNrOOE9pUa935TbIs-YwJ_TQMZ3BX-HlbhVIlhDTZlZk6tlK9e5clqjyNuu265Quv69mMAuBluBxtGXA53aCp-rDbBvVmjQoqEH98hD9YRMpEuomXIO7oYj6_NY/s320/disk_backup19.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
If we drill down to the directory we defined in our disk storage unit- you will see a number of files there after you run your first backup. These files ocntian the actual image data, and image metadata (headers etc) that would normally be written to tape. The exact number and type of files per backup depends on what settings and features are enabled (if you have Bare Metal Restore options turned on, there will be extra metadata files above and beyond what is needed for normal backups).&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
&amp;nbsp;&lt;/div&gt;
&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: left;&quot;&gt;
While I am simplifiying things somewhat, this is the general idea. There are several settings and tunings that can be done, and options other than Basic Disk to give access to advanced features such as deduplication. Despite these subtelties, the general process is still surprisingly simple.&lt;/div&gt;
</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/847980104020344151/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/11/set-up-backups-to-disk-for-netbackup-7x.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/847980104020344151'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/847980104020344151'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/11/set-up-backups-to-disk-for-netbackup-7x.html' title='Set up Backups to Disk for NetBackup 7.x'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEincfvdx1Q5bQNgT6KRohUFABETA1UQlVIDqGffZ1NXfEpt6iJB9t_Rjw0Qrnc9ECuN4gh58ytqN7L4XdM_xnGdwGO81jbfP-rstkmP0NTHs474M5mwpeueH21XAOdj9J0EXBe73_3cdTo/s72-c/NetBackup_GUI_jnbSA.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-7206369778242911197</id><published>2012-10-16T21:41:00.000-04:00</published><updated>2012-10-16T21:41:24.240-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="hardware"/><category scheme="http://www.blogger.com/atom/ns#" term="reset"/><category scheme="http://www.blogger.com/atom/ns#" term="RSC"/><category scheme="http://www.blogger.com/atom/ns#" term="rscadm"/><category scheme="http://www.blogger.com/atom/ns#" term="SC"/><category scheme="http://www.blogger.com/atom/ns#" term="scadm"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 9"/><category scheme="http://www.blogger.com/atom/ns#" term="SPARC"/><category scheme="http://www.blogger.com/atom/ns#" term="Sun"/><category scheme="http://www.blogger.com/atom/ns#" term="v240"/><category scheme="http://www.blogger.com/atom/ns#" term="v440"/><category scheme="http://www.blogger.com/atom/ns#" term="v480"/><title type='text'>Reset RSC/SC admin password on a Sun v440 or v480</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj7yWWxzfFreZTkj-n7ehp8ldBIf16FpxOBnjjizeHs2nxMXIiOipBQ3MVEwdrbkqy79W-wuTccmeiKT1-oaesXy3MGaDZY_Vs1mjFbraRi0pgXnbDxdH50KSnCVUjFuNQ-tfshRwYq-3A/s1600/sun-logo.png&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; cssfloat: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;198&quot; nea=&quot;true&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj7yWWxzfFreZTkj-n7ehp8ldBIf16FpxOBnjjizeHs2nxMXIiOipBQ3MVEwdrbkqy79W-wuTccmeiKT1-oaesXy3MGaDZY_Vs1mjFbraRi0pgXnbDxdH50KSnCVUjFuNQ-tfshRwYq-3A/s200/sun-logo.png&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
We&#39;ve all worked in shops where a &quot;new&quot; server purchase meant you were getting a used system that was refurbished (or not). One of my recent projects keyed on leveraging old used equipment from other departments/facilities to keep costs under control, and to leverage the fact that we have the in-house talent to perform maintenance as long as management is willing to pay for replacement parts if something breaks.&lt;br /&gt;
&lt;br /&gt;
Of course, this meant we would have mostly older Sun equipment- from the Sun Fire series- still very good servers with a lot of processing power, but older. Since they were used by parties unknown before being sent to the trash heap (then straight into my greedy little hands from there), there was a good bet that the small stack of v240s, v440s, and v480s I was inheriting would have an admin password set on their system controller cards that I have almost zero chance of guessing. I intend to run these systems without monitor and keyboard, and access them remotely through their SC (or RSC) ports- so I need admin access to the service processors. The solution- hack your way in. It&#39;s actually much easier than it sounds- basically keeping with the Sun philosphy that if you have physical access to the server, you can do what you want with the hardware...&lt;br /&gt;
&lt;br /&gt;
The v240 and v440&amp;nbsp;class of&amp;nbsp;servers are relatively easy. This class also includes close relatives such as the v210. These systems have an integrated System Controller on their mainboards ratehr than a seperate RSC card like the ones you would find on a v480 or v490. The serial management port that belongs to the SC is the only serial port on the system. Luckily, Sun thought there might be a problem if you forgot the SC password (or never had it in the first place), so if you connect a serial terminal to the SC serial mangement port, and leave the session idle for a few minutes- it will drop you back into a console session so you can manipulate the OBP variables and/or login to a running operating system- everything except actually access the system controller prompt where you can perform poweron/poweroff oeprations without touching the front panel. Note that this only happens with the serial interface- the network management port will ONLY present the SC login prompt- to work around this you need physical access to the machine, or access to the serial management port via a terminal server.&lt;br /&gt;
&lt;br /&gt;
Once you have access to a running operating system, you can run the &quot;scadm&quot; command to change some of the basic settings (such as user id definitions and passwords) on the SC itself. Typically, if you were in my situation, you would use this command to reset the password for the &quot;admin&quot; account to something you know. On Solaris 10, and Solaris 9, this command lives in the /usr/platform directory that corresponds to the hardware you are working with. Before you go trying to change a password, you should verify that you know what the admin user name is- it is normally &quot;admin&quot; by default, but this account could have been removed and replaced with another name by some enterprising admin in the server&#39;s past life- so first check by running:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/sbin/scadm usershow&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
This will show you a list of all SC users, and what their permissions are. Most of the time, you will only see one user named &quot;admin&quot; with the permissions listed as &quot;cuar&quot; - if not, the user id you want is the one that has the permissions &quot;cuar&quot;. Once you identify the admin user, you can run the following to change the password to whatever you want it to be:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/sbin/scadm userpassword admin&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(if the admin user is something other than &quot;admin&quot;, obviously, replace &quot;admin&quot; with that user id!)&lt;br /&gt;
&lt;br /&gt;
You will be prompted with the usual two step process- define a password then confirm it. You should now be able to login to the SC.&lt;br /&gt;
&lt;br /&gt;
Alternately, if there is no &quot;admin&quot; user, you can create one, and define whatever password you desire:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/sbin/scadm useradd admin&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/sbin/scadm userperm admin cuar&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/sbin/scadm userpassword admin&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Once you have password access to the SC, you can use the native SC commands to configure the SC&#39;s network management port for remote access- but that is another article.&lt;br /&gt;
&lt;br /&gt;
Resetting the password on a system with an older style RSC on it&#39;s own card rather than integrated onto the board- such as the ones on a v480 or v490, is a little more difficult, but similar in theory. These systems have the serial management port on the RSC card, and they also have a separate serial port. If you have netork access to the running operating system, you can skip this part- but if your only choice is to access the console directly, you&#39;ll need to redirect the console output to the standard serial port. &lt;br /&gt;
&lt;br /&gt;
The first method requires access to the running OS (which sort of defeats the purpose of the exercise)- login to the OS and use the eeprom command to change your OBP variable to use ttya for input and output rather than the rsc-console:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# eeprom input-device=ttya&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# eeprom output-device-ttya&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
After your next boot, the console will be accessible on the standard serial port.&lt;br /&gt;
&lt;br /&gt;
If you cannot access the runnin goperating system, you can still force the system to extend console to the standard serial port by turning the physical keyswitch to the &quot;diag&quot; position, and hard power-cycling the system. After running extended diagnostics, the system will stop at an OBP &quot;ok&quot; prompt on the standard serial port- you may then siwtch the input and output variables to ttya:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;ok setenv input-device ttya&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;ok setenv output-device ttya&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
If you know the root password, you can just reboot the system then login as root on the serial console, or you can just put a fresh load of Solaris 9 or 10 on the system from dvd or from a jumpstart server. Before you cycle the machine again, remember to put the keyswitch back to the normal &quot;run&quot; position or you&#39;ll have to endure extended diags again...&lt;br /&gt;
&lt;br /&gt;
Once you have an operating system running, and you have root access, comes the next challenge. The RSC management software is not actually installed by default. If you are running Solaris 10, you will need to go to &lt;a href=&quot;http://support.oracle.com/&quot;&gt;http://support.oracle.com&lt;/a&gt; - log in, and hope you have a valid Oracle service contract, and the patience to dig up the rsc utilties software from their legacy downloads (it&#39;s not easy to find- in fact, I gave up and went another route). &lt;br /&gt;
&lt;br /&gt;
For Soalris 9, if you have the full media kit distribution, the RSC utility is actually shipped on the Software Supplement CD - there is a directory on this CD containing the RSC utiltiy, and an installer. The RSC directory is usually something similar to /cdrom/cdrom0/RSC-2.2.1 (exact version number will vary depending on the release of Solaris 9 you are dealing with). Luckily, the software from the Solaris 9 cd will work, even if you are runnign Solaris 10- so you can avoid having to deal with Oracle&#39;s support site if you have a Solaris 9 media kit handy. When I attempted to run the installer program on Solaris 10, it didn&#39;t work, but if you drill down into the &quot;Product&quot; subdirectory where the actual packages are, you can install the three required packages manually with pkgadd-&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /cdrom/cdrom0/RSC-2.2.1/Product&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# pkgadd -d . SUNWrsc&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# pkgadd -d . SUNWrscd&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# pkgadd -d . SUNWrscj&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
You will see many other packages in this directory- these are all copies of the three packages listed above, just for different languages/locales.&lt;br /&gt;
&lt;br /&gt;
Once you have the packages installed, you can use the rscadm command line utiltity to change a user id, or completely reconfigure the RSC from scratch. In most cases it may be easier to just reconfigure the whole thing:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/rsc/rsc-config&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
This comman will launch a series of questions that will guide you through all setup steps- including network settings for the network management port, and adding an admin user / password. It takes several minutes to run, but will start you off with a complete new set of known RSC settings.&lt;br /&gt;
&lt;br /&gt;
Alternately, you can use rscadm to change an existing password, similar to the procedure used for the v240/v440 class of servers:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/rsc/rscadm usershow&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(look for the userid with &quot;cuar&quot; permissions- usually &quot;admin&quot;)&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/rsc/rscadm userpassword admin&lt;/span&gt;&lt;br /&gt;
(change the password)&lt;br /&gt;
&lt;br /&gt;
Or, just create a new admin user if one does not exist already-&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/rsc/rscadm useradd &amp;lt;username&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/rsc/rscadm userperm &amp;lt;username&amp;gt; cuar&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# /usr/platform/`uname -i`/rsc/rscadm userpassword &amp;lt;username&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
If you changed the input and ouput OBP variables, and you need to see the console on the RSC card, you need to change these variables back to &quot;rsc-console&quot; - either with the eeprom command, or from the OK prompt:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# eeprom input-device=rsc-console&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# eeprom output-device=rsc-console&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
or &lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;ok setenv input-device rsc-console&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;ok setenv output-device rsc-console&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
You will need to reboot before the change will take effect. You will now be able to use the RSC card&#39;s serial port to access the console, and if you set up the RSC card&#39;s network management port, you will be able to access the console remotely by IP address as well. </content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/7206369778242911197/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/10/reset-rscsc-admin-password-on-sun-v440.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/7206369778242911197'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/7206369778242911197'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/10/reset-rscsc-admin-password-on-sun-v440.html' title='Reset RSC/SC admin password on a Sun v440 or v480'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj7yWWxzfFreZTkj-n7ehp8ldBIf16FpxOBnjjizeHs2nxMXIiOipBQ3MVEwdrbkqy79W-wuTccmeiKT1-oaesXy3MGaDZY_Vs1mjFbraRi0pgXnbDxdH50KSnCVUjFuNQ-tfshRwYq-3A/s72-c/sun-logo.png" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-765581903169394469</id><published>2012-10-10T08:04:00.000-04:00</published><updated>2012-10-10T08:04:36.775-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Avamar"/><category scheme="http://www.blogger.com/atom/ns#" term="Bugs"/><category scheme="http://www.blogger.com/atom/ns#" term="Disaster Recovery"/><category scheme="http://www.blogger.com/atom/ns#" term="EMC"/><category scheme="http://www.blogger.com/atom/ns#" term="mminfo"/><category scheme="http://www.blogger.com/atom/ns#" term="Networker"/><category scheme="http://www.blogger.com/atom/ns#" term="recover"/><category scheme="http://www.blogger.com/atom/ns#" term="save set"/><category scheme="http://www.blogger.com/atom/ns#" term="ssid"/><category scheme="http://www.blogger.com/atom/ns#" term="system state"/><category scheme="http://www.blogger.com/atom/ns#" term="Windows 2003 Enterprise"/><category scheme="http://www.blogger.com/atom/ns#" term="winworkr"/><title type='text'>EMC Networker - Windows 2003 Enterprise, missing system state files.</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEifHogNv89BgqW25-xwawqrW_YuGMON5uJoKNmFWnU6SG_e4Df2EodN-yOvFG70juDwggmHC5sU0drzb4x5m3U8CP6W7om4Z7YVZ5pYaI6ytLrV2-rK9fGmgVj9AiDS82dKN1pR7hozGds/s1600/EMCNetWorker.JPG&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;198&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEifHogNv89BgqW25-xwawqrW_YuGMON5uJoKNmFWnU6SG_e4Df2EodN-yOvFG70juDwggmHC5sU0drzb4x5m3U8CP6W7om4Z7YVZ5pYaI6ytLrV2-rK9fGmgVj9AiDS82dKN1pR7hozGds/s200/EMCNetWorker.JPG&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
Sometimes, it takes a UNIX expert to solve a windows problem. The team was running a proof-of-concept recovering several problematic servers running Windows 2003 Enterprise backed up via Networker 7.6.3 using system state. Windows is typically not my thing- I was attached to make sure the Networker server, running on a Solaris 10 system, was up and ready to go. Once my server was up and ready to go, I was ready to sit back and watch the local windows experts struggle with the dreaded &quot;blue screen of death&quot; for the rest of the day.&lt;br /&gt;
