<?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-6031958966034419131</id><updated>2026-11-09T14:31:03.514-08:00</updated><category term="FreeBSD Thinkpad"/><category term="xorg mouse habu debian"/><title type='text'>John&#39;s Tech Notes</title><subtitle type='html'>Notes on things I have done so that I might remember them later.</subtitle><link rel='http://schemas.google.com/g/2005#feed' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/posts/default'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default?redirect=false'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/'/><link rel='hub' href='http://pubsubhubbub.appspot.com/'/><link rel='next' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default?start-index=26&amp;max-results=25&amp;redirect=false'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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>51</openSearch:totalResults><openSearch:startIndex>1</openSearch:startIndex><openSearch:itemsPerPage>25</openSearch:itemsPerPage><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-4464995903148107482</id><published>2026-05-01T09:57:00.000-07:00</published><updated>2026-05-01T09:57:23.932-07:00</updated><title type='text'>zfs, docker, systemd startup</title><content type='html'>Various places on the internet suggest putting docker filesystem things for configurations in /home. On my setup I have /home setup as a bind mount into my zfs volume.

Whenever I rebooted my system, most of my docker containers are missing their configs and I have to restart them to get them to work.

First attempt, use the &quot;wait-for-zfs&quot; systemd trick in /etc/systemd/system/docker-wait-for-zfs.service (and run systemctl enable docker-wait-for-zfs.service):

&lt;pre&gt;
[Unit]
Description=Wait for ZFS before starting Docker
Before=docker.service
After=zfs.target
Requires=zfs.target

[Service]
Type=oneshot
ExecStart=/bin/true
RemainAfterExit=yes

[Install]
RequiredBy=docker.service
WantedBy=multi-user.target
&lt;/pre&gt;

That helped docker start later, but still didn&#39;t work.

Looking through &lt;pre&gt;sudo journalctl -b&lt;/pre&gt; I noticed some logs about mounting /home.

So I added home.mount to the After list. The docker jobs start correctly, but not until after I login for the first time.

Tweaking /etc/fstab to add x-systemd.automount and tweak my zfs dependencies to zfs.target rather than zfs-import.target, this seems to work.
&lt;pre&gt;
/tank/home /home none bind,defaults,nofail,x-systemd.after=zfs.target,x-systemd.requires=/tank/home,x-systemd.automount 0 0
&lt;/pre&gt;

Trying to add mount dependencies in docker-wait-for-zfs seemed to make docker not start well. But maybe it wasn&#39;t shaped right.
</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/4464995903148107482/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/4464995903148107482' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/4464995903148107482'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/4464995903148107482'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2026/05/zfs-docker-systemd-startup.html' title='zfs, docker, systemd startup'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-5881255990298231801</id><published>2026-03-14T08:44:00.000-07:00</published><updated>2026-03-14T08:44:49.149-07:00</updated><title type='text'>Trying Xfinity xb10</title><content type='html'>&lt;p&gt;
Opnsense with WAN Groups on a Protectcli FW4C
&lt;/p&gt;

&lt;p&gt;
Google Speed Test
&lt;ul&gt;
&lt;li&gt;Centurylink Gigabit &lt;-&gt; 942/923&lt;/li&gt;
&lt;li&gt;Comcast Regular (xb8-t) &lt;-&gt; 919/115&lt;/li&gt;
&lt;li&gt;Comcast FTTaH (xb10) &lt;-&gt; 910/904&lt;/li&gt;
&lt;/ul&gt;
&lt;/p&gt;

&lt;p&gt;
  The FW4c wasn&#39;t able to manage 2Gbps service when I tried it but seems to be ok at 1G.
&lt;/p&gt;

&lt;p&gt;
  The xb10 10G port negotiated to the 2.5G of the FW4C.
&lt;/p&gt;

&lt;p&gt;
  Seems to be working as advertised.
&lt;/p&gt;
 </content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/5881255990298231801/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/5881255990298231801' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5881255990298231801'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5881255990298231801'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2026/03/trying-xfinity-xb10.html' title='Trying Xfinity xb10'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-5042194029314703800</id><published>2025-12-27T11:44:00.000-08:00</published><updated>2025-12-27T11:44:46.967-08:00</updated><title type='text'>Pbx adventures</title><content type='html'>Trying to run some voip devices on a vlan on the network with something running asterix on docker.

Seems like the modern way to go is system-networkd, but I don&#39;t feel like migrating just now.


&lt;pre&gt;
apt-get install bridge-utils

auto vlan2
iface vlan2 inet manual
        vlan-raw-device ens2

auto br2
iface br2 inet static
        bridge_ports vlan2
        bridge_stp on
        bridge_maxwait 2
        bridge_hw vlan2 # won&#39;t go UP o/w
        address 192.168.48.2
        netmask 255.255.255.0
&lt;/pre&gt;


Then for docker

&lt;pre&gt;
sudo docker network create -d ipvlan \
  --subnet=192.168.48.0/24 \
  --gateway=192.168.48.1 \
  --ip-range=192.168.48.128/25 \
  -o ipvlan_mode=l2 \
  -o parent=br2 ipvlan2
&lt;/pre&gt;



Then using mikopbx after creating a pbx user via the https://docs.mikopbx.com/mikopbx/english/setup/docker/docker-installation instructions (but using a different www user):

&lt;pre&gt;
sudo docker run \
  --network=ipvlan2 \
  --name mikopbk \
  --hostname mikopbx \
  -v /home/pbx/cf:/cf \
  -v /home/pbx/storage:/storage
  -e SSH_PORT=23 \
  -e ID_WWW_USER=&quot;$(id -u pbx)&quot; \
  -e ID_WWW_GROUP=$(id -g pbx)&quot; \
  -it -d --restart always \
  ghcr.io/mikopbx/mikopbx-x86-64
&lt;/pre&gt;

And then when I forget the admin password I set, &lt;pre&gt;sudo docker exec -it mikopbx /etc/rc/console_menu&lt;/pre&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/5042194029314703800/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/5042194029314703800' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5042194029314703800'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5042194029314703800'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2025/12/pbx-adventures.html' title='Pbx adventures'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-5048397587810029362</id><published>2025-12-27T11:31:00.000-08:00</published><updated>2025-12-27T11:31:56.302-08:00</updated><title type='text'>Trixie Upgrade</title><content type='html'>Grub reinstall failed, had to recover to get it working again, roughly following https://wiki.debian.org/GrubEFIReinstall with a USB

postgres is trying to start earlier that the zfs of the db, used 

&lt;pre&gt;
systemctl edit --full postegresql
&lt;/pre&gt;

to add 

&lt;pre&gt;
[Unit] 
After=zfs-mount.service
&lt;/pre&gt;

Somehow postgres had also selected Navajo as the timezone and it wasn&#39;t working.