&lt;br /&gt;
So the windows guys restored h data drives with no problems, but the fun started when they attempted to complete the first recovery by applying the system state backup. My understanding of windows system state is somewhat limited - &amp;nbsp;I know that it is a a way to store and recover the system back to the state it was in when the backup was taken. It includes items foreign to us UNIX folks, such as the windows registry, key system files, and other obscure items that do not exist in the UNIX &quot;everything is a file&quot; world.&lt;br /&gt;
&lt;br /&gt;
I also know that a valid system state backup should have three components, SYSTEM FILES, SYSTEM DB, and SYSTEM STATE.&lt;br /&gt;
&lt;br /&gt;
So our windows folks fired up the handy winworkr program- the windows based GUI for Networker clients, which I found interesting since most UNIX folks do just about all of their Networker restores using the command line utilities. They set the browse time to the date we expected to recover from and sighed....&lt;br /&gt;
&lt;br /&gt;
The SYSTEM STATE file was there but SYSTEM FILES and SYSTEM DB were not. They checked other backup dates, but they were all missing one or more of the three critical components. This is when I found out that there was a history of incomplete system state backups with this client, and that the client&#39;s admins would insist that the backups were there and complete, but no matter what, our team could not see them o matter what they set for a browse date or time.&lt;br /&gt;
&lt;br /&gt;
Now they had my attention- why would this client&#39;s Networker admins (who are far better at Networker than I am) insist that the backups were there when it obviously looked like it wasn&#39;t?&lt;br /&gt;
&lt;br /&gt;
The short answer is that they were absolutely right, the backups were there. On the Networker server, &amp;nbsp;I ran a familiar command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mminfo -s myserver -c myclient -v -t &#39;10/04/12&#39;&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
The output clearly showed all three pieces of the system state backup there, each as their own save set. I showed the windows guys, they scratched their heads. We looked at and fiddled with winworkr some more, and still saw the same nonsensical results. The server knew the backups were there, but the client GUI would not show them.&lt;br /&gt;
&lt;br /&gt;
Just for the sheer entertainment of it all, I said &quot;let&#39;s try to restore it from the command line&quot;- a statement sure to get the windows guys to look at you like you just grew a second head. To most windows folks, the command line is like Classical Latin - a dead language that only exists for those crazy UNIX admins, and their slightly less strange cousins, the LINUX admins... Once we got past that awkward moment, they decided to give it a shot.&lt;br /&gt;
&lt;br /&gt;
First I suggested that we run an mminfo on the windows client, just to see if the results would match what I saw on the master server... And it did match, so I decided to press my advantage and suggest we restore by ssid from the command line, since mminfo is telling us what the ssids we need are- so, we went ahead and ran:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# recover -s myserver -S &amp;lt;ssid from mminfo output&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
...three times- once for each system state component using the corresponding ssid number. After the restores finished, the system came up and was stable after going through the usual post-system state cleanup tasks such as adjusting the network configuration to match the hardware we were recovering the system to.&lt;br /&gt;
&lt;br /&gt;
Apparently, the Networker command line utilities work almost exactly the same in both the UNIX and Windows clients- or at least close enough for our purposes. It was clear that the backups we needed existed on the backup server, and were recoverable, but something was preventing the windows client GUI from displaying them correctly.&lt;br /&gt;
&lt;br /&gt;
We still do not know the full answer- but after going back and examining the minfo output for each similar case we encountered, it appears that the condition happens when there are multiple backups on the same date- in each case mminfo showed the three system state backup components listed twice for each date,&amp;nbsp; once in the early morning, and once late at night- possibly due to delays in the previous night&#39;s backup schedule, so that the backup finished after midnight, then a more timely run in the evening, so the two backup sets appear on the same calendar date. We believe there is a bug in how the windows UI selects which backups from the mminfo output to display- and it appears that it is triggered by having two system state backups for the same client occurr on the same calendar day. Shifting around the browse date/time to time stamps both before and after the first backup did not seem to change matters- parts of the system state backup would not appear regardless of the selected date.&lt;br /&gt;
&lt;br /&gt;
Will EMC ever investigate further and come up with a fix? Maybe, maybe not- perhaps they already have fixed it in later maintenance pack levels or minor releases- either way, the command line workaround produces perfectly acceptable results, and seems much more bulletproof. I wouldn&#39;t hold my breath for a bugfix though- the impression I geet from my other dealings with EMC is that they intend to gradually phase Networker out and convince their custoemrs to convert to Avamar as a backup/recovery solution instead. Only time will tell.</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/765581903169394469/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/10/emc-networker-windows-2003-enterprise.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/765581903169394469'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/765581903169394469'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/10/emc-networker-windows-2003-enterprise.html' title='EMC Networker - Windows 2003 Enterprise, missing system state files.'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEifHogNv89BgqW25-xwawqrW_YuGMON5uJoKNmFWnU6SG_e4Df2EodN-yOvFG70juDwggmHC5sU0drzb4x5m3U8CP6W7om4Z7YVZ5pYaI6ytLrV2-rK9fGmgVj9AiDS82dKN1pR7hozGds/s72-c/EMCNetWorker.JPG" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-8721694896058255713</id><published>2012-08-24T16:33:00.000-04:00</published><updated>2012-08-31T16:00:46.735-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Oracle"/><category scheme="http://www.blogger.com/atom/ns#" term="patches"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 11"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 8"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 9"/><category scheme="http://www.blogger.com/atom/ns#" term="Sun"/><category scheme="http://www.blogger.com/atom/ns#" term="sunsolve"/><title type='text'>Sunsolve website returns as a User-driven site</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhMRPJevTfOhyphenhyphenIbd_teQBisfUa_qg8YmyZVQKxF8x8swST3nRnG8u-PtMaKefLka-Eh2da-7ecomEMwsvI4wNza41kYyM4UjouB0LO9osPnRsRea-dcfCq-PHHR_uVANhypusQHvaSAhGE/s1600/wesunsolve.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; cssfloat: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhMRPJevTfOhyphenhyphenIbd_teQBisfUa_qg8YmyZVQKxF8x8swST3nRnG8u-PtMaKefLka-Eh2da-7ecomEMwsvI4wNza41kYyM4UjouB0LO9osPnRsRea-dcfCq-PHHR_uVANhypusQHvaSAhGE/s1600/wesunsolve.jpg&quot; yda=&quot;true&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
Just about any experienced Solaris Admin who was around prior to&amp;nbsp;Oracle&#39;s buyout of Sun Microsystems, relied heavily on information found on the&amp;nbsp;old sunsolve website- especially for information about current patches. When Oracle took over, one of the first things to happen was that they absorbed Sun&#39;s old web properties and converted them over to&amp;nbsp;fit Oracle&#39;s existing web infrastructure. Sunsolve was a casualty&amp;nbsp;of this merge, replaced by the &quot;Oracle Technology Network&quot; - which many users were not happy with.&lt;br /&gt;
&lt;br /&gt;
Now, anindependent Solaris enthusiast named Thomas Gouverneur&amp;nbsp;has resurrected an unofficial, non-Oracle sponsored version of sunsolve called &lt;a href=&quot;http://wesunsolve.net/&quot; target=&quot;_blank&quot;&gt;&quot;wesunsolve&quot;&lt;/a&gt;&amp;nbsp;- this new site has the old-style look and feel of sunsolve, and incorporates patch metadata information (now collected independently)- re-creating some of the most used features of the original sunsolve site. I think the entire Solaris community owes wesunsolve a round of thanks for resurrecting this incredibly useful tool! Thanks!</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/8721694896058255713/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/08/sunsolve-website-returns-as-user-driven.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/8721694896058255713'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/8721694896058255713'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/08/sunsolve-website-returns-as-user-driven.html' title='Sunsolve website returns as a User-driven site'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhMRPJevTfOhyphenhyphenIbd_teQBisfUa_qg8YmyZVQKxF8x8swST3nRnG8u-PtMaKefLka-Eh2da-7ecomEMwsvI4wNza41kYyM4UjouB0LO9osPnRsRea-dcfCq-PHHR_uVANhypusQHvaSAhGE/s72-c/wesunsolve.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-5464699584594447838</id><published>2012-08-21T19:20:00.002-04:00</published><updated>2012-08-21T19:20:19.417-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Containers"/><category scheme="http://www.blogger.com/atom/ns#" term="LDOMs"/><category scheme="http://www.blogger.com/atom/ns#" term="Logical Domains"/><category scheme="http://www.blogger.com/atom/ns#" term="Oracle"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><category scheme="http://www.blogger.com/atom/ns#" term="Virtualization"/><category scheme="http://www.blogger.com/atom/ns#" term="Zones"/><title type='text'>Resolve Multicast Socket failures for Solaris 10 Logical Domains</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s1600/oralogo-small.gif&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; cssfloat: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;26&quot; mda=&quot;true&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s200/oralogo-small.gif&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
Solaris 10 was a late entry into the world of server virtualization. As such, the Solaris virtualization meethods are somewhat quirky- and niether of them can really be called true virtualization.&lt;br /&gt;
&lt;br /&gt;
On the one hand, you have Containers (aka zones) - where you partially virtualize by splitting off a partial new instance of the kernel, and partially simulate virtualization by sandboxing or logically isolating processes from each other. &lt;br /&gt;
&lt;br /&gt;
On the other hand you have LDoms- which again is not quite real virtualization. It is part physical partitioning of resources, and part virtual allocation of resources. Each LDom is essentially it&#39;s own complete server, although a central control domain handles dishing out virtual chunks of resources such as disk and network connections, while processing resources and memory are assigned in a fashion more like a physical server partition. In essence, LDoms bear a strong similarity to an RS/6000 LPar using a VIO server to assign virtual resources to physical hadware partitions.&lt;br /&gt;
&lt;br /&gt;
IBM, however, has been doing this with their RS/6000 servers for years - Sun is a fairly recent entry, and can be plagued with allmanner of bizzare issues.&lt;br /&gt;
&lt;br /&gt;
One of which, I ran into recently. I cannot explain exactly how the system got into this state, or why the borderline insane solution actually works, but I&#39;ll take it, and a dose of whatever the guy on the OTN forums who came up with the solution was smoking when he figured it out.&lt;br /&gt;
&lt;br /&gt;
The problem- while configuring networking virtual switch resources for an Ldom setup, I ran into a seemingly impassable error-&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# ldm add-vsw net-dev=e1000g0 primary-vsw0 primary&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Automatic MAC allocation not available&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Please check your network.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Automatic MAC allocation not available&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Nothing we tried would resolve this- not reverting the service processor to default config, rebooting, resetting, etc. A little research showed that the logical domain manager will assign MAC addresses to new resources as needed, but also sends out multicast messages to check to see if the selected MAC address is already in use. The error is supposed to happen when this check times out after thirty seconds. I got this error immediately with no waiting.&lt;br /&gt;
&lt;br /&gt;
A little more research turned up a somewhat bizzare solution. I still do not know what was really going on, or why meddling with files normally not meddled with would fix this, but it worked.&lt;br /&gt;
&lt;br /&gt;
Later, on another system we bumped into the same error in a slightly different setting, while provisioning a new Ldom- &lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;# ldm add-domain newldom&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;Multicast socket creation failed. Please set LDom MAC addresses manually.&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
This time the error message brings up the point that we also have the option to manually assign a MAC address, which could bypass the broken multicast duplicate check step... however, this was not the workaround I used.&lt;br /&gt;
&lt;br /&gt;
Here&#39;s what I did- I edited the /etc/path_to_inst file. Normally touching this file is considered borderline insane (execpt for a few rare disaster recovery situations where you need to rebuild the device tree). You then search out all entries that correspond to your network interfaces, and any virtual switch devices (vsw) that you&#39;ve already defined, and delete them. (you should probably save a copy of the path_to_inst_file before you edit, a mistake here could potentially render your system unusable). &lt;br /&gt;
&lt;br /&gt;
After saving your edits, run a devfsadm with the cleanup and verbose flags&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# devfsadm -Cvvv&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
which should clean up links in your /dev and /devices directories, then perform a reconfigure reboot-&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# reboot -- -r&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
When the system comes up, you&#39;ll be able to use the ldm command without the annoying Multicast errors. I&#39;ve sucessfully used this fix a few times in a row. As I said before,and will say again, I cannot truthfully say that I know why it works, or what was really wrong in the first place, but it works, and that&#39;s enough for me!&lt;br /&gt;
&lt;br /&gt;
The real credit goes to whoever &quot;SunForumsGuest21&quot; is for coming up with the solution &lt;a href=&quot;https://forums.oracle.com/forums/thread.jspa?threadID=1947825&quot;&gt;here&lt;/a&gt;. Thanks, you got me out of a jam.</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/5464699584594447838/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/08/resolve-multicast-socket-failures-for.html#comment-form' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5464699584594447838'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5464699584594447838'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/08/resolve-multicast-socket-failures-for.html' title='Resolve Multicast Socket failures for Solaris 10 Logical Domains'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s72-c/oralogo-small.gif" height="72" width="72"/><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-5582147624973367608</id><published>2012-07-03T10:35:00.000-04:00</published><updated>2012-07-03T10:38:34.898-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Disaster Recovery"/><category scheme="http://www.blogger.com/atom/ns#" term="Oracle"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 10"/><category scheme="http://www.blogger.com/atom/ns#" term="SPARC"/><category scheme="http://www.blogger.com/atom/ns#" term="ZFS"/><title type='text'>Clone/DR a Solaris 10 system using zfs snapshots</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s1600/zfs_feature_2.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; cssfloat: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;160&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s200/zfs_feature_2.jpg&quot; vca=&quot;true&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
The ZFS file system adds a wide array of powerful features and capabilities to Solaris 10 (and 11). The option to build a system using a zfs-based root file system is frequently overlooked by legacy system administrators in favor of more traditional ufs based root- mainly due to familiarity issues. This, however, cuts away some of the more powerful capabilities of Solaris in the process.&lt;br /&gt;
&lt;br /&gt;
Here is an example of what I am talking about- A ZFS root-based system has the option of using the ZFS snapshot / ZFS send/receive mechanism to clone a system- in a manner similar to flash archive. Why is this technique important, when flash archive recovery is a mature, proven technology that is actually a cleaner and more concise cloning method? The answer is Solaris 11 - under Solaris 11 flash archives are no more- removing the primary method for system cloning/recovery that has been considered the &quot;best practice&quot; method for recovery on Solaris since it was introduced in Solaris 8. Since Solaris 11 uses a ZFS root file system scheme as a standard- a similar method may prove&amp;nbsp;to be useful in Solaris 11 as an alternative to building a system from a new install and attempting to rebuild package inventory from a manifest that may or may not be up to date- a model that is somtimes&amp;nbsp;problematic in the linux world.&lt;br /&gt;
&lt;br /&gt;
This method is also useful in that it makes it easy to recover virtual containers if all the data for the container is included in the root zfs pool (rpool). This model can greatly simplify the process of recovering systems that have been appropriately backed up even if they have a large number of Solaris containers embedded in them.&lt;br /&gt;
&lt;br /&gt;
The first challenge is the setup of the original system. If you need to bring across you containers, they will need to have their zone root directories contained in the root&amp;nbsp;zfs pool (usually named rpool). I built my sample system on a sun e420r using a default zfs-based file system layout. I then created a zfs file system called &quot;rpool/testzone&quot; to act as my zone root, and set permissions to &quot;700&quot; as recommended by Oracle:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zfs create rpool/testzone&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# chmod 700 /rpool/testzone&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
I then configured a simple &quot;whole root&quot; zone that uses a single network interface-&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zonecfg -z testzone&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;testzone: No such zone configured&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Use &#39;create&#39; to begin configuring a new zone.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone&amp;gt; create -b&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone&amp;gt; set zonepath=/rpool/testzone&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone&amp;gt; set autoboot=true&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone&amp;gt; add net&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone:net&amp;gt; set address=10.1.10.12/24&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone:net&amp;gt; set physical=hme0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone:net&amp;gt; set defrouter=10.1.10.1&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone:net&amp;gt; end&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone&amp;gt; commit&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;zonecfg:testzone&amp;gt; exit&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Then installed the new zone:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zoneadm -z testzone install&lt;/span&gt;&lt;br /&gt;
(this install may take ~15 minutes or more)&lt;br /&gt;
&lt;br /&gt;
And start up the new zone:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zoneadm -z testzone boot&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
When the zone starts up, you will need to connect to the zone&#39;s console and complete the standard Soalris sysidcfg questions - i.e. confirming the network settings, and setting the root password.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zlogin -C testzone&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Follow the prompts, then login as root when done.&lt;br /&gt;
&lt;br /&gt;
Just so we have something to prove success, I started up an instance of the apache webserver-&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /etc/apache2&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cp httpd.conf-example httpd.conf&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# svcadm enable apache2&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
We may now verify that our test zone is running by pointing a web browser at it, or manually querying the web server using telnet:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# telnet 10.1.10.12 80&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Trying 10.1.10.12 ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Connected to 10.1.10.12.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Escape character is &#39;^]&#39;.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;GET /&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(you should see the html code for the default apache start page)&lt;br /&gt;
&lt;br /&gt;
Exit out of the zloogin session, and make sure you are working on the global zone.&lt;br /&gt;
&lt;br /&gt;
In order to back this monster up, we will need to take a zfs snapshot- however, by default a snapshot will only create an image of a single file system or dataset. If you look at a zfs list for a zfs rooted server, you will see that there are several nested datasets under the top level rpool - all of which we will need in order to recover/clone our system. The folks who designed ZFS already understood this, so there is a recursive option for zfs snapshot that will create snapshots for all file systems under the named dataset- so we will be able to create our image with one snapshot command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zfs snapshot -r rpool@testsnap&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(where rpool is the top-level dataset, and testsnap is the name of the snapshot)&lt;br /&gt;
&lt;br /&gt;
If you run a zfs list - you will see that a snapshot with the name &quot;testsnap&quot; has been created for rpool, and for every dataset nested beneath it.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;# zfs list&lt;br /&gt;NAME&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; USED&amp;nbsp; AVAIL&amp;nbsp; REFER&amp;nbsp; MOUNTPOINT&lt;br /&gt;rpool&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 15.5G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp;&amp;nbsp; 109K&amp;nbsp; /rpool&lt;br /&gt;rpool@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 20K&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp; 107K&amp;nbsp; -&lt;br /&gt;rpool/ROOT&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 4.54G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp; legacy&lt;br /&gt;rpool/ROOT@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp; -&lt;br /&gt;rpool/ROOT/s10s_u10wos_17b&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 4.54G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp; 4.49G&amp;nbsp; /&lt;br /&gt;rpool/ROOT/s10s_u10wos_17b@testsnap&amp;nbsp; 47.7M&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp; 4.48G&amp;nbsp; -&lt;br /&gt;rpool/dump&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 1.00G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp; 1.00G&amp;nbsp; -&lt;br /&gt;rpool/dump@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 16K&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp; 1.00G&amp;nbsp; -&lt;br /&gt;rpool/export&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 81K&amp;nbsp;&amp;nbsp; 118G&amp;nbsp;&amp;nbsp;&amp;nbsp; 32K&amp;nbsp; /export&lt;br /&gt;rpool/export@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 18K&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp;&amp;nbsp; 32K&amp;nbsp; -&lt;br /&gt;rpool/export/home&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp;&amp;nbsp; 118G&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp; /export/home&lt;br /&gt;rpool/export/home@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp; -&lt;br /&gt;rpool/swap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 2.06G&amp;nbsp;&amp;nbsp; 120G&amp;nbsp;&amp;nbsp;&amp;nbsp; 16K&amp;nbsp; -&lt;br /&gt;rpool/swap@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp;&amp;nbsp; 16K&amp;nbsp; -&lt;br /&gt;rpool/testzone&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 3.94G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp; 3.94G&amp;nbsp; /rpool/testzone&lt;br /&gt;rpool/testzone@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 2.51M&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp; 3.94G&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
In order to use these snapshots as a backup/image, we will need to put the data in a more portable form. We may use the zfs send command with the &quot;Replication&quot; option to dump all of our snapshots into a file. The &quot;Replication&quot; option is similar to the &quot;recursive&quot; snapshot we used above- it will basically combine all the related, nested snapshots into one output stream that we can redirect to a file, so we can manage one file containing the entire OS image rather than several files.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zfs send -R rpool@testsnap &amp;gt; /export/home/snapshots/myhost.snapshot&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
The file this will create can be very large depending on how much actual data is contained in your root pool, and in any zones contained within it. Make sure you have sufficient space on whatever file system you place the file in.&lt;br /&gt;
&lt;br /&gt;
This file is essentially an image of your entire OS and any zones contained within the root pool. You may move it to replicated storage, or back it up with your enterprise backup solution as needed. In our example, we will assume that the file will be available on an NFS share when we attempt to recover it.&lt;br /&gt;
&lt;br /&gt;
In order to recover our snapshot, we will need a &quot;seed&quot; system to start with. It is possible to boot from a DVD and proceed from there, but you have more fallback and troubleshooting options if you start with a base install of Solaris 10 using ufs root file systems. We can build a zfs root pool on a second disk on the system, populate it with our snapshot file, and proceed from there.&lt;br /&gt;
&lt;br /&gt;
First you will need to select a disk to use, and make note of it&#39;s device name. In our example, our &quot;seed&quot; system is a Sun v880, and we will use the second internal disk- c1t1d0 as the new root disk.&lt;br /&gt;
&lt;br /&gt;
Before proceeding, we need to make sure this disk has no existing data- such as bootblocks, zfs pools, ufs file systems, etc on it. To ensure the disk is clean, you may overwrite the disk and it&#39;s disk label withjunk data using the below command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# dd if=/dev/zero of=/dev/rdsk/c1t1d0s2&lt;/span&gt;&lt;br /&gt;
(let the command run for about a minute then control-c to exit)&lt;br /&gt;
&lt;br /&gt;
You should then use the format -e command to partition and label the disk. You should put the entire disk in slice 0, and make sure you give the disk an SMI label if prompted.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# format -e&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
After the format/label process, you should be able to verify with the zdb command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zdb -l /dev/rdsk/c1t1d0s0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
You should get a message similar to &quot;failed to unpack label 0&quot;- if you see what appears to be statistics and settings for a file system, go back and re-wipre the disk before proceeding.&lt;br /&gt;
&lt;br /&gt;
If the original system uses a different version of&amp;nbsp;zpool than your seed system, you will need to know the&amp;nbsp;zpool version of the original system. To get this, run the following on the original system (not the system you are recovering to!):&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zpool get version rpool&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Take note of the version number- you will need it when you create the zpool on the new system.&lt;br /&gt;
&lt;br /&gt;
On the system you are recovering to, create the target root pool you will recover your snapshot to. Note that if you are sure that the &quot;seed&quot; OS/bootable DVD uses the same&amp;nbsp;zpool version as the home system, you may omit the &quot;-o version=&quot; portion of the command, otherwise, replace the value with the version number you discovered above:&lt;br /&gt;
&lt;br /&gt;
&lt;div style=&quot;text-align: left;&quot;&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zpool create -f -o failmode=continue -o version=29 -R /a -m legacy -o cachefile=/etc/zfs/zpool.cache rpool c1t1d0s0&lt;/span&gt;&lt;/div&gt;
&lt;br /&gt;
Now we move on to the heavy lifting. If the directory containing the zfs send file from the home machine is available as an NFS share, mount it now:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mount nfshost:/share/snapshots /mnt&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(or copy the file to the local OS if necessary)&lt;br /&gt;
&lt;br /&gt;
Recover the snapshot into the newly created pool using the zfs &quot;receive&quot; command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cat /mnt/myhost.snapshot | zfs receive -Fdu rpool&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
At this point, the data continaed in the original rpool will be there. We can mount the new root directory and make any modifications needed to account for the new hardware. Other documentation I&#39;ve read about this method stated that you can directly mount the root dataset- however, I have found that it is easier to unmount the pieces of the new root that zfs has attempted to automatically mount, then go back and mount the root. If you do a zfs list, you will see output similar to below:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;# zfs list &lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;NAME&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; USED&amp;nbsp; AVAIL&amp;nbsp; REFER&amp;nbsp; MOUNTPOINT&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 15.5G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp;&amp;nbsp; 109K&amp;nbsp; /a/rpool&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 20K&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp; 107K&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/ROOT&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 4.54G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp; /a/rpool/ROOT&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/ROOT@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/ROOT/s10s_u10wos_17b&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 4.54G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp; 4.49G&amp;nbsp; /a&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/ROOT/s10s_u10wos_17b@testsnap&amp;nbsp; 47.7M&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp; 4.48G&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/dump&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 1.00G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp; 1.00G&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/dump@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 16K&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp; 1.00G&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/export&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 81K&amp;nbsp;&amp;nbsp; 118G&amp;nbsp;&amp;nbsp;&amp;nbsp; 32K&amp;nbsp; /a/rpool/export&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/export@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 18K&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp;&amp;nbsp; 32K&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/export/home&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp;&amp;nbsp; 118G&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp; /a/rpool/export/home&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/export/home@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp;&amp;nbsp; 31K&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/swap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 2.06G&amp;nbsp;&amp;nbsp; 120G&amp;nbsp;&amp;nbsp;&amp;nbsp; 16K&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/swap@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 0&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp;&amp;nbsp;&amp;nbsp; 16K&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/testzone&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 3.94G&amp;nbsp;&amp;nbsp; 118G&amp;nbsp; 3.94G&amp;nbsp; /a/rpool/testzone&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace; font-size: x-small;&quot;&gt;rpool/testzone@testsnap&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; 2.52M&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp;&amp;nbsp; -&amp;nbsp; 3.94G&amp;nbsp; -&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