NFSv4 export handling seems to have changed. Fixed specific /etc/exports


Some netatalk services I don&#39;t use are sad, so disabling.

&lt;pre&gt;
systemctl mask timelord
systemctl mask a2boot
&lt;/pre&gt;

Apple said that they are deprecating appletalk for time machine in the future, so we have to see how to get that working on samba with the fruit modules.</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/5048397587810029362/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/5048397587810029362' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5048397587810029362'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5048397587810029362'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2025/12/trixie-upgrade.html' title='Trixie Upgrade'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-8164417420598492178</id><published>2024-12-02T10:39:00.004-08:00</published><updated>2024-12-02T10:39:32.929-08:00</updated><title type='text'>USB Hubs Everywhere</title><content type='html'>&lt;p&gt;
For work I had a decent setup connecting a couple monitors to a thunderbolt dock with ethernet, some usb ports, and then a usb hub connecting my random collection of keyboard, mouse, and accessories.

&lt;p&gt;
When I transitioned to a newer M3 Macbook, half of the things on the usb hub stopped working.

&lt;p&gt;
Messing with &lt;pre&gt;ioreg -p IOUSB&lt;/pre&gt;, the output is approximately the same. I had a hazy memory of there being a USB hub depth limit and I saw that some of the devices that were failing had internal USB hubs. The output of &lt;pre&gt;ioreg&lt;/pre&gt; didn&#39;t show a depth difference, but the Intel to ARM platform change may have changed something.

&lt;p&gt;
Within the TBT200 dock, the USB2.0 ports are on a hub are nested under the USB3.0 hub. Moving my accessory external USB hub to a USB-C to USB converter moved it up the hierarchy and my attached devices started working again.

&lt;pre&gt;
+-o Root 
  +-o AppleT8122USBXHCI
  +-o AppleT8122USBXHCI
  | +-o USB2.0 Hub
  | | +-o ITE BillBoard
  | | +-o TBT200 Hub1
  | | | +-o TBT200 Hub2
  | | |   +-o (PREVIOUS USB2.0 Hub Position) 
  | | | +-o TBT200 Dock
  | | | +-o TBT200 Dock
  | | | +-o hg_0005
  | | | +-o USB2.0 Hub (New Position)
  | | |   +-o USB2.0 Hub
  | | |   | +-o Logitech Webcam
  | | |   | +-o Headphones
  | | |   +-o USB 2.0 Hub
  | | |   | +-o USB-Keyboard
  | | |   +-o mouse
  | | +-o BillBoard Device
  | +-o USB3.0 Hub
  |   +-o TBT200 Hub1
  |     +-o TBT200 Dock Ethernet
  |     +-o TBT200 Hub2
  |     +-o USB3.0 Hub
  |       +-o USB3.0 Hub
  +-o AppleT8122USBXHCI
  &lt;/pre&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/8164417420598492178/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/8164417420598492178' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/8164417420598492178'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/8164417420598492178'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2024/12/usb-hubs-everywhere.html' title='USB Hubs Everywhere'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-1927530449142566415</id><published>2024-06-15T13:50:00.000-07:00</published><updated>2024-06-15T13:50:48.273-07:00</updated><title type='text'>(opn|pf)sense and bridged vlans</title><content type='html'>After testing regular bridges went ok, I tried to setup the vlans I like to use. I originally thought it was failing because of the anti-spoofing rules on the interfaces (digging through pfctl -s rules -v). Tweaking the &lt;a href=&quot;https://docs.netgate.com/pfsense/en/latest/bridges/firewall.html#figure-bridging-filter-tunables&quot;&gt;bridge filtering tunables&lt;/a&gt; didn&#39;t seem to work.

&lt;p&gt;
Mucking with tcpdump -n -n -e vlan -i igc1, I noticed that incoming vlan packet replies came in on the interface but never the respective bridge. An &lt;a href=&quot;https://forum.opnsense.org/index.php?topic=22660.msg107966#msg107966&quot;&gt;opnsense thread&lt;/a&gt; points to a pfsense/freebsd issue that you &lt;a href=&quot;https://redmine.pfsense.org/issues/11139&quot;&gt;can&#39;t mix untagged and tagged vlans with bridges&lt;/a&gt;.

</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/1927530449142566415/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/1927530449142566415' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/1927530449142566415'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/1927530449142566415'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2024/06/opnpfsense-and-bridged-vlans.html' title='(opn|pf)sense and bridged vlans'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-7791467442888662168</id><published>2024-04-13T14:30:00.000-07:00</published><updated>2024-04-13T14:30:22.644-07:00</updated><title type='text'>Benchmarking 2.5gbe</title><content type='html'>Lots of the forums give a lot of gruff on pfsense and bridged interfaces. I was curious to see how it does.
&lt;br/&gt;
I have a Protectli Vault FW4C Intel J3710 which has 4 2.5gbe interfaces and one WAN port.
&lt;br/&gt;
Running iperf3 on my NAS server with a I225-LM PCI card in it.
&lt;br/&gt;
On the other end I have either a TBT200 corsair dock w/2.5gbe that has an rtl8152 chipset. Or an Asus USB3 adapter with the same rtl8152. Both have similar performance.
&lt;br/&gt;
It looks like the per-packet overhead is high on the bridge. 