First unmount any filesystems listed as mounted above:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zfs unmount -a&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Take a quick look at /a - if there are any subdirectories under it, this may block the mount you need to make- so you may have to remove them (make sure they are empty and unmounted first!).&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# rmdir /a/rpool&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Look at the zfs list output above- the file system that shows &quot;/a&quot; as it&#39;s mountpoint is the actual root file system- it&#39;s exact name will vary from build to build. This is the file system we want to mount manually.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zfs mount rpool/ROOT/s10s_u10wos_17b&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Here we can go in and make minor changes to the recovered operating system prior to booting- since we are moving to different hardware, items such as network interfaces, or disk devices, etc may need to be changed to their equivalents on the new hardware. This is also a good time to manually rebuild the device tree of the new root file system using the device files from the &quot;seed&quot;&quot; operating system. There are other ways of cleaning up the device tree, while more simple, they generally are not 100% clean and you will have artifacts such as disk controller numbers and network interface instance numbers offset (i.e. ce0 may appear as ce4, c1t2d0 may appear as c7t2d0 in the recovered operating system). The steps below are a reliable and complete method for fixing this:&lt;br /&gt;
&lt;br /&gt;
(move aside /dev and /devices on the recovered OS and create empty top-level directories for them)&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /a&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mv dev dev.DR&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mv devices devices.DR&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mkdir -m 755 /a/dev&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mkdir -m 755 /a/devices&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(use find and cpio to copy the contents of /dev and /devices on the &quot;seed&quot; OS to the recovered OS)&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /dev&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# find . -depth -print | cpio -pdmv /a/dev&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /devices&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# find . -depth -print | cpio -pdmv /a/devices&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(copy the path_to_inst file from the &quot;seed&quot; OS to the recovered OS)&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /a/etc&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mv path_to_inst path_to_inst.DR&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /etc&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cp path_to_inst /a/etc&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(create a file called &quot;reconfigure&quot; in the recovered root directory so the system will perform a sanity check on the device configuration on the next reboot)&lt;br /&gt;
# touch /a/reconfigure&lt;br /&gt;
&lt;br /&gt;
When we boot the recovered OS, the reconfigure file will force the equivalent of a &quot;boot -r&quot; which will perform a consistency check on our device configuration.&lt;br /&gt;
&lt;br /&gt;
While we have the recovered root file system mounted, this is also a good time to check our /a/etc/vfstab file, and comment out any mounts for file systems that we have not yet built/recovered (if any). You should also look at your /a/etc/hostname.* files and move any of them that need to be changed to accomodate the new hardware- for instance, the original system may have had an /etc/hostname.hme0 - but the new hardware may not have an hme0 interface - if the equivalent interface is eri0 on the new hadware, you may clean this up with:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mv /a/etc/hostname.hme0 /a/etc/hostname.eri0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Additionally, if your recovered system hosts any non-global zones, you may adjust their network interface assignments as well - typically the configuration files for your zones are .xml files found in /a/etc/zones - for example, if we have a zone called &quot;testzone&quot; we would:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# vi /a/etc/zones/testzone.xml&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(note- the comments in the file will warn you not to edit this file! - under normal circumstances, this would be true, but recovering a system this way is not a normal circumstance. Since to global zone OS is not even booted, and the non-global zone we are editing is down, we can freely edit this file as long as we are careful and do not make mistakes!)&lt;br /&gt;
&lt;br /&gt;
Look for lines similar to the below- there should be at least one &quot;network&quot; tab - carefully edit the interface name (physical=&quot;hme0&quot;) portion and change it to the correct interface on the new hardware (in our example, change hme0 to eri0)&lt;br /&gt;
&lt;br /&gt;
Change:&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;lt;zone name=&quot;testzone&quot; zonepath=&quot;/rpool/testzone&quot; autoboot=&quot;true&quot;&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp; &amp;lt;network address=&quot;10.1.10.12/24&quot; physical=&quot;hme0&quot;/&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;lt;/zone&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
to:&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;lt;zone name=&quot;testzone&quot; zonepath=&quot;/rpool/testzone&quot; autoboot=&quot;true&quot;&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;nbsp; &amp;lt;network address=&quot;10.1.10.12/24&quot; physical=&quot;eri0&quot;/&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;&amp;lt;/zone&amp;gt;&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Save and exit the file.&lt;br /&gt;
&lt;br /&gt;
Before we can reboot and test our recovered OS, we still need to make the recovered zpool bootable. First we need to unmount the root file system currently on /a&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# cd /&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zfs unmount rpool/ROOT/s10s_u10wos_17b&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
We now need to set the bootfs property of the pool &quot;rpool&quot; to the dataset we previously mounted on /a.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# zpool set bootfs=rpool/ROOT/s10s_u10wos_17b rpool&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Next we need to install the zfs bootblock on the disk device we used to build our rpool on (in our example this is c1t1d0s0)&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# installboot -F zfs /usr/platform/`uname -i`/lib/fs/zfs/bootblk /dev/rdsk/c1t1d0s0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
We will also need to set our &quot;boot-device&quot; eeprom variable to the corresponding device path or device alias- on most Sun/Oracle systems- the second internal disk is aliased to &quot;disk1&quot; by default (and the third is disk2 etc...), so, in our example, we can use the disk1 alias:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# eeprom boot-device=disk1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
In theory, if we have done our job correctly, all we should have to do to complete recovery of our original system and the non-global zone running on it is to simply reboot the system:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# init 6&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
There may be additional items to troubleshoot if tehre are portions of the original system that live outside of the rpool- there may be file systems to create and restore, and other applications to start. In our example, however, all of our storage, including our non-global zone, were contained in the pool &quot;rpool&quot; - everything should start on boot.&lt;br /&gt;
&lt;br /&gt;
We an test that the web server on our non-global zone is up again by pointing a web browser at the non-global zone&#39;s IP address, or by a telnet to port 80 on the non-global zone&#39;s IP address:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# telnet 10.1.10.12 80&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Trying 10.1.10.12 ...&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Connected to 10.1.10.12.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;Escape character is &#39;^]&#39;.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;GET /&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(you should see the html code for the default apache start page)&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;em&gt;Common pitfalls- &lt;/em&gt;&lt;br /&gt;
&lt;em&gt;&lt;br /&gt;&lt;/em&gt;&lt;br /&gt;
&lt;em&gt;Original OS should be Full + OEM in order to be recoverable on alternate hardware&lt;/em&gt;&lt;br /&gt;
&lt;em&gt;&lt;br /&gt;&lt;/em&gt;&lt;br /&gt;
&lt;em&gt;It is possible to have a container that is not hosted within the root pool, or have some data in rpool, and some in other datasets- this technique only covers recovering rpool! &lt;/em&gt;&lt;br /&gt;
&lt;em&gt;&lt;br /&gt;&lt;/em&gt;&lt;br /&gt;
&lt;em&gt;ZFS version of the DVD or seed OS must match the zfs version of the original system (or you must use a seed OS with a higher version and know the version number of the original- so it can be specified when the&amp;nbsp;new root pool is created.)&lt;/em&gt;&lt;br /&gt;
&lt;em&gt;&lt;br /&gt;&lt;/em&gt;&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/5582147624973367608/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/07/clonedr-solaris-10-system-using-zfs.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5582147624973367608'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/5582147624973367608'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/07/clonedr-solaris-10-system-using-zfs.html' title='Clone/DR a Solaris 10 system using zfs snapshots'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s72-c/zfs_feature_2.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-4263213890408166331</id><published>2012-06-11T10:27:00.000-04:00</published><updated>2012-06-11T10:28:49.706-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="DHCP"/><category scheme="http://www.blogger.com/atom/ns#" term="dladm"/><category scheme="http://www.blogger.com/atom/ns#" term="IP Address"/><category scheme="http://www.blogger.com/atom/ns#" term="ipadm"/><category scheme="http://www.blogger.com/atom/ns#" term="Network"/><category scheme="http://www.blogger.com/atom/ns#" term="Oracle"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris 11"/><category scheme="http://www.blogger.com/atom/ns#" term="Static"/><title type='text'>Static IP Configuration in Solaris 11</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s1600/oralogo-small.gif&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;26&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s200/oralogo-small.gif&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
Many basic system administration tasks that most of us take for granted, change drastically when you move from Solaris 10 to Solaris 11 - and some parts of those tasks have even changed between the Solaris 11 &quot;Express&quot; pre-release version and the first full production release. One task that one would assume would be very simple to figure out on the fly, or with a quick google search, is the simple process of setting and changing a static IP address on an interface. Do not take this step for granted- if you are unfamiliar with Solaris 11, or think you can rely on your extensive knowledge of Solaris 10, you will be mistaken, and may spend more time than you&#39;d care to admit scratching your head in frustration.&lt;br /&gt;
&lt;br /&gt;
The first change is that Solaris 11 comes out of the box, configured to use DHCP to assign network addresses. This is the default, and in fact, Solaris 11 simply assumes that you have a live DHCP service available on your network. You may know all the other commands to configure your network interface, &amp;nbsp;but if you forget that Solaris 11 defaults to, and assumes that there is, DHCP, you will be dead in the water before you start.&lt;br /&gt;
&lt;br /&gt;
Automated network configuration is handled by a Solaris service called &quot;nwam&quot; or &quot;NetWork Auto Magic&quot; - this service is turned on by default, and must be disabled before you can even think about setting a static IP address. Furthermore, you need to turn on the &quot;default&quot; physical networking service as well (why is it called &quot;default&quot; if it&#39;s not on by default? Solaris is anything but consistent in this respect!).&lt;br /&gt;
&lt;br /&gt;
To disable NWAM:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# svcadm disable network/physical:nwam&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
To enable &quot;default&quot; network services (which are not the default settings!):&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# svcadm enable network/physical:default&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
According to the official documentation, these two tasks are the equivalent of running a single netadm command- I haven&#39;t actually used this one myself, but it&#39;s supposed to be:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# netadm enable -p ncp defaultfixed&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Once we&#39;ve enable static addressing, we can continue. If you are not sure what interface to use, or which interfaces are actually connected, you can check with the command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# dladm show-phys&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(this is equivalent to &quot;dladm show-dev&quot; in Solaris 10)&lt;br /&gt;
&lt;br /&gt;
You can also use the traditional &quot;ifconfig -a&quot; to see which interfaces are plumbed, but be advised- ifconfig will only display information, you cannot actually change any meaningful settings with it in Solaris 11.&lt;br /&gt;
&lt;br /&gt;
If your interface is not plumbed, you will need to enable/plumb it (&quot;create&quot; in Solaris 11 terminology):&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ipadm create-ip net0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
(replace net0 with your network interface). This command used to be &quot;ipadm create-if&quot; in the Express version of Solaris 11.&lt;br /&gt;
&lt;br /&gt;
Next we will need to create an &quot;address object&quot; on the interface - this is the equivalent of using ifconfig to set an address (if config cannot do this any more!)&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ipadm create-addr -T static -a local=192.168.1.10/24 net0/v4static&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Note that we set the netmask/broadcast using CIDR notation- expressing the number of bits in the netmask with the number after the &quot;/&quot;. Also take note of the address name at the end of the command- this is a user-defined name in two parts- the part before the &quot;/&quot; needs to be the exact name of the network interface you created in the previous step, and the part after the slash can be whatever the user (that&#39;s you) want to refer to this address as (I usually use &quot;v4static&quot; so I know it&#39;s a statically defined IPv4 address- and because it&#39;s what I saw in the first example of this command I saw in the manuals).&lt;br /&gt;
&lt;br /&gt;
The &quot;ipadm create-addr&quot; is supposed to enable the address by default. If you run the familiar &quot;ifconfig -a&quot; you should see the kind of output you would expect to see if you were on Solaris 10. The &quot;correct&quot; Solaris 11 way to check is:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ipadm show-addr&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
If your address object comes up as disabled, you may need to manually enable it (in practice I&#39;ve had to do this in some cases but not others):&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ipadm enable-addr -t net0/v4static&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
That said- I&#39;ve had issues getting all this to work consistently - I believe there may be other undocumented issues/bugs with the static addressing process. I am not confident that they will be addressed in a timely manner since &quot;DHCP&quot; seems to be what Oracle is pushing as the &quot;right&quot; way to go in their haste to make Solaris 11 appeal to the linux crowd.&lt;br /&gt;
&lt;br /&gt;