&lt;table&gt;
  &lt;tr&gt;&lt;th&gt;Server&lt;/th&gt;&lt;th&gt;intermediate&lt;/th&gt;&lt;th&gt;Client&lt;/th&gt;&lt;th&gt;MTU&lt;/th&gt;&lt;th&gt;Throughput&lt;/th&gt;&lt;/tr&gt;
  &lt;tr&gt;&lt;td&gt;I225-LM (pci)&lt;/td&gt;&lt;td&gt;passthrough&lt;/td&gt;&lt;td&gt;rtl8152 (usb-c)&lt;/td&gt;&lt;td&gt;1500&lt;/td&gt;&lt;td&gt;2.35gbps&lt;/td&gt;&lt;/tr&gt;
  &lt;tr&gt;&lt;td&gt;&lt;/td&gt;&lt;td&gt;passthrough&lt;/td&gt;&lt;td&gt;&lt;/td&gt;&lt;td&gt;9000&lt;/td&gt;&lt;td&gt;2.47gbps&lt;/td&gt;&lt;/tr&gt;
  &lt;tr&gt;&lt;td&gt;I225-LM (pci)&lt;/td&gt;&lt;td&gt;fw4c bridge (~30% CPU)&lt;/td&gt;&lt;td&gt;rtl8152 (usb-c)&lt;/td&gt;&lt;td&gt;1500&lt;/td&gt;&lt;td&gt;1.2gbps&lt;/td&gt;&lt;/tr&gt;
  &lt;tr&gt;&lt;td&gt;&lt;/td&gt;&lt;td&gt;fw4c bridge (~14% CPU)&lt;/td&gt;&lt;td&gt;&lt;/td&gt;&lt;td&gt;9000&lt;/td&gt;&lt;td&gt;2.47gbps&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;

Setting MTUs with &lt;pre&gt;ip link set dev eth0 mtu 9000&lt;/pre&gt;

Now, can my NAS actually serve 2.5gbps? Maaaybe. 

&lt;pre&gt;dd if=a_60gb_file bs=1M of=/dev/null status=progress&lt;/pre&gt; from of a zfs ssd mirror hit 800MBps (6.4gbps) for read speed. SCP over the network ran around 230-260MBps (2.5gbps is 310MBps).
&lt;br/&gt;
The HDD mirror pair runs closer to 115-140MBps which is a marginally higher than 1Gbps can reach (125MBps).
</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/7791467442888662168/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/7791467442888662168' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/7791467442888662168'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/7791467442888662168'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2024/04/benchmarking-25gbe.html' title='Benchmarking 2.5gbe'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-2575759515836579951</id><published>2024-04-12T22:19:00.000-07:00</published><updated>2024-04-12T22:19:11.260-07:00</updated><title type='text'>pfsense bridges and vlans</title><content type='html'>Bridging on pfsense is a little funky. There is a lot of anti-bridge sentiment on the forums which doesn&#39;t translate well to answers.

If you set a gateway on the lan port it applies the wan rules to it. Initially I though this was the issue when I was testing but that appears to be a red-herring.

On at least v2.7, if you create a bridge and assign it an ip address, there is a firewall rule that is added that drops traffic to that subnet that isn&#39;t destined to the bridge

&lt;pre&gt;$ pfctl -v -s rules | grep 192.168.0.0 -A 2
block drop in log on ! bridge0 inet from 192.168.0.0/24 to any ridentifier 1000002520
  [ Evaluations: 5784      Packets: 2872      Bytes: 294921      States: 0     ]
  [ Inserted: uid 0 pid 44267 State Creations: 0     ]
&lt;/pre&gt;

Poking around in /var/log/filter.log I found plent of entries that suggest the rule is being evaluated at the interface rather than the bridge.

&lt;pre&gt;...63,,,1000002520,igc1,match,block,in,4,0x0,,64,53798,0,DF,1,icmp,84,192.168.0.1,192.168.0.7,request,17734,177264&lt;/pre&gt;


&lt;div&gt;I found the &lt;a href=&quot;https://docs.netgate.com/pfsense/en/latest/bridges/firewall.html#bridging-and-firewalling&quot;&gt;pfsense documentation&lt;/a&gt; on the sysctl that influences how the rules are evaluated. Setting net.link.bridge.pfil_bridge to 1 and net.link.bridge.pfil_member to 0 seems to have remedied the immediate issue of not begin able to ping across the bridge based on the generated firewall rules.&lt;/div&gt;
&lt;br/&gt;
&lt;div&gt;Setting up vlans on the bridges is more involved than other places. You can build the untagged bridge based the OPT* interfaces. To make a bridge you can set up the tagged vlan for each interface, then mark it as enabled, then assign it to an OPT*, then go make a bridge out of the OPT*. If you set the description that carries around which is better than having the OPT mapping on a sticky note.
&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/2575759515836579951/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/2575759515836579951' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/2575759515836579951'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/2575759515836579951'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2024/04/pfsense-bridges-and-vlans.html' title='pfsense bridges and vlans'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-7545470445138544864</id><published>2022-11-23T13:47:00.000-08:00</published><updated>2022-11-23T13:47:59.917-08:00</updated><title type='text'>Brother ADS2200 Linux</title><content type='html'>Replacing my ADS2000 with a ADS2200 I see it comes with drivers last updated in 2017 &lt;a href=&quot;https://support.brother.com/g/b/downloadend.aspx?c=us&amp;amp;lang=en&amp;amp;prod=ads2200_all&amp;amp;os=128&amp;amp;dlid=dlf103479_000&amp;amp;flang=4&amp;amp;type3=566&quot;&gt;on the Brother Website&lt;/a&gt;.

However, when running &lt;blockquote&gt;scanimage -L&lt;/blockquote&gt; I didn&#39;t get an output.

Looking at the driver file I see we are missing some dependencies

&lt;pre&gt;% ldd /usr/lib/x86_64-linux-gnu/sane/libsane-BrotherADS2200.so
        linux-vdso.so.1 (0x00007ffd981fb000)
        libusb-1.0.so.0 =&gt; /usr/lib/x86_64-linux-gnu/libusb-1.0.so.0 (0x00007f4bfc14b000)
        libjpeg.so.8 =&gt; not found
        libstdc++.so.6 =&gt; /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f4bfbf7e000)
        libgcc_s.so.1 =&gt; /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f4bfbf64000)
        libc.so.6 =&gt; /lib/x86_64-linux-gnu/libc.so.6 (0x00007f4bfbd8f000)
        libudev.so.1 =&gt; /usr/lib/x86_64-linux-gnu/libudev.so.1 (0x00007f4bfbd67000)
        libpthread.so.0 =&gt; /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f4bfbd43000)
        /lib64/ld-linux-x86-64.so.2 (0x00007f4bfc43c000)
        libm.so.6 =&gt; /lib/x86_64-linux-gnu/libm.so.6 (0x00007f4bfbbff000)
&lt;/pre&gt;

&lt;a href=&quot;https://unix.stackexchange.com/questions/687633/why-is-libjpeg-so-8-missing-from-debian&quot;&gt;Debian doesn&#39;t come with libjpeg8&lt;/a&gt; anymore, but you can pull from &lt;a href=&quot;http://snapshot.debian.org/package/libjpeg8/8d1-2/&quot;&gt;an older snapshot&lt;/a&gt;
</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/7545470445138544864/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/7545470445138544864' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/7545470445138544864'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/7545470445138544864'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2022/11/brother-ads2200-linux.html' title='Brother ADS2200 Linux'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-162427125206225923</id><published>2016-04-06T06:56:00.000-07:00</published><updated>2016-05-14T08:58:57.306-07:00</updated><title type='text'>Wiping out zfs snapshots</title><content type='html'>My zfs volume filled up the other day, in part because I had zfs-auto-snapshot going and it picked up some rather large files that I move around one of the filesystems.

You can list all snapshot space usage with:
&lt;pre style=&quot;font-family: fixed&quot;&gt;&lt;code style=&quot;color:black;word-wrap:normal;&quot;&gt;
zfs list -t snapshot -o space
&lt;/code&gt;&lt;/pre&gt;

Then filtering the list of snapshots down to the fs in question:
&lt;pre style=&quot;font-family: fixed&quot;&gt;&lt;code style=&quot;color:black;word-wrap:normal;&quot;&gt;
zfs list -t snapshot -o space | grep scratch | awk &#39;{print $1}&#39; | xargs -n 1 zfs destroy
&lt;/code&gt;&lt;/pre&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/162427125206225923/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/162427125206225923' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/162427125206225923'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/162427125206225923'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2016/04/wiping-out-zfs-snapshots.html' title='Wiping out zfs snapshots'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-9125249113051836372</id><published>2015-05-02T12:03:00.004-07:00</published><updated>2015-05-02T12:03:49.163-07:00</updated><title type='text'>Disable lid-close-suspend on debian/systemd.</title><content type='html'>I found the basis &lt;a href=&quot;https://ask.fedoraproject.org/en/question/27808/preventing-lid-close-suspension/&quot;&gt;this on the fedora site&lt;/a&gt; and am replicating it here because I have never configured systemd before.

Edit &lt;pre&gt;/etc/systemd/logind.conf&lt;/pre&gt;
Uncomment &lt;i&gt;HandleLidSwitch&lt;/i&gt; and set it to &quot;&lt;i&gt;ignore&lt;/i&gt;&quot;  (more details in man logind.conf).
Then run &lt;pre&gt;systemctl restart systemd-logind&lt;/pre&gt;

My systems certainly boot faster, but I do remain skeptical of all the magic contained within systemd.</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/9125249113051836372/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/9125249113051836372' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/9125249113051836372'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/9125249113051836372'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2015/05/disable-lid-close-suspend-on.html' title='Disable lid-close-suspend on debian/systemd.'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-1134874256633977313</id><published>2015-01-31T11:59:00.002-08:00</published><updated>2015-01-31T11:59:25.510-08:00</updated><title type='text'>More on low power system bandwidth.</title><content type='html'>I picked up a &lt;a href=&quot;http://ameridroid.com/products/odroid-c1&quot;&gt;ODroid-C1&lt;/a&gt; for fun, in part because it has a gigabit ethernet adapter &lt;a href=&quot;http://hardkernel.com/main/products/prdt_info.php?g_code=G141578608433&amp;tab_idx=2&quot;&gt;as part of the SoC&lt;/a&gt;.  In theory you can usb-attach some storage to and call it a server.

I have a desktop with an Intel GbE NIC that I am testing against.  Making a table from including my earlier numbers for systems I can no longer access:
&lt;table&gt;
&lt;tr&gt;&lt;th&gt;Thingy&lt;/th&gt;&lt;th&gt;approximate iperf throughput&lt;/th&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td&gt;2010 Macbook Pro&lt;/td&gt;&lt;td&gt;928Mbps&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td&gt;Odroid-C1&lt;/td&gt;&lt;td&gt;920Mbps (tx 400Mbps)&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td&gt;Atom330 w/Intel GbE&lt;/td&gt;&lt;td&gt;670Mbps&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td&gt;Atom330 w/onboard GbE&lt;/td&gt;&lt;td&gt;580Mbps&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td&gt;AtomN450 w/USB2.0 GbE&lt;/td&gt;&lt;td&gt;240Mbps (tx 220Mbps)&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td&gt;TP-LINK TL-WDR4300&lt;/td&gt;&lt;td&gt;205Mbps&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td&gt;WZR-HP-G300NH&lt;/td&gt;&lt;td&gt;200Mbps&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;

I&#39;ve noticed that the Odroid and others appear to be transmission limited.  I spent a little while tweaking tcp algorithms and parameters without any significant change.  My current guess is driver/irq behavior.  I&#39;m not going down that road today.