If you go through all the steps, and nothing appears to be working, the only advice I can give is to delete the ip interface with:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ipadm delete-ip net0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Then check to make sure that svc:/network/physical:nwam is disabled, and that svc:/network/physical:default is enabled, and start over again at the &quot;create-ip&quot; step. I&#39;ve had this routine work with no issues in one case, and be a nightmare of inconsistent behavior on others- on one run through I ended up with a disabled interface the first two times I ran the procedure, then backed out again and suddenly everything worked. Either the order of operations is extremely sensitive, or there are simply bugs in the current production release. Only time (and future updates) will tell.</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/4263213890408166331/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/06/static-ip-configuration-in-solaris-11.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4263213890408166331'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4263213890408166331'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/06/static-ip-configuration-in-solaris-11.html' title='Static IP Configuration in Solaris 11'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjc1b4bz5CrL3uClZMVSGnbibm_crU6e2q7qdBWxdTUkFlIfq4W9iXmroRTBUmglsR1i9CZ8JfhPT5GIgIBdtnQ_XvhksNA5qDx_1DBpWPFA1vq5duOrgFKZH5Bkam_8Lx0q2Jd5NR9g4A/s72-c/oralogo-small.gif" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-4830369851658464287</id><published>2012-04-14T14:05:00.002-04:00</published><updated>2012-04-14T14:05:05.728-04:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="File System"/><category scheme="http://www.blogger.com/atom/ns#" term="File Systems"/><category scheme="http://www.blogger.com/atom/ns#" term="HP-UX"/><category scheme="http://www.blogger.com/atom/ns#" term="Logical Volume Management"/><category scheme="http://www.blogger.com/atom/ns#" term="Vxfs"/><title type='text'>Grow a File System in HP-UX</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s1600/hp_logo.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; cssfloat: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;188&quot; nda=&quot;true&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s200/hp_logo.jpg&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
HP-UX has well-integrated volume management. Command line utilities, and the SAM management interface make the process relatively simple. The catch is that the SAM interface can become a crutch- in some cases it can make some processes too easy- to the point where an admin who relies on SAM exclusively may no longer remember the command line methods- which are essential for performing deep troubleshooting.&lt;br /&gt;
&lt;br /&gt;
In our example, we&#39;ll look at a fairly common occurrance in a Disaster Recovery scenario - a file system has proven to be too small, and has filled up during a restore. In most cases, modern enterprise backup products are able to pick up a restore from the point where it failed, but the problem still remains- you will need to attach more disk, then add it to the existing file system without destroying the existing data. Once the disk is attached we will need to know a few commands in order to complete the task.&lt;br /&gt;
&lt;br /&gt;
On the more recent versions of HP-UX, you may be able to expand the file system live and in place using the SAM inteface under the &quot;file systems&quot; and &quot;logical volumes&quot; menu items- however- if this fails to expand the file system, you will need to know the manual process.&lt;br /&gt;
&lt;br /&gt;
For purposes of demonstration, we will use a file system built on the &quot;vg01&quot; volume group, using the logical volume &quot;lv01&quot; and mounted on the directory &quot;/myfs&quot; - this file system is currently 100gb in size, and we have attached a new 50gb disk as /dev/dsk/c1t0d0 which we will add to the filesystem. We assume that the disk is already attached and is visible in ioscan (and insf -e has been run if required).&lt;br /&gt;
&lt;br /&gt;
First we will need to initialize our new disk with the &quot;pvcreate&quot; command:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# pvcreate -f /dev/rdsk/c1t0d0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Next, we need to add (extend) our new disk to the volume group the file system lives on.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# vgextend /dev/vg01 /dev/dsk/c1t0d0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Next we need to change the size of the logical volume that the file system lives on- and we need to provide the &quot;lvextend&quot; command with the new desired size of the file system stated in megabytes- in our example, we want the new file system to be 150gb - multiply by 1024 to convert this to megabytes, and we have 153600 megabytes.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# lvextend -L 153600 /dev/vg01/lv01&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Finally, we need to update the actual file system itself, so it understands that it now lives on a lrger logical volume than it did before. If your version of HP-UX does not support on-the-fly file system updates, a &quot;bdf&quot; command will not show the change until this &quot;extendfs&quot; command is run. The challenge is that the file system must be unmounted before we can run &quot;extendfs&quot;. The safe way to do this is to comment the file system out of the &quot;/etc/fstab&quot; file and reboot (then uncomment it and run mountall when finished). You may, however, be able to avoid a reboot by first using the &quot;fuser&quot; command to stop all processes that may be accessing the file system, and unmounting it by hand. This process, however, may fail or hang in some cases, so be aware that you may need to doctor up your fstab and reboot.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# fuser -cku /myfs&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# umount /myfs&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
If you are sucessfull, you may continue on to the &quot;extendfs&quot; step, otherwise, edit the /etc/fstab file to comment out the entry for /myfs, and reboot. Once you have either suceesded in unmounting /myfs, or rebooting without /myfs mounted, run the extendfs command- note that you are using the &quot;raw&quot; device name for the logical volume on the command line.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# extendfs /dev/vg01/rlv01&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Once done, uncomment the entry for /myfs in the /etc/fstab (if needed) and remount the file system.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &amp;quot;Courier New&amp;quot;, Courier, monospace;&quot;&gt;# mount /myfs&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
You should now see the total size of /myfs listed as 150gb ahen you run a &quot;bdf&quot; to check your work.&lt;br /&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/4830369851658464287/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/04/grow-file-system-in-hp-ux.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4830369851658464287'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4830369851658464287'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/04/grow-file-system-in-hp-ux.html' title='Grow a File System in HP-UX'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s72-c/hp_logo.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-8681583456965784207</id><published>2012-03-09T20:33:00.000-05:00</published><updated>2012-08-09T12:20:03.440-04:00</updated><title type='text'>EMC Networker (Legato) - how to find SSIDs for save set recovery</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEifHogNv89BgqW25-xwawqrW_YuGMON5uJoKNmFWnU6SG_e4Df2EodN-yOvFG70juDwggmHC5sU0drzb4x5m3U8CP6W7om4Z7YVZ5pYaI6ytLrV2-rK9fGmgVj9AiDS82dKN1pR7hozGds/s1600/EMCNetWorker.JPG&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;198&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEifHogNv89BgqW25-xwawqrW_YuGMON5uJoKNmFWnU6SG_e4Df2EodN-yOvFG70juDwggmHC5sU0drzb4x5m3U8CP6W7om4Z7YVZ5pYaI6ytLrV2-rK9fGmgVj9AiDS82dKN1pR7hozGds/s200/EMCNetWorker.JPG&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
A quick and dirty method to run restores in Legato from the command line that works on most UNIX or Linux operating sysyems. This method is especially useful if you are limited to a serial console, or a telnet session with no access to a graphical console session.&lt;br /&gt;
&lt;br /&gt;
It is also useful if you are performing a recovery for a client that has not given you a whole lot of information to work with. Replace &quot;backup_server&quot; with the hostname of your backup server, replace &quot;client_hostname&quot; with the hostname of the client you are restoring (usually the system you are logged into), and replace &quot;/directory&quot; with the directory you need to restore.&lt;br /&gt;
&lt;br /&gt;
The first &quot;mminfo&quot; command will display all full backups found for the specified directory. The second &quot;mminfo&quot; command will display info for all backups found for the given host during the specified &quot;time spec&quot; (a time spec is either a date, or a string such a s &quot;2 weeks ago&quot;)- this version of the command will show all backups of all types (full and incremental) for all file systems associated with the client. Both commands are useful to know, and can help you figure out what you need to restore in a disaster recovery situation.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# mminfo -s backup_server&amp;nbsp;&lt;backup server=&quot;&quot;&gt;-v -q &#39;client=client_hostname&lt;client name=&quot;&quot;&gt;,name=/directory&lt;directory&gt;,level=full&#39; | more&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;&lt;/span&gt;&lt;br /&gt;
&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;directory&gt;&lt;br /&gt;&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;
&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;directory&gt;&amp;nbsp;Or&amp;nbsp;&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;&lt;br /&gt;
&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;directory&gt;&lt;br /&gt;&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;
&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;directory&gt;&amp;nbsp;&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# mminfo -s backup_server&amp;nbsp;&lt;backup server=&quot;&quot;&gt;-c client_hostname -v&lt;client name=&quot;&quot;&gt;&amp;nbsp;-t &#39;time spec&lt;time spec=&quot;&quot;&gt;&#39; &amp;nbsp;| more&lt;/time&gt;&lt;/client&gt;&lt;/backup&gt;&lt;/span&gt;&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;&lt;br /&gt;
&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;directory&gt;&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;time spec=&quot;&quot;&gt;&lt;br /&gt;&lt;/time&gt;&lt;/client&gt;&lt;/backup&gt;&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;
&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;directory&gt;&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;time spec=&quot;&quot;&gt;In either case, once you identify the backup save set you need, write down the save set id (or SSID number), them run:&lt;/time&gt;&lt;/client&gt;&lt;/backup&gt;&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;&lt;br /&gt;
&lt;br /&gt;
&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;directory&gt;&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;time spec=&quot;&quot;&gt;&lt;span style=&quot;font-family: Verdana, sans-serif; font-size: x-small;&quot;&gt;&amp;nbsp;&lt;/span&gt;&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# recover -s backup_server&amp;nbsp;&lt;/span&gt;&lt;backup server=&quot;&quot;&gt;&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;-S SSID_number&lt;/span&gt;&lt;ssid&gt;&lt;/ssid&gt;&lt;/backup&gt;&lt;/time&gt;&lt;/client&gt;&lt;/backup&gt;&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;&lt;br /&gt;
&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;directory&gt;&lt;backup server=&quot;&quot;&gt;&lt;client name=&quot;&quot;&gt;&lt;time spec=&quot;&quot;&gt;&lt;backup server=&quot;&quot;&gt;&lt;br /&gt;&lt;/backup&gt;&lt;/time&gt;&lt;/client&gt;&lt;/backup&gt;&lt;/directory&gt;&lt;/client&gt;&lt;/backup&gt;
replace &quot;backup_server&quot; with the hostname of the backup server, and &quot;SSID_number&quot; with the SSID number you wrote down.&lt;br /&gt;
&lt;br /&gt;
Since this is a save set restore, you do not get the option of picking out specific files (at least not using this exact method). This method restores the entire specified directory only. This may not be the best method for day-to-day restores, but it is very useful for a complete rebuild or disaster recovery operation.&lt;br /&gt;
&lt;br /&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/8681583456965784207/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/03/emc-networker-legato-how-to-find-ssids.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/8681583456965784207'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/8681583456965784207'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/03/emc-networker-legato-how-to-find-ssids.html' title='EMC Networker (Legato) - how to find SSIDs for save set recovery'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEifHogNv89BgqW25-xwawqrW_YuGMON5uJoKNmFWnU6SG_e4Df2EodN-yOvFG70juDwggmHC5sU0drzb4x5m3U8CP6W7om4Z7YVZ5pYaI6ytLrV2-rK9fGmgVj9AiDS82dKN1pR7hozGds/s72-c/EMCNetWorker.JPG" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-6534646442475600140</id><published>2012-02-13T22:26:00.000-05:00</published><updated>2012-02-13T22:26:27.217-05:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="File System"/><category scheme="http://www.blogger.com/atom/ns#" term="Logical Volume Management"/><category scheme="http://www.blogger.com/atom/ns#" term="Oracle"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="Storage"/><category scheme="http://www.blogger.com/atom/ns#" term="Sun"/><category scheme="http://www.blogger.com/atom/ns#" term="ZFS"/><title type='text'>Getting started with ZFS file systems</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s1600/zfs_feature_2.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;256&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s320/zfs_feature_2.jpg&quot; width=&quot;320&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
ZFS has been out in the world long enough that is should be part of every Solaris admin&#39;s every day bag of tricks. It was first introduced along with Solaris 10, and it&#39;s continued integration and ease of use/administration has slowly caused it to become much more common. It&#39;s acceptance has been gradual, but steady- in contrast to the rapid paradigm shift I expected to see once I learned how useful, efficient, and user-friendly it is. Now, several years later, we are seeing ZFS used to manage the core root of the Solaris operating system- Solaris 11, in fact, installs on a ZFS based root by default, and it is gaining considerable traction in the realm of NAS filers as well.&lt;br /&gt;
&lt;br /&gt;
Personally, I feel that ZFS is probably the best, and most complete feat of engineering to come out of Sun&#39;s R&amp;amp;D department in recent memory. It is a remarkable achievement considering that most of Sun&#39;s new releases in the last years leading up to their buyout by Oracle have always felt unpolished or unfinished. ZFS was a strong product out of the box, and has continued to grow and improve ever since. That said, if you don&#39;t already know the basics, here are a few tips and concepts to help get you started...&lt;br /&gt;
&lt;br /&gt;
&amp;nbsp;&lt;b&gt;1) You only need to know two commands.&lt;/b&gt;&lt;br /&gt;
&lt;br /&gt;
The only commands you need to administer ZFS file systems are &quot;zpool&quot; and &quot;zfs&quot;. It&#39;s actually a little more difficult than that in practice- each command has a large number of sub-commands and flags/options- but in principal, that&#39;s it.&lt;br /&gt;
&lt;br /&gt;
The &quot;zpool&quot; command is used to create and manage your &quot;zpools&quot; - which, in ZFS terms, is essentially a logical volume group or disk group (for those of you familiar with Veritas). This is where you determine the low level properties for each pool of storage devices you assemble- how they are mirrored/striped, or organized into raid-5 groups.&lt;br /&gt;
&lt;br /&gt;
The &quot;zfs&quot; command is used to create and manage actual file systems (logical volumes really). Here is where you can define properties such as mount point, sharing behavior, disk quotas, etc...&lt;br /&gt;
&lt;br /&gt;
ZFS naming conventions are arranged as a hierarchy. Each object is referred to as a dataset- you may create a zpool named &quot;mypool&quot; that has two additional zfs file systems built on top of it called &quot;fs1&quot; and &quot;fs2&quot; - the proper dataset names for these file systems would be &quot;mypool/fs1&quot; and &quot;mypool/fs2&quot; respectively.&lt;br /&gt;
&lt;br /&gt;
&lt;b&gt;2) Create zpools.&lt;/b&gt;&lt;br /&gt;
&lt;br /&gt;
Creating a simple zpool is a very easy task. The commands may seem similar to metainit (for those familiar with Solaris Volume Manager). A zpool can be built using entire disk devices (recommended), disk slices, or even flat files existing on another file system. Below are a few basic examples:&lt;br /&gt;
&lt;br /&gt;
Create a simple pool using one disk:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zpool create mypool c1t1d0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