Running a loopback iperf test (iperf -c 127.0.0.1 -r) gets ~19Gbps on the MBP, 64Gbps on my i7-3770k, 3.4Gbps on the N450 atom, and 4.4Gbps on the Odroid.

scp to /dev/null the odroid hits around 23.3MBps, ~11MBps for the atom.</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/1134874256633977313/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/1134874256633977313' title='2 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/1134874256633977313'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/1134874256633977313'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2015/01/more-on-low-power-system-bandwidth.html' title='More on low power system bandwidth.'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>2</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-6552441570042777386</id><published>2014-11-08T10:38:00.000-08:00</published><updated>2014-11-08T10:38:52.942-08:00</updated><title type='text'>R.A.T.7 Mouse + Linux = Focus challenge.</title><content type='html'>My ~7 year old SSD drive ate it and I lost my system configs. &amp;nbsp;After re-building, I found that I was having focus problems within my current window manager (&lt;a href=&quot;http://enlightenment.org/&quot; target=&quot;_blank&quot;&gt;e17&lt;/a&gt;). &amp;nbsp;Because I end up building it from scratch because of Debian inertia, I blamed the custom build.&lt;br /&gt;
&lt;br /&gt;
Several months and much pain later, I&#39;ve discovered that it was actually the input configuration.&lt;br /&gt;
&lt;br /&gt;
Thanks to an &lt;a href=&quot;https://bbs.archlinux.org/viewtopic.php?id=183251&quot; target=&quot;_blank&quot;&gt;ArchLinux post&lt;/a&gt;, I now have the magic xorg.conf juju to make it work again:&lt;br /&gt;
&lt;br /&gt;
&lt;pre&gt;
Section &quot;InputClass&quot;
Identifier &quot;R.A.T.&quot;
MatchProduct &quot;R.A.T.7|R.A.T.9&quot;
MatchDevicePath &quot;/dev/input/event*&quot;
Option &quot;Buttons&quot; &quot;17&quot;
Option &quot;ButtonMapping&quot; &quot;1 2 3 4 5 0 0 8 9 7 6 12 0 0 0 16 17&quot;
Option &quot;AutoReleaseButtons&quot; &quot;13 14 15&quot;
Option &quot;ZAxisMapping&quot; &quot;4 5 6 7&quot;
EndSection
&lt;/pre&gt;