If you create a simple zpool using more than one disk, it is automatically striped across all disks. A single disk pool can become a striped pool by simply adding more disks.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zpool create mypool c1t1d0 c2t1d0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
A Mirrored pool gives fault tolerance by forcing two (or more) disks to contain duplicate data- so if one disk fails, the other can still maintain the integrity of the entire pool. &lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zpool create mypool mirror c1t1d0 c2t1d0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
RAID5 arrays give a balance of performance enhancement through striping, and fault tolerance where a single disk can fail, but the array will still be intact. Instead of having a 1:1 copy, RAID5 achieves this by distributing parity information across all devices in the array. The ZFS implementation of RAID5 is called &quot;raidz&quot; - a raidz pool can be created if you have a minimum of three devices. There is also a variant called &quot;raidz2&quot; that contains double the amount of parity information, and can withstand a larger number of disk failures without interruption of data integrity- at the cost of requiring a larger number of disks for the same amount of storage capacity.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zpool create mypool raidz c1t1d0 c2t1d0 c3t1d0&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Note that a top-level zfs file system is created with default settings when a zpool is created. By default, this file system is mounted at a mount point with the same name as the pool- so if we create a zpool called &quot;mypool&quot; it will be mounted as /mypool.&lt;br /&gt;
&lt;br /&gt;
&lt;b&gt;3) Create more zfs file systems, modify properties&lt;/b&gt;&lt;br /&gt;
&lt;br /&gt;
When you create a zpool, you also create a top-level file system on it. You may, however, want to build multiple file systems on the same zpool - this is where the zfs command comes in.&lt;br /&gt;
&lt;br /&gt;
Lets assume we want to crate two more file systems under the pool &quot;mypool&quot; - this is very easy:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs create mypool/newvol1&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs create mypool/newvol2&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
The &quot;mypool/newvol1&quot; name is the name of the dataset corresponding to the new file system. It is the name of the zpool itself, followed by a slash, then the file system name. Since we did not specify any options, the file systems are created with default options- meaning they have no disk quotas, and may dynamically grow as long as there is free disk space in the associated zpool. By default, the file systems are mounted at a mount point corresponding to the dataset name, so our two zfs file systems will be mounted as /mypool/newvol1 and /mypool/newvol2 respectively.&lt;br /&gt;
&lt;br /&gt;
The most common file system property you will want to modify is the mount point. Suppose you intended newvol1 to be a volume that will hold user home directories normally found under /export/home. You can change the mountpoint option using the zfs set subcommand:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs set mountpoint=/export/home mypool/newvol1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
You can also set the mountpoint property when you first create a zfs file system:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs create -o mountpoint=/export/home mypool/newvol1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
The next most common properties you may want to manipulate are quotas and reservations. A quota is a maximum size that zfs will allow the file system to grow to. Quotas can help you manage your storage, by preventing a file system from consuming all available space in a pool with no warning. &amp;nbsp;It is also a convenient way to audit user disk usage- if the user or application needs more disk space than the quota you have set, they will have to ask for it. It also prevents one zfs file system on a pool from consuming space that other file systems may need:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs set quota=10G mypool/newvol1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
This sets the quota (maximum size) for newvol1 to 10 gigabytes.&lt;br /&gt;
&lt;br /&gt;
As with mountpoints, you can set a quota when you create a file system as well:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs create -o quota=10G mypool/newvol1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
A reservation is similar to a quota in that it helps manage space usage. Reservations set a minimum size for a file system - the amount of disk apce specified by the reservation is the minimum amount of space that will be allocated to the file system regardless of how much data it actually contains. If you set a reservation of 5 gigabytes, that space will always be set aside for your file system- other file systems on the same zpool will not be able to dynamically grow into the reserved space. In some production situations, this may help - if you know a file system normally has about 1 gigabyte of data, but needs to grow to 5 gigabytes temporarily during nightly processing, setting a 5 gigabyte reservation may be a good idea so you can always guarantee&lt;br /&gt;
that you will have enough space available to cover your peak usage.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs set reservation=5G mypool/newvol1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
Again, this property can be set at creation time:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs create -o reservation 5G mypool/newvol1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
You may have both a quota and a reservation on a single file system. There are a large number of properties that can be manipulated in a similar fashion. To list all the properties of a file system, use the zfs get subcommand.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs get all mypool/newvol1&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
&lt;b&gt;4) Informational commands&lt;/b&gt;&lt;br /&gt;
&lt;br /&gt;
You will, of course, need to check the status of your zpools and zfs filesystems from time to time. One of the more useful methods is a &quot;zpool status&quot; - which will give a general status of all your zpools, and will show if any hardware or data corruption errors have been encountered.&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zpool status&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp; pool: shared&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp;state: ONLINE&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp;scrub: none requested&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;config:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; NAME &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp;STATE &amp;nbsp; &amp;nbsp; READ WRITE CKSUM&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; shared &amp;nbsp; &amp;nbsp; &amp;nbsp;ONLINE &amp;nbsp; &amp;nbsp; &amp;nbsp; 0 &amp;nbsp; &amp;nbsp; 0 &amp;nbsp; &amp;nbsp; 0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; c0t1d0s0 &amp;nbsp;ONLINE &amp;nbsp; &amp;nbsp; &amp;nbsp; 0 &amp;nbsp; &amp;nbsp; 0 &amp;nbsp; &amp;nbsp; 0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; c0t2d0s0 &amp;nbsp;ONLINE &amp;nbsp; &amp;nbsp; &amp;nbsp; 0 &amp;nbsp; &amp;nbsp; 0 &amp;nbsp; &amp;nbsp; 0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; c0t3d0s0 &amp;nbsp;ONLINE &amp;nbsp; &amp;nbsp; &amp;nbsp; 0 &amp;nbsp; &amp;nbsp; 0 &amp;nbsp; &amp;nbsp; 0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;errors: No known data errors&lt;/span&gt;&lt;br /&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
To check I/O performance, you can use &quot;zpool iostat&quot; to view activity in a format similar to the unix &quot;iostat&quot; command:&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;#&amp;nbsp;&amp;nbsp;zpool iostat 1&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp;capacity &amp;nbsp; &amp;nbsp; operations &amp;nbsp; &amp;nbsp;bandwidth&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;pool &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; used &amp;nbsp;avail &amp;nbsp; read &amp;nbsp;write &amp;nbsp; read &amp;nbsp;write&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;---------- &amp;nbsp;----- &amp;nbsp;----- &amp;nbsp;----- &amp;nbsp;----- &amp;nbsp;----- &amp;nbsp;-----&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;shared &amp;nbsp; &amp;nbsp; &amp;nbsp;3.76G &amp;nbsp; 200G &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;8 &amp;nbsp; &amp;nbsp;108&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;shared &amp;nbsp; &amp;nbsp; &amp;nbsp;3.76G &amp;nbsp; 200G &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;shared &amp;nbsp; &amp;nbsp; &amp;nbsp;3.76G &amp;nbsp; 200G &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;shared &amp;nbsp; &amp;nbsp; &amp;nbsp;3.76G &amp;nbsp; 200G &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;0&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;^C&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;#&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
To view file system capacity, use &quot;zfs list&quot; to view usage similar to the solaris &quot;df&quot; command (df will also give you similar statistics):&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;#&amp;nbsp;&amp;nbsp;zfs list&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;NAME &amp;nbsp; &amp;nbsp; USED &amp;nbsp;AVAIL &amp;nbsp;REFER &amp;nbsp;MOUNTPOINT&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;shared &amp;nbsp;3.76G &amp;nbsp; 197G &amp;nbsp;3.76G &amp;nbsp;/shared&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;br /&gt;
&lt;b&gt;5) Create a snapshot&lt;/b&gt;&lt;br /&gt;
&lt;br /&gt;
File system snapshots are a quick and easy way to provide rudimentary data protection. A Snapshot is a point-in-time image of the parent file system. Snapshots have the advantage of using very little disk space- initially they occupy no space, and as the parent file system changes over time, the snapshot will retain a copy of any files that have changed since the snapshot was taken. Snapshots are very easy to create- simply use the &quot;zfs snapshot&quot; sub-command. &amp;nbsp;A snapshot name begins with the name of the file system/dataset itself, and &quot;@&quot; symbol, then a snapshot identifier. The identifier can be any arbitrary name, but it is customary to make it something representing the date and/or time of the snapshot:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;#&amp;nbsp;zfs snapshot shared@02122012&lt;/span&gt;&lt;br /&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
Snapshots will show up in a &quot;zfs list&quot; - notice how the &quot;used&quot; space for the newly created snapshot is 0 - indicating that no files have changed since the snapshot was taken. The &quot;refer&quot; field claim that the snapshot is 3.76gb - also correct- the image refers to 3.76gb of data (the size of the parent file system) - but since nothing has changed since the snapshot time, there is no need to store a duplicate of all this data. Data is only duplicated when something changes in the original file system.&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;#&amp;nbsp;zfs list&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;NAME &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp;USED &amp;nbsp;AVAIL &amp;nbsp;REFER &amp;nbsp;MOUNTPOINT&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;shared &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; &amp;nbsp; 3.76G &amp;nbsp; 197G &amp;nbsp;3.76G &amp;nbsp;/shared&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;shared@02122012 &amp;nbsp; &amp;nbsp; &amp;nbsp;0 &amp;nbsp; &amp;nbsp; &amp;nbsp;- &amp;nbsp;3.76G &amp;nbsp;-&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
If you turn on the &quot;snapdir&quot; option, you will be able to access a read-only version of the snapshot:&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# zfs set -o snapdir=visible shared&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
The &quot;snapdir&quot; can be found in the &quot;.zfs&quot; directory under the file system&#39;s top level directory.&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# cd /shared/.zfs/snapshot&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ls&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;02122012&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
This provides a quick way to recover a lost file back to the state it was in when the snapshot was taken. After creating a snapshot, if you delete a file in the parent file system, you can drill down through the snapshot directory and find a copy exactly as it was at the time the snapshot was taken:&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ls -l /shared/testfile&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;-rw-r--r-- &amp;nbsp; 1 root &amp;nbsp; root &amp;nbsp; &amp;nbsp;12567 Feb 11 2012 testfile&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# rm /shared/testfile&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
We can restore this file by drilling down into the snapshot directory&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# cd /shared/.zfs/snapshot/02122012&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ls -l testfile&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;-rw-r--r-- &amp;nbsp; 1 root &amp;nbsp; root &amp;nbsp; &amp;nbsp;12567 Feb 11 2012 testfile
&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# cp testfile /shared&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
This guide, while lengthy, is not exhaustive. There are many other features to explore, but these represent the most basic you&#39;ll need to get started. The beauty and strength of zfs is in it&#39;s simplicity- it takes complex concepts, and gives them simple and elegant implementations. The ease of use alone makes zfs an extremely powerful tool to add to your arsenal.&amp;nbsp;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/6534646442475600140/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/02/getting-started-with-zfs-file-systems.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/6534646442475600140'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/6534646442475600140'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/02/getting-started-with-zfs-file-systems.html' title='Getting started with ZFS file systems'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg6f0hPzTzvucqFDyOBqtEvQb8FNNVXYomVpIj8PoQxNmpDJ-iQRH4psH-rt7-6NxJEn6OkOGiSb6fNBVphi7EtPcHkCRfv3lB_Qr29zYCzgiVDGT9feXbEIh31org7ejM5l6yqe7fFmKU/s72-c/zfs_feature_2.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-3910339562375562104</id><published>2012-01-30T11:54:00.001-05:00</published><updated>2012-01-30T11:55:25.446-05:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Bugs"/><category scheme="http://www.blogger.com/atom/ns#" term="HP-UX"/><category scheme="http://www.blogger.com/atom/ns#" term="ignite"/><title type='text'>HP-UX make_sys_image bug</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s1600/hp_logo.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;187&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s200/hp_logo.jpg&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
I stumbled on what appears to be a bug in the make_sys_image command under HP-UX 11i 11.31. Admittedly, this may be an issue with the patch level I was running at, but it does appear to be very real.&lt;br /&gt;
&lt;br /&gt;
The make_sys_image command is really just a utility script, written to help create an ignite image of a system so it can be cloned or recovered entirely- similar to the Solaris &quot;flarcreate&quot; command. While I am relatively new to supporting HP-UX on a regular basis, we have a standard recommended method for producing ignite images that has been used for years. I knew that an ignite image of the system I was working with would need about 3-4 gb of free space for the final file, and possibly more to use as working space for it&#39;s compression routines. The recommended method is:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# make_sys_image -v -u -s local -d /image&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
For the target directory /image, I built a very large vxfs file system- about 30gb, on a new volume group on an unused disk, making sure that the&lt;a href=&quot;http://var-adm.blogspot.com/2012/01/hp-ux-useful-tip-enable-largefiles-on.html&quot;&gt; largefiles option is turned on&lt;/a&gt;.&lt;br /&gt;
&lt;br /&gt;
Much to my surprise, the make_sys_image failed- the error indicated that script detected that there was not enough space for gzip to store the archive. This is strange, since I have nearly 10 times the required space, and I used the exact same command, and file system size to build an ignite image on a serve running 11.23 about an hour earlier.&lt;br /&gt;
&lt;br /&gt;
After investigating things a little, and determining that I was not insane, that there really was no space issue. I also actually read the man page, instead of running the command blindly, and found that the &quot;-u&quot; option tells the script to check the amount of available disk space before attempting to write the file to disk- so I took a guess, and ran the script without the -u option, and it worked with no problems. I double checked myself and made sure that the image I produced would actually extract and produce the expected results.&lt;br /&gt;