This reminds me that I do not miss setting modelines for CRT monitors back in the day.  Kids have it so easy these days.</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/6552441570042777386/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/6552441570042777386' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/6552441570042777386'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/6552441570042777386'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2014/11/rat7-mouse-linux-focus-challenge.html' title='R.A.T.7 Mouse + Linux = Focus challenge.'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-9137094201393815175</id><published>2014-05-11T16:03:00.000-07:00</published><updated>2014-05-11T16:03:41.367-07:00</updated><title type='text'>Speeding up file copies to an Atom server.</title><content type='html'>&lt;span style=&quot;font-family: inherit;&quot;&gt;I have a small Atom machine that I use for temporary file storage and upnp streaming and the like. &amp;nbsp;It has a lot of things going against it in terms of cpu speed and network connectivity, but it sips power much more so than anything else I have and can serve to cache backups before uploading them somewhere else.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;&lt;br /&gt;&lt;/span&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Copying files has been pretty slow. &amp;nbsp;The builtin NIC happens to be a 100Mbps device. &amp;nbsp;I attached a USB2.0 gigabit adapter which should put it closer to the USB2.0 max throughput of 480Mbps. &amp;nbsp;However, out of the box you scp files at around ~5MBps (~40Mbps).&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;&lt;br /&gt;&lt;/span&gt;
&lt;h3&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Step 1: Relaxing security&lt;/span&gt;&lt;/h3&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Because most of my copies go over the local LAN, and I expect my local LAN to be relatively secure, I decided to use the fastes cipher rather than the strongest cipher. &amp;nbsp;Other parts of the Internet and self-testing suggest that arcfour is the fastest, albeit least secure cipher we have. This improves throughput around 4x to ~20MBps (~160Mbps).&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Using scp directly, this looks like &lt;/span&gt;&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;scp -c arcfour place:/something somewhereelse&lt;/span&gt;&lt;/div&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;If you don&#39;t like to remember what to type, or aren&#39;t calling ssh directly (e.g. rsync), then put it in your ~/.ssh/config:&lt;/span&gt;&lt;/div&gt;
&lt;blockquote class=&quot;tr_bq&quot;&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;Host the_host&lt;br /&gt;&amp;nbsp; Ciphers arcfour&lt;/span&gt;&lt;/blockquote&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;At this point, ssh is still pegging the CPU. &amp;nbsp;What else is to be done?&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;&lt;br /&gt;&lt;/span&gt;
&lt;h3&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Step 2: Recompiling Everything&lt;/span&gt;&lt;/h3&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Now to see if re-compiling ssl/ssh has any impact on performance.&lt;/span&gt;&lt;br /&gt;
&lt;h4&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Openssl (1.0.1e)&lt;/span&gt;&lt;/h4&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;apt-get source openssl&lt;/span&gt;&lt;br /&gt;
Modify the &lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;Configure&lt;/span&gt; script by changing the &lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;debian-amd64 &lt;/span&gt;entry to have &lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;-march=atom&lt;/span&gt;&lt;br /&gt;
Run &lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;dpkg-source --commit&lt;/span&gt; to make the build setup sane.&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;&lt;br /&gt;&lt;/span&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Execute&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: Courier New, Courier, monospace;&quot;&gt;dpkg-buildpackage&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;and wait patiently (you should see -march= in the build output).&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;&lt;br /&gt;&lt;/span&gt;
&lt;h4&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Openssh (6.0p)&lt;/span&gt;&lt;/h4&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;apt-get source openssh&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;add to CFLAGS in debian/rules&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;dpkg-buildpackage&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;&lt;br /&gt;&lt;/span&gt;
&lt;h4&gt;
Drumroll&lt;/h4&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;... no change. &amp;nbsp;This is somewhat expected given that the optimization of the libraries is often through inline assembly and compiler tuning isn&#39;t going to bring very much to the table for our little in-order execution chum.&lt;/span&gt;&lt;br /&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;&lt;br /&gt;&lt;/span&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Just for reference, this was on Debian 7 w/gcc 4.7.2 (Debian 4.7.2-5) with an Atom N450 @1.66GHz.&lt;/span&gt;&lt;br /&gt;
&lt;h3&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Step 3: Don&#39;t bother with Step 2.&lt;/span&gt;&lt;/h3&gt;
&lt;div&gt;
&lt;span style=&quot;font-family: inherit;&quot;&gt;Stick with the cipher suite tweaks.&lt;/span&gt;&lt;/div&gt;
</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/9137094201393815175/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/9137094201393815175' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/9137094201393815175'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/9137094201393815175'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2014/05/speeding-up-file-copies-to-atom-server.html' title='Speeding up file copies to an Atom server.'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-2826905944832295567</id><published>2012-08-06T16:08:00.002-07:00</published><updated>2012-08-06T16:08:38.014-07:00</updated><title type='text'>Throughput testing on low-power systems</title><content type='html'>I&#39;ve been thinking about trying to make a low-power storage server for a while. &amp;nbsp;When I look through particularly low power hardware such as the trim slice or beagle board, I get a little upset that many of them don&#39;t have gigabit network adapters. &amp;nbsp;Thinking that this might be because no one uses them that way, I keep wishing that they&#39;d get gigabit adapters. &amp;nbsp;The other possibility is that the processors can&#39;t drive gigabit. &amp;nbsp;I did some simple testing with hardware I have laying around to see what happened.&lt;br /&gt;
&lt;br /&gt;
To test things out I used iperf over a local network. &amp;nbsp;The hardware switch on my access point can do gigabit, but apparently the processor on the&amp;nbsp;&lt;a href=&quot;http://wiki.openwrt.org/toh/buffalo/wzr-hp-g300h&quot; target=&quot;_blank&quot;&gt;Buffalo WZR-HP-G300NH&lt;/a&gt;&amp;nbsp;can only manage&amp;nbsp;~&lt;b&gt;200Mbps&lt;/b&gt;. &lt;br /&gt;
&lt;br /&gt;
Dinking around with a&amp;nbsp;&lt;a href=&quot;http://ark.intel.com/products/35641/Intel-Atom-Processor-330-(1M-Cache-1_60-GHz-533-MHz-FSB)&quot; target=&quot;_blank&quot;&gt;Dual Core Atom 330 &amp;amp; 1.66GHz&lt;/a&gt;&amp;nbsp;I had on hand, I got about &lt;b&gt;580Mbps&lt;/b&gt; with the embedded gigabit nice, and with a Intel Pro/1000 card I got an extra 100Mbps at around &lt;b&gt;670Mbps&lt;/b&gt; total. &lt;br /&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