&lt;br /&gt;
Maybe it&#39;s a quirk of whatever patch level I&#39;m running, maybe it was the file system type I was using- but one thing is clear- the &quot;-u&quot; option fails miserably under 11.31, but seems to work as advertised on 11.23. From now on, my standard method will be:&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# make_sys_image -v -s local -d /image&lt;/span&gt;
&lt;br /&gt;
&lt;br /&gt;
I&#39;ll worry about making sure there is enough space on my own!</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/3910339562375562104/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/01/hp-ux-makesysimage-bug.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/3910339562375562104'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/3910339562375562104'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/01/hp-ux-makesysimage-bug.html' title='HP-UX make_sys_image bug'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s72-c/hp_logo.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-4909881276919979462</id><published>2012-01-21T09:14:00.004-05:00</published><updated>2012-01-21T09:15:42.171-05:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="File Systems"/><category scheme="http://www.blogger.com/atom/ns#" term="HP-UX"/><category scheme="http://www.blogger.com/atom/ns#" term="Vxfs"/><title type='text'>HP-UX useful tip - enable largefiles on the fly.</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s1600/hp_logo.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;188&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s200/hp_logo.jpg&quot; width=&quot;200&quot; /&gt;&lt;/a&gt;&lt;/div&gt;
&amp;nbsp;While Solaris is my normal UNIX of choice, my work commitments and projects mean that I often delve into other UNIX flavors- and when you are used to using one style UNIX over the other, the similarities (hey it&#39;s still UNIX) can often lead to more mistakes than anything else- it&#39;s very easy to forget that the assumptions you can make while working with one UNIX are no longer valid when you switch to another. Case in point- Solaris file systems always have &quot;largefiles&quot; support (support for files larger than 2gb) turned on by default. HP-UX 11i is another story altogether...&lt;br /&gt;
&lt;br /&gt;
My situation was this- I built a very basic web services stack on an itanium server running HP-UX 11i to use as a basic set of test data for a project. I wanted to hang onto an image of the server in case my development project made a mess of it, or in case I needed a similar testbed for other projects, so used the HP-UX logical volume management commands to build a new vxfs file system, and set out to dump an ignite image to it with the make_sys_image command.&lt;br /&gt;
&lt;br /&gt;
Luckily, I was paying attention to the output- I happened to notice that an error message scrolled by from the gzip compression step of the make_sys_image run claiming &quot;gzip: stdout: File too large&quot;. I knew I did not run out of disk space- the target file system I built was easily 5 times as much space as I could ever possibly need. I also noticed that my ignite image appeared to be exactly 2gb in size- and I expected it to be significantly larger.&lt;br /&gt;
&lt;br /&gt;
In retrospect, I should have realized what had happened right away, however, coming from a Solaris background, the problem was something I would never normally see. HP-UX vxfs file systems do not have &quot;largefiles&quot; turned on by default. Under Solaris, the opposite is true, so I don&#39;t think I&#39;ve seen a &quot;largefiles&quot; limitation problem on a Solaris system in at least 10 years... A few google searches later, I had my answer.&lt;br /&gt;
&lt;br /&gt;
Under HP-UX, fixing this is fairly easy. If you&#39;re not worried about losing data, you could always destroy and rebuild the file system with the largefiles option turned on, or you can do it the easy way-&lt;br /&gt;
&lt;br /&gt;
&lt;div style=&quot;text-align: center;&quot;&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# fsadm -F vxfs -o largefiles /mountpoint&lt;/span&gt;&lt;/div&gt;
&lt;br /&gt;
&quot;vxfs&quot; is the file system type, &quot;largfiles&quot; is obviously the &quot;largefiles&quot; support option, and &quot;/mountpoint&quot; is the directory where your file system is mounted. With no additional steps needed, a second run of make_sys_image was now able to produce a valid image.</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/4909881276919979462/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2012/01/hp-ux-useful-tip-enable-largefiles-on.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4909881276919979462'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/4909881276919979462'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2012/01/hp-ux-useful-tip-enable-largefiles-on.html' title='HP-UX useful tip - enable largefiles on the fly.'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiXSfQSTHK6O9jIYzyTtj1WlBxGuvEFJa9iFfNWNvjnhyphenhyphenmnTPBdcu2qJAuZV2acqwsf9h8fs1pyb236P4wIfaAAosIbWAj6EA100AvmF6Tzo2TX1ot6NLviazHV-6mHyRvriIYjtTvAeDY/s72-c/hp_logo.jpg" height="72" width="72"/><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-1407002244064487228.post-7310207303300285865</id><published>2011-12-21T10:56:00.000-05:00</published><updated>2011-12-21T14:31:55.638-05:00</updated><category scheme="http://www.blogger.com/atom/ns#" term="Disaster Recovery"/><category scheme="http://www.blogger.com/atom/ns#" term="Netbackup"/><category scheme="http://www.blogger.com/atom/ns#" term="Solaris"/><category scheme="http://www.blogger.com/atom/ns#" term="SPARC"/><category scheme="http://www.blogger.com/atom/ns#" term="Sun"/><title type='text'>5 NetBackup DR tips on Solaris</title><content type='html'>&lt;div class=&quot;separator&quot; style=&quot;clear: both; text-align: center;&quot;&gt;
&lt;a href=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEincfvdx1Q5bQNgT6KRohUFABETA1UQlVIDqGffZ1NXfEpt6iJB9t_Rjw0Qrnc9ECuN4gh58ytqN7L4XdM_xnGdwGO81jbfP-rstkmP0NTHs474M5mwpeueH21XAOdj9J0EXBe73_3cdTo/s1600/NetBackup_GUI_jnbSA.jpg&quot; imageanchor=&quot;1&quot; style=&quot;clear: left; float: left; margin-bottom: 1em; margin-right: 1em;&quot;&gt;&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;img border=&quot;0&quot; height=&quot;240&quot; src=&quot;https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEincfvdx1Q5bQNgT6KRohUFABETA1UQlVIDqGffZ1NXfEpt6iJB9t_Rjw0Qrnc9ECuN4gh58ytqN7L4XdM_xnGdwGO81jbfP-rstkmP0NTHs474M5mwpeueH21XAOdj9J0EXBe73_3cdTo/s320/NetBackup_GUI_jnbSA.jpg&quot; width=&quot;320&quot; /&gt;&lt;/span&gt;&lt;/a&gt;&lt;/div&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;NetBackup has saved the day many times, allowing for quick recovery of lost files, or even entire systems. In some cases it may be used to retrieve old data for legal/auditing purposes- either way, it&#39;s a popular enterprise level solution for backup and recovery.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;But what do you do when you need to recover the backup environment itself? If some freak disaster, such as a meteor turning your data center into a big hold in the ground, happens, you&#39;ll probably lose your precious backup server as well as your production systems. You&#39;ll need to know what to do if all you&#39;re left with is a truckload of backup tapes held at your vital records vendor, a disaster booking with your Disaster Recovery provider, and a very short Recovery Time Objective to meet. The clock is ticking, and the CIO is already calling to demand that your get your tier 1 applications back online yesterday. These five tips should help you get past most of the common problems associated with recovering a NetBackup environment from the ground up.&lt;/span&gt;&lt;br /&gt;
&lt;a name=&#39;more&#39;&gt;&lt;/a&gt;&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;b&gt;&lt;i&gt;Question 1-&lt;/i&gt; I see there are x number of tape drives listed in /dev/rmt, but when I tell NetBackup to auto configure devices, it does not see all the tape drives, (or sometimes the robotic arm)- your hardware must be broken. Fix it. NOW.&lt;/b&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;This is probably one of the most common complaints I&#39;ve seen in NetBackup-based recoveries- especially ones using older Solaris releases (such as Solaris 8 or 9) before the Sun Leadville driver came along and made dealing with fiber/SAN attached devices less troublesome. What is happening here has nothing to do with the hardware being broken. Basically, Solaris sees all the devices associated with the silo that it should see. That alone is proof that the library is not &amp;nbsp;broken, or that your SAN zoning is incorrect. What is happening is that the default settings for the SCSI pass through driver, or sg driver, used by NetBackup is not configured for a wide enough range of target and lun numbers to cover all your devices.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Fixing this is simple. you can do it with a few simple commands-&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# mv /kernel/drv/sg.conf /kernel/drv/sg.conf.ORIG&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# cd /usr/openv/volmgr/bin/driver&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ../sg.build all -mt 10 -ml 10&amp;nbsp;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;(builds sg driver settings for up to 10 targets with 10 luns per target - change these numbers as needed for your environment- 10 and 10 is usually more than enough to cover most environments)&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Finish the process-&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# ./sg.install&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;In most cases, you will not need to reboot- but if you are using an older operating system and very old HBA drivers, just reboot to save yourself some headaches. You can also re-instantiate the sg driver using modunload and modload, followed by an update_drv:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# modinfo | grep sg&lt;/span&gt;&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&amp;nbsp;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;find the sg driver, note the id number at the beginning of the line&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# modunload -i (id number)&lt;/span&gt;&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&amp;nbsp;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Unloads the kernel module indicated by the id number&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;# modload /kernel/drv/sg&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Reloads the sg driver into the kernel&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# update_drv sg&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Dynamically re-reads the driver configuration file - redundant if you&#39;ve already done a modunload/modload, but useful to know.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;You may need to rerun devfsadm to make sure all the links under /dev/sg are built properly.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Verify that NetBackup can see all devices now- NetBackup has two utilities to scan for and get information about devices- sgscan and scan. sgscan is the original, &amp;nbsp;it simply reports all devices seen by the sg driver, and lists some basic information about them (including serial numbers etc). Scan gives a little more information, and when scan sees a robotic library device, it will list the &amp;nbsp;serial numbers for each drive number, and will also list serial numbers for each drive&#39;s individual entry. This is basically how the NetBackup GUI matches up tape drive device names with their position within the library- so if scan can see it, NetBackup will. Run either of the below commands to verify what devices the sg driver is seeing:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# /usr/openv/volmgr/bin/sgscan&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;or&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# /usr/openv/volmgr/bin/scan&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;You may need to go back and re-run the process again with larger target and lun numbers, but probably not. If you see what you need to see, go back to the GUI and run the device configuration wizard again.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;b&gt;&lt;i&gt;Question 2-&lt;/i&gt; I can start recovering my catalog using the DR file, but the job fails/hangs when actually trying to mount the tape...&lt;/b&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;If your production environment uses multiple media servers in addition to your master server, your catalog tape may actaully be &quot;owned&quot; by a media server that is not being built/recovered in your DR environment. For a normal restore you could just alias the media server to the master by adding a FORCE_RESTORE_MEDIA_SERVER directive to your bp.conf file, but this will not work for a catalog restore. There is actually a very easy solution for this- simply edit the DR file and change any references to the media server to the host name of the master server, and the catalog tape will now think that it is owned by the master. In the file, you will see several lines similar to the ones below:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# FRAG: c# f# K rem mt den fn id/path host bs off md dwo f_flags f_unused1 exp mpx rl chkpt rsm_nbr seq_no media_subtype keep_date copy_date i_unused1&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;FRAGMENT 1 -1 68571 0 2 6 3 TT2112 &lt;b&gt;&lt;span style=&quot;color: red;&quot;&gt;media_srv&lt;/span&gt;&lt;/b&gt; 1048576 117797 1318513833 8 0 *NULL* 0 0 0 0 0 0 1 0 0 0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;FRAGMENT 1 1 120508992 0 2 6 2 TT2112 &lt;b&gt;&lt;span style=&quot;color: red;&quot;&gt;media_srv&lt;/span&gt;&lt;/b&gt; 1048576 110 1318513833 8 0 *NULL* 1323870795 0 65540 0 0 0 1 0 1318526617 0&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;#DR_MEDIA_REC: ver rtype mtype/den host c# p# classes# classes kbytes nimg vimg hsize off subtype client_type sched_type run_time id/path id_ext&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;DR_MEDIA_REC 0 0 6 &lt;b&gt;&lt;span style=&quot;color: red;&quot;&gt;media_srv&lt;/span&gt;&lt;/b&gt; 1 1 0 120685275 3 3 1024 117866 0 35 0 1318513995 TT2112 TT2112&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;DR_MEDIA_REC 0 0 6 &lt;b&gt;&lt;span style=&quot;color: red;&quot;&gt;media_srv&lt;/span&gt;&lt;/b&gt; 1 1 0 107712 1 1 1024 110 0 7 0 1318513833 TT2112 TT2112&lt;/span&gt;&lt;br /&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Note the highlighted entries- these are the host names for the media server that has ownership of the image or images needed. simply change them to the name of your master server in the DR environment. Also, scan the rest of the file, and change any other similar references- the exact format of the DR file varies slightly depending on which version of NetBackup you are using.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;b&gt;&lt;i&gt;Question 3-&lt;/i&gt; I know which tape is my catalog tape, but I do not have a DR file- how do I recover my catalog?&lt;/b&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;The catalog is still recoverable as long as you know which tape or tapes your catalog backup lives on- hopefully you&#39;ve pre-planned things well enough so that you are putting your catalog backups on easy to identify tapes- maybe assign them to a range of barcode labels that start with a very different character than normal, so you can easily pick it out from a batch of 1200 tapes in an emergency.&amp;nbsp;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;The solution here is to use the bpimport command to scan the tape, and essentially perform a stage 1 tape import on the catalog tape. The DR file itself is actually just a dump of raw data from the catalog database- the same information that gets pulled from the image headers on a tape when you do a stage 1 import. This information is used to tell netbackup which tape to use, and how to find the catalog image on the tape itself. Assuming you know that the catalog tape is media id &quot;TT2112&quot; you can use bpimport to generate the DR file for you, as seen below:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# bpimport -drfile TT2112 -drfile_dest /var/tmp&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;This command will read the tape, pull the information it needs to build a DR file, and dump it in the /var/tmp directory. Once you&#39;ve obtained the DR file this way, you can recover your catalog normally. It takes a little extra time to scan through the tape and pick out the image headers, but it beats attempting to manually import 1000+ tapes. NOTE- this only works with Netbackup 6.0 MP 6 or later!&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;b&gt;&lt;i&gt;Question 4-&lt;/i&gt; My catalog restore was successful, but now what do I do?