USB2.0 is rated at 480Mbps, which translates to 60MBps. &amp;nbsp;A full gigabit is 125MBps. &amp;nbsp;&lt;a href=&quot;http://www.tomshardware.com/charts/hdd-charts-2012/-01-Read-Throughput-Average-h2benchw-3.16,2901.html&quot; target=&quot;_blank&quot;&gt;Typical hard drive speed ratings&lt;/a&gt; for sequential throughput are between 80MBps and 160MBps. &amp;nbsp;&lt;a href=&quot;http://www.tomshardware.com/charts/2012-mobile-hdd-charts/-01-Read-Throughput-Average-h2benchw-3.16,2988.html&quot; target=&quot;_blank&quot;&gt;Mobile hard drive speed ratings&lt;/a&gt; are between 60MBps and 100MBps. &lt;a href=&quot;http://www.tomshardware.com/charts/ssd-charts-2011/AS-SSD-Sequential-Read,2782.html&quot; target=&quot;_blank&quot;&gt;&amp;nbsp;SSDs are way past&lt;/a&gt; any of those, which means you probably want 10gigE, but then you wouldn&#39;t be paying so much attention to power if you can afford the equipment. &amp;nbsp;&lt;/div&gt;
&lt;div&gt;
&lt;br /&gt;&lt;/div&gt;
&lt;div&gt;
So, if you want to drive one or more hard drives at sequential throughput, you&#39;re going to want something beefier than an Atom. &amp;nbsp;I&#39;ve seen some NASes rated around 40W. &amp;nbsp;My desktop full of hard drives pulls ~60W at idle. &amp;nbsp;Assuming a power cost of $0.18/kWhr, it costs $94/year to run the desktop for a year (in reality it is off most of the time). &amp;nbsp;If a lower power system costs $500 to $1000 and it consumes 40W, then it would break even in 16 to 32 years.&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/2826905944832295567/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/2826905944832295567' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/2826905944832295567'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/2826905944832295567'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2012/08/throughput-testing-on-low-power-systems.html' title='Throughput testing on low-power systems'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-5220036532123813537</id><published>2012-05-18T10:41:00.002-07:00</published><updated>2012-05-18T10:41:27.645-07:00</updated><title type='text'>Vim tweaks for filename completion.</title><content type='html'>I have yet to delve deeply into the configuration of vim. &amp;nbsp;Someday I will, but I figured out something interesting today. &amp;nbsp;I&#39;ve been writing a lot of LaTeX lately, and I figured out how to make some tweaks to my auto-complete configuration (because it usually picks the wrong file on the first try).&lt;br /&gt;
&lt;br /&gt;
To make changes specific to the &lt;i&gt;tex&lt;/i&gt; filetype, without being overridden, you can make a file in &lt;i&gt;~/.vim/after/ftplugin/tex.vim (&lt;/i&gt;&lt;a href=&quot;http://vimdoc.sourceforge.net/htmldoc/filetype.html&quot;&gt;ref here&lt;/a&gt;) with content that looks something like:&lt;br /&gt;
&lt;blockquote class=&quot;tr_bq&quot;&gt;
if has(&quot;wildmenu&quot;)&lt;br /&gt;&amp;nbsp; &amp;nbsp; set wildignore+=*.fls,*.lof,*.lot,*.pdf,*.aux&lt;br /&gt;&amp;nbsp; &amp;nbsp; set wildmenu&lt;br /&gt;endif&lt;br /&gt;set spell&lt;/blockquote&gt;
I found a nice &lt;a href=&quot;http://blog.sanctum.geek.nz/vim-filename-completion/&quot;&gt;wildmenu description&lt;/a&gt; on another part of the internet.</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/5220036532123813537/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/5220036532123813537' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5220036532123813537'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5220036532123813537'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2012/05/vim-tweaks-for-filename-completion.html' title='Vim tweaks for filename completion.'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-5049474740828055633</id><published>2012-05-08T15:44:00.000-07:00</published><updated>2012-05-08T15:44:05.756-07:00</updated><title type='text'>XFonts</title><content type='html'>During the dissertation procrastination period, I made my XTerm fonts more beautiful. &amp;nbsp;I think the major challenge is that for LCDs you want sub-pixel rendering and not anti-aliasing (playing computer games means you keep turning up the anti-aliasing number more and more). &amp;nbsp;I found some neat documentation on the &lt;a href=&quot;https://wiki.archlinux.org/index.php/Font_Configuration#Subpixel_rendering&quot;&gt;xfont configuration&lt;/a&gt;&amp;nbsp;which indicates that sub-pixel rendering and anti-aliasing are incompatible. &amp;nbsp;Thus, all the examples/forums that suggest that you turn on all the knobs mean that your display looks funny. &amp;nbsp;(With anti-aliasing on with the monitors I&#39;m using right now I get weird green/red borders on some letters). &amp;nbsp;Font arguments aside, my .XResources file now looks like:&lt;br /&gt;
&lt;br /&gt;
&lt;blockquote class=&quot;tr_bq&quot;&gt;
&amp;nbsp; xterm*faceName: xft:DejaVu Sans Mono&lt;br /&gt;
&amp;nbsp; xterm*faceSize: 12pt&lt;br /&gt;
&amp;nbsp; xterm*metaSendsEscape: true&lt;br /&gt;
&amp;nbsp; xterm*altSendsEscape: true&lt;br /&gt;
&amp;nbsp; xterm*reverseVideo: &amp;nbsp; on&lt;br /&gt;
&amp;nbsp; xterm*utf8: 2&lt;br /&gt;
&amp;nbsp; Xft.dpi: 100&lt;br /&gt;
&amp;nbsp; Xft.lcdfilter: lcddefault&lt;br /&gt;
&amp;nbsp; Xft.rgba: rgb&lt;/blockquote&gt;
&lt;div&gt;
Note the lack of any antialias/hinting/autohint settings there.&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/5049474740828055633/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/5049474740828055633' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5049474740828055633'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5049474740828055633'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2012/05/xfonts.html' title='XFonts'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-3575163298445953049</id><published>2011-05-26T14:33:00.000-07:00</published><updated>2011-05-26T14:35:03.398-07:00</updated><title type='text'>Bulk copy</title><content type='html'>Copying one disk to many disks:&lt;br /&gt;&lt;pre&gt;&lt;br /&gt;dd if=/dev/sda bs=1M | tee &gt;(dd of=/dev/sdb bs=1M) &gt;(dd of=/dev/sdc bs=1M) &gt; /dev/null&lt;br /&gt;&lt;/pre&gt;&lt;br /&gt;Duplicating the bs argument is important for throughput.</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/3575163298445953049/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/3575163298445953049' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/3575163298445953049'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/3575163298445953049'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2011/05/bulk-copy.html' title='Bulk copy'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-7450920600487527482</id><published>2010-12-07T11:32:00.001-08:00</published><updated>2010-12-07T11:36:23.914-08:00</updated><title type='text'>Installing iWork 09 over iWork 09 Trial</title><content type='html'>I wanted to play the pdf (yay beamer)-&gt;keynote game to get a nice presentation display at the low low price of a student discounted iWork.  I had installed and updated the trial earlier, and when I installed the version from the DVD it would ask me for a serial number, which does not exist on any of the packaging.  The instructions below the serial number prompt said &quot;Install from DVD and follow the on-screen instructions&quot;.  This clearly hadn&#39;t worked.  Nor did just deleting /Applications/iWork 09.  You actually need to also delete /Library/Application Support/iWork 09. (I reinstalled at this point, I don&#39;t know if ordering is important).</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/7450920600487527482/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/7450920600487527482' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/7450920600487527482'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/7450920600487527482'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2010/12/installing-iwork-09-over-iwork-09-trial.html' title='Installing iWork 09 over iWork 09 Trial'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-9196412993742697906</id><published>2010-07-29T16:28:00.000-07:00</published><updated>2010-07-29T17:11:55.020-07:00</updated><title type='text'>Disable persistent device naming</title><content type='html'>&lt;div&gt;Persistent device naming is all fun and games until you start moving a boot drive between machines frequently.