&lt;/b&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Your catalog restore will bring in all the image info needed to find what it needs on your tapes, but it will also wipe out any device configuration you&#39;ve done before recovering the catalog, and pulled in all your device and media server configuration from the production environment- but in DR, most environments use a very simplified version of the backup environment. Where there may be three or four media servers, and possibly multiple tape libraries at home, in the DR world you will &amp;nbsp;perform your recovery using a single master server, probably with only one tape library, that is very likely a different model and/or configuration than the one(s) in the production environment. Why is that the norm? Simply put, time. In a disaster the goal needs to be getting up and running quickly NOT exactly duplicating a complex backup environment. You need to get tape spinning as soon as possible. In most cases, the performance gains you can have by building out multiple media servers are wiped out by the extra time it takes to build, configure, and troubleshoot them. If you need performance, you have other options, such as a faster network connection, or multiple network connections teaming together to load balance traffic, or just use a server with more cpu/memory or mroe I/O cards to spread the tape drive load. In most cases, simplicity removes enough potential stumbling blocks that you can move ahead more efficiently, and get your restores running sooner.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;In any case, you&#39;ll need to clean up your device configuration, and account for the fact that you do not have a full complement of media servers. You&#39;ll also need to disable any backup jobs, as they are obviously not needed, and will probably just get in the way when you are performing a recovery. You can always turn these back on later.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;1) Shut down netbackup as soon as the catalog is recovered.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# /usr/openv/netbackup/bin/bp.kill_all&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;2) Add FORCE_MEDIA_RESTORE_SERVER directives to bp.conf&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;If you have any media servers in your production environment that will not exist in the recovery environment, add an entry to bp.conf for each media server that will tell NetBackup to use the master server instead- suppose you have two media servers at hom names media_1 and media_2 , you would need to add the following two lines:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;FORCE_MEDIA_RESTORE_SERVER = media_1 master&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;FORCE_MEDIA_RESTORE_SERVER = media_2 master&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;(replace &quot;master&quot; with the host name of your master server) Now if you run a restore that needs an image assigned to one of the above media servers, the master will be able to handle the request. Without this addition, NetBackup will try in vain to send the restore request to the non-existent media server until the job times out and fails.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;3) Perform a partial startup of the emm database (without the job manager) so you can perform operations on the database without jobs attempting to kick off.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# /usr/openv/netbackup/bin/nbdbms_stop_start start&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# /usr/openv/netbackup/bin/nbemm&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;4) Deactivate all media servers not being recovered&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;span style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;# /usr/openv/netbackup/bin/admincmd/nbemmcmd -updatehost -machinename&amp;nbsp;(media server name)&amp;nbsp;&lt;/span&gt;&lt;media style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;-machinestateop set_admin_pause -machinetype media -masterserver (master server name)&lt;/media&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;Repeat this command for each media server you need to deactivate. Then startup the nbevtmgr and bpdm processes:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;tt style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span class=&quot;FIXEDFONT&quot;&gt;# /usr/openv/netbackup/bin/nbevtmgr&lt;/span&gt;&lt;/tt&gt;&lt;/span&gt;&lt;br /&gt;
&lt;tt style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span class=&quot;FIXEDFONT&quot; style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# /usr/openv/netbackup/bin/initbpdbm&lt;/span&gt;&lt;/tt&gt;
&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;5) Delete all storage units. This is easier to do from the GUI- just highlight your storage units, right click, and select delete. If you do it this way, make sure you delete any storage unit groups first, then delete all storage units. This can also be done from a command line:&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;#&amp;nbsp;&lt;span style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;/usr/openv/netbackup/bin/admincmd/bpstulist -go | cut -f 1 -d &#39; &#39; &amp;gt; /tmp/stu_groups&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;background-color: white; color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace; line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;tt style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span class=&quot;FIXEDFONT&quot; style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# /usr/openv/netbackup/bin/admincmd/bpstulist | cut -f 1 -d &#39; &#39; &amp;gt; /tmp/stu_list&lt;/span&gt;&lt;/tt&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;br style=&quot;background-color: white; color: #333333; line-height: 20px;&quot; /&gt;&lt;tt style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span class=&quot;FIXEDFONT&quot;&gt;# for i in `cat /tmp/stu_groups` ; do echo &quot;/usr/openv/netbackup/bin/admincmd/bpstudel -group $i&quot; ; done &amp;gt;&amp;gt; /tmp/delete_stu_groups&lt;/span&gt;&lt;/tt&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;br style=&quot;background-color: white; color: #333333; line-height: 20px;&quot; /&gt;&lt;tt style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span class=&quot;FIXEDFONT&quot;&gt;# for i in `cat /tmp/stu_list` ; do echo &quot;/usr/openv/netbackup/bin/admincmd/bpstudel -label $i&quot; ; done &amp;gt;&amp;gt; /tmp/delete_stus&lt;/span&gt;&lt;/tt&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;br style=&quot;background-color: white; color: #333333; line-height: 20px;&quot; /&gt;&lt;tt style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span class=&quot;FIXEDFONT&quot;&gt;# sh /tmp/delete_stu_groups&lt;/span&gt;&lt;/tt&gt;
&lt;/span&gt;&lt;br /&gt;
&lt;tt style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span class=&quot;FIXEDFONT&quot; style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/tt&gt;&lt;br /&gt;
&lt;tt style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span class=&quot;FIXEDFONT&quot; style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;# sh /tmp/delete_stus&lt;/span&gt;&lt;/tt&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;6) Delete all tape devices / libraries (this is actually easier to do from command line than from the GUI)&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;background-color: white; color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace; line-height: 20px;&quot;&gt;# /usr/openv/netbackup/bin/admincmd/nbemmcmd -deletealldevices -allrecords&lt;/span&gt;
&lt;br /&gt;
&lt;span style=&quot;background-color: white; color: #333333; line-height: 20px;&quot;&gt;&lt;span style=&quot;font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;6) Stop and restart NetBackup&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333;&quot;&gt;&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace; line-height: 20px;&quot;&gt;# /usr/openv/netbackup/bin/bp.kill_all&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333;&quot;&gt;&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace; line-height: 20px;&quot;&gt;# /usr/openv/netbackup/bin/bp.start_all&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;7) Disable the job scheduler&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333;&quot;&gt;&lt;span style=&quot;font-family: &#39;Courier New&#39;, Courier, monospace; line-height: 20px;&quot;&gt;# /usr/openv/netbackup/bin/admincmd/nbpemreq -suspend_scheduling&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;Remember- this command only lasts until NetBackup is shutdown and restarted again. If you restart the NetBackup services, before you are ready for backups to start again, you&#39;ll need to run this command again. To turn scheduling back on either bounce NetBackup, or run:&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace; line-height: 20px;&quot;&gt;# /usr/openv/netbackup/bin/admincmd/nbpemreq -resume_scheduling&lt;/span&gt;
&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif; line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif; line-height: 20px;&quot;&gt;You can get the same effect by disabling all backup policies, and not have to worry about the restart issue- but you may run into problems later- some database restores require that the associated policy is active, so this avoids having to remember to reactivate certain policies later on.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;8) Re-run the device configuration wizard (from the GUI) this can be done from the command line as well, but the GUI is easier to deal with. This is the same standard routine you used to configure your tape library before you recovered the catalog, so we will not elaborate further here.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;9) Move all tapes to &quot;non-robotic&quot;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;The catalog restore will have brought in definitions for all tapes in the home environment. Those tapes may/will still be associated with their production robotic libraries even though we&#39;ve deleted those libraries, so any jobs run will attempt to contact a non-existent library! In the Media pane of the GUI, highlight all tapes, right click, and select move. Click on the checkbox next to &quot;tapes are in a robotic library&quot; then click the checkbox again to uncheck it. Click on OK to continue. The operation may take several minutes, and may give you a timeout warning- this timeout is just a warning- the operation will still run in the background. Wait several minutes, then refresh the media pane, and you should see the tapes are no longer associated with any robotic libraries.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;10) Re-run robot inventory. Make sure all the tapes you need are loaded (you can still add more later, but it&#39;s easier if you get them all now), then rerun your robot inventory, this will associate all of the tapes currently in the robot with the correct robot, so NetBackup can find them.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;11) Run a test restore. You should be good to go, run a small test restore of something small such as the /var/adm/messages file on one of your servers to make sure everything is working properly.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;b&gt;&lt;i&gt;Question 5-&lt;/i&gt; I did everything you told me, but when I run restores, the jobs are requesting tapes that I don&#39;t have!&lt;/b&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;Verify that you are attempting to restore from backup jobs that match the range of dates covered by your tapes- you may be unintentionally asking to restore data from a backup that is older than the set of tapes you currently have on-site.&amp;nbsp;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;If you know your range of dates matches your tape set, you may actually have secondary copies of the original backup tapes - some production shops prefer to keep all of their archived tapes available on-site in case they need them quickly, and will clone copies of the tape set that gets sen off-site- in these cases, both sets of tapes are valid, but only one is defined as the &quot;primary copy&quot; - all jobs will always attempt to use the primary copy. You will need to promote your secondary copies to &quot;primary&quot; status before NetBackup will attempt to use them.&amp;nbsp;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;There are a few tricks to get around them - most of them are tricky, and I don&#39;t really consider any of them to be very clean. You can search the image database and selectively promote the images you have on your tapes to primary - doing this is somewhat convoluted and difficult to describe without showing you on a live system, and you can very easily miss some images and end up playing cleanup later.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;You can manually select which copy you want to pull a restore from, but only if you know the copy number you need, and you are running your restore specifically from the command line without the GUI to help. Since it&#39;s generally impractical to run restores from the command line most of the time, this is not a good solution.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;You can globally set the primary copy number- but this will only work if *all* the restores you need come from the same set of copies. With this fix in place, you will not be able to access restores that come from any other set, even the primary copies, until you change the setting or return it to default. You also need to know which set of copies you need (usually it will be copy 2). To use copy two as primary by this method do:&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;# echo 2 &amp;gt; /usr/openv/netbackup/ALT_RESTORE_COPY_NUMBER&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;Replace 2 with the copy number you need. To revert back to default, simply change the name of this file to something else or delete it.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;You may also use the bpchangeprimary command to promote a copy from the command line- this is probably the easiest method for a beginner, since you don&#39;t need to figure out what to click on or highlight in &amp;nbsp;order to do it. All you need to know is the client name, the copy number you need to be using (probably 2 if you have to guess), and a &quot;start date&quot; (the command will apply the change to all images after the date you specify as the start date). For example:&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;# bpchangeprimary -copy 2 -cl mynbuclient -sd 11/24/2011&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;In english, this says to use mark copy 2 as primary for all images of the client named &quot;mynbuclient&quot; taken after november 24, 2011. You&#39;ll need to run this command for every client you need to restore from an alternate copy, but this can easily be scripted if you can put a list of all your affected clients in a text file called something like &quot;/tmp/clients&quot; -&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;# for x in `cat /tmp/clients` ; do&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&amp;gt; echo &quot;Working on ${x}&quot;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&amp;gt; bpchangeprimary -copy 2 -cl ${x} -sd 11/24/2011&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Courier New&#39;, Courier, monospace;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&amp;gt; done&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;Be warned, this may take a while to run! For more information on the bpchangeprimary command, see the &lt;a href=&quot;http://www.symantec.com/business/support/index?page=content&amp;amp;id=HOWTO43646&quot;&gt;manual page listing&lt;/a&gt; on symantec&#39;s official site.&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;color: #333333; font-family: &#39;Trebuchet MS&#39;, sans-serif;&quot;&gt;&lt;span style=&quot;line-height: 20px;&quot;&gt;&lt;br /&gt;&lt;/span&gt;&lt;/span&gt;</content><link rel='replies' type='application/atom+xml' href='http://var-adm.blogspot.com/feeds/7310207303300285865/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://var-adm.blogspot.com/2011/12/5-netbackup-dr-tips-on-solaris.html#comment-form' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/7310207303300285865'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/1407002244064487228/posts/default/7310207303300285865'/><link rel='alternate' type='text/html' href='http://var-adm.blogspot.com/2011/12/5-netbackup-dr-tips-on-solaris.html' title='5 NetBackup DR tips on Solaris'/><author><name>Anonymous</name><uri>http://www.blogger.com/profile/00970993989722934266</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='https://img1.blogblog.com/img/b16-rounded.gif'/></author><media:thumbnail xmlns:media="http://search.yahoo.com/mrss/" url="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEincfvdx1Q5bQNgT6KRohUFABETA1UQlVIDqGffZ1NXfEpt6iJB9t_Rjw0Qrnc9ECuN4gh58ytqN7L4XdM_xnGdwGO81jbfP-rstkmP0NTHs474M5mwpeueH21XAOdj9J0EXBe73_3cdTo/s72-c/NetBackup_GUI_jnbSA.jpg" height="72" width="72"/><thr:total>0</thr:total></entry></feed>