&lt;/div&gt;&lt;div&gt;&lt;br /&gt;&lt;/div&gt;&lt;div&gt;/etc/udev/rules.d/00-no-persistent-net.rules&lt;/div&gt;&lt;div&gt;&lt;code&gt;&lt;/code&gt;&lt;/div&gt;&lt;code&gt;&lt;div&gt;&lt;div&gt;# Disable net renaming&lt;/div&gt;&lt;div&gt;SUBSYSTEM==&quot;net&quot;, ACTION==&quot;add&quot;, DRIVERS==&quot;?*&quot;, NAME=&quot;$name&quot;&lt;/div&gt;&lt;/div&gt;&lt;/code&gt;&lt;div&gt;&lt;/div&gt;&lt;div&gt;&lt;br /&gt;&lt;/div&gt;&lt;div&gt;Then go and empty out the 70-persistent-net.rules&lt;/div&gt;&lt;div&gt;&lt;br /&gt;&lt;/div&gt;&lt;a href=&quot;http://linux.derkeiler.com/Mailing-Lists/Debian/2007-07/msg01606.html&quot;&gt;adapted from here&lt;/a&gt; (Modern udev doesn&#39;t like NAME=&quot;%k&quot;)</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/9196412993742697906/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/9196412993742697906' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/9196412993742697906'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/9196412993742697906'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2010/07/disable.html' title='Disable persistent device naming'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-4124861286317255607</id><published>2010-05-07T03:19:00.000-07:00</published><updated>2010-05-07T03:20:31.008-07:00</updated><title type='text'>multicast dns keeps getting installed on my machines...</title><content type='html'>/etc/apt/preferences.d/avahi-daemon-is-annoying:&lt;br /&gt;&lt;pre&gt;&lt;br /&gt;Package: avahi-daemon&lt;br /&gt;Pin: release a=blacklist-this-thing&lt;br /&gt;Pin-Priority: 1001&lt;br /&gt;&lt;/pre&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/4124861286317255607/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/4124861286317255607' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/4124861286317255607'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/4124861286317255607'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2010/05/multicast-dns-keeps-getting-installed.html' title='multicast dns keeps getting installed on my machines...'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-5668073108517069234</id><published>2010-03-08T13:11:00.000-08:00</published><updated>2010-03-08T13:12:31.547-08:00</updated><title type='text'>Down with flash cookies</title><content type='html'>Because clearing your browser history isn&#39;t enough.&lt;br /&gt;&lt;br /&gt;crontab entry:&lt;br /&gt;&lt;pre&gt;&lt;br /&gt;0 6 * * * find ~/.macromedia -name &#39;*.sol&#39; -delete&lt;br /&gt;&lt;/pre&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/5668073108517069234/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/5668073108517069234' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5668073108517069234'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/5668073108517069234'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2010/03/down-with-flash-cookies.html' title='Down with flash cookies'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-441842192864663611</id><published>2010-01-29T14:08:00.000-08:00</published><updated>2010-01-29T14:10:39.902-08:00</updated><title type='text'>Wiping out host keys for all likely instances.</title><content type='html'>For when an ssh host key changes and you don&#39;t want to run &lt;i&gt;ssh-keygen -R&lt;/i&gt; until it stops complaining.  I know it works on zsh, I believe it should be bash compatible.&lt;br /&gt;&lt;br /&gt;&lt;pre&gt;&lt;br /&gt;ssh-wipe ()&lt;br /&gt;{&lt;br /&gt;    addr=`host $1 | awk &#39;{print $4}&#39;`&lt;br /&gt;    names=`host $addr | awk &#39;{print substr($5, 0, -1)}&#39;`&lt;br /&gt;&lt;br /&gt;    for x in $1 $addr `host $addr | awk &#39;{print $5}&#39;`&lt;br /&gt;    do&lt;br /&gt;        echo ${x%%.}&lt;br /&gt;        ssh-keygen -R ${x%%.}&lt;br /&gt;    done&lt;br /&gt;}&lt;br /&gt;&lt;/pre&gt;</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/441842192864663611/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/441842192864663611' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/441842192864663611'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/441842192864663611'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2010/01/wiping-out-host-keys-for-all-likely.html' title='Wiping out host keys for all likely instances.'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-4543812561071375125</id><published>2010-01-13T13:37:00.000-08:00</published><updated>2010-01-28T15:20:20.707-08:00</updated><title type='text'>Xorg DRI with a Radeon HD 4650 on Debian</title><content type='html'>X was being slow on this fancy desktop, and I noticed a &quot;Direct Rendering Disabled&quot; message in the log.  Dmesg indicated a failure to load the firmware &quot;platform r600_cp.0: firmware: requesting radeon/RV730_cp.bin&quot;.  A simple&lt;br /&gt;&lt;pre&gt;&lt;br /&gt;apt-get install linux-firmware&lt;br /&gt;&lt;/pre&gt;&lt;br /&gt;corrected that problem, however, trying to run glxgears results in&lt;br /&gt;&lt;pre&gt;&lt;br /&gt;drmRadeonCmdBuffer: -22. Kernel failed to parse or rejected command stream. See dmesg for more info.&lt;br /&gt;&lt;/pre&gt;&lt;br /&gt;&lt;br /&gt;I noticed a few MTRR errors, and after adding &lt;i&gt;enable_mtrr_cleanup&lt;/i&gt; to &lt;i&gt;GRUB_CMDLIN_LINUX_DEFAULT&lt;/i&gt; in /etc/defaults/grub, rebooting, fixing the fact that a recent grub2 upgrade (boo grub2) had hosed itself (you should run &#39;grub-setup &quot;(hd0)&quot;&#39; before rebooting if you haven&#39;t for a while and don&#39;t like getting a rescue cd out).&lt;br /&gt;&lt;br /&gt;This fixed the MTRR errors, but the drmRadeonCmdBuffer issue remains, and is related to Mesa and the drm driver being out of sync (https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/419434).&lt;br /&gt;&lt;br /&gt;For the moment, X is faster, with no GL.&lt;br /&gt;&lt;br /&gt;[[Edit, works now w/drivers from testing (squeeze)]]</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/4543812561071375125/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/4543812561071375125' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/4543812561071375125'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/4543812561071375125'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2010/01/xorg-dri-with-radeon-hd-4650-on-debian.html' title='Xorg DRI with a Radeon HD 4650 on Debian'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-6031958966034419131.post-2360371386529814867</id><published>2009-12-13T16:41:00.000-08:00</published><updated>2011-05-26T14:38:35.683-07:00</updated><title type='text'>One-liner to extract vmlinuz to vmlinux</title><content type='html'>The hex string is the gzip offset.  I&#39;ve seen variations of this around, many using od.  (The od approach doesn&#39;t work when the magic string spans output lines).&lt;br /&gt;&lt;br /&gt;&lt;code&gt;&lt;br /&gt;x=/boot/vmlinuz-2.6.26-2-amd64 dd if=$x skip=`grep -a -b -o -m 1 -e $&#39;\x1f\x8b\x08\x00 $x | cut -d: -f 1` bs=1 | zcat &gt; /tmp/vmlinux&lt;br /&gt;&lt;/code&gt;&lt;br /&gt;&lt;br /&gt;(edited to fix \xf1 -&gt; \x1f).</content><link rel='replies' type='application/atom+xml' href='http://jcmdev0.blogspot.com/feeds/2360371386529814867/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment/fullpage/post/6031958966034419131/2360371386529814867' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/2360371386529814867'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/6031958966034419131/posts/default/2360371386529814867'/><link rel='alternate' type='text/html' href='http://jcmdev0.blogspot.com/2009/12/one-liner-to-extract-vmlinuz-to-vmlinux.html' title='One-liner to extract vmlinuz to vmlinux'/><author><name>jcmdev0</name><uri>http://www.blogger.com/profile/09764886910679742615</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><thr:total>0</thr:total></entry></feed>