Lethe's Tuxforge

Wireshark: Couldn't run /usr/bin/dumpcap

Lethe — Sun, 10 Jun 2012 16:29:11 +0000

This morning I tried to run Wireshark to capture some traffic for troubleshooting purposes on one of my lab machines and all I’ve got was the error Couldn’t run/usr/bin/dumpcap.

As this was a fresh installation of both Wireshark and Linux, I’ve upgraded to Fedora 17 only this morning, I never faced this error so I did as instructed in the error message and verified that my account was indeed part of the Wireshark group :

Issueing the command id told me that my account was indeed part of the wireshark group, next I tried to run the command as root and the error did not appear so what I did next was going into /usr/sbin/ directory looking for the dumpcap executable file, issuing a simple ls has shown the following :

As you can see in the image the Wireshark group is the owner of the file but the eXcute permission is missing from the file so to solve the Couldn’t run /usr/bin/dumpcap error simply issue the command :

 chmod +x /usr/sbin/dumpcap

And you will be able to run Wireshark as a standard user without the error Couldn’t run /usr/bin/dumpcap showing up.

I hope you will find the post useful and will take a second to leave me a comment or retweet it.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

Linux Smart Admin tips - Ssh passwordless authentication

Lethe — Thu, 08 Mar 2012 12:29:12 +0000

Ssh passwordless authentication – Why?

There are a variety of reasons why you would and should use ssh passwordless authentication, personally I tend to jump from one system to another quite frequently and probably you know better than me remembering a large number of password or even typing the same one over and over is not fun at all.

Apart from personal preference there is a small bit of added security in using ssh passwordless authentication, as the same implies not using a password means that one is not traveling through the wire (or the air if on a wireless connection) adding a layer of security

Last but bot least there are situation where you need one system to be able to login to a remote one without using any password, think of a cron job or a backup script that runs at night.

Got the idea? Like it I guess, so let’s see how to be a Linux smart admin and implement ssh passwordless authentication.

Ssh passwordless authentication- How to

The first step in implementing ssh passwordless authentication is generating a pair of private/public key that, as we will see, will be copied to the remote system for passwordless authentication via ssh, the command to create the keys is pretty straightforward, as the user for which you wish to enable ssh passwordless authentication type :

 ssh-keygen

You will be asked a few questions similar to the following :


Generating public/private rsa key pair.
Enter file in which to save the key (/home/lethe/.ssh/id_rsa):
Enter passphrase (empty for no passphrase):
Enter same passphrase again:
Your identification has been saved in /home/lethe/.ssh/id_rsa.
Your public key has been saved in /home/lethe/.ssh/id_rsa.pub.

For the sake of this example I’ve left the passphrase empty, you can choose to use one for further protection but keep in mind every time you will be connecting to the remote system you will be asked for it.

Ok the next, and final, step is that of copying the key to the system where we want to connect, there are few ways to accomplish this but for the sake of simplicity we will use the network to copy the key to a remote system like :

 ssh-copy-id -i /home/lethe/.ssh/id_rsa.pub lethe@system02

If this is the first time connecting to the remote system you will be asked to accept its RSA key, once accepted you will be asked for the remote user’s password and then the following message will be displayed to confirm everything worked as inteded :


Now try logging into the machine, with "ssh 'lethe@system02'", and check in:

.ssh/authorized_keys

to make sure we haven't added extra keys that you weren't expecting.

To test everything ssh to the remote machine and you will be immediately greeted by the motd without being asked for any password.

I hope you liked the article and will find it useful, if so take a second share it via the +1 button or Tweet it.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

Linux Smart Admin Tips - Tar extract to directory | tar extract to directory

Lethe — Wed, 07 Mar 2012 11:10:45 +0000

Starting with this post I commit myself to write every now and then posts that depict Linux “good” practices and habits, not overcomplicated posts about a specific argument but more a series of quick tips to show good Linux Admin habits.

Tar extract to directory – The long way

In your life as a Linux Admin or simple user you’ve found yourself in the situation when you need to extract the content of a tar file somewhere, what lot of people do is something like the following :

 cp somefile.tar.gz /tmp/

tar zxvf somefile.tar.gz

Ok while there is nothing technically wrong in the above commands it is… cumbersome!

Tar extract to directory – The short (and right) way

As I’ve said there is nothing wrong with the above method just there is a better and quicker way to do it which is the smart admin habit, so let’s see how to accomplish the same result as above with just with command :

 tar zxvf somefile.tar.gz -C /tmp/

There you go. What I did was tell tar to extract the file directly into the /tmp/ directory. See how easy and saving quite some typing?

Hope you will find this quick tip useful, don’t forget to subscribe to the blog as more will follow.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

Linux VLAN configuration

Lethe — Sun, 04 Mar 2012 18:30:49 +0000

Linux VLAN Configuration – The basics

VLAN is the short acronym for “Virtual LAN” which basically translates into creating multiple, separated LAN segments on the same physical switch without the need to use multiple network switches/cables to break up the network into smaller and more manageable entities. VLAN are created for a variety of reasons among which :

Network performance
Network security
Manageability

A complete and detailed description of VLAN is out of the scope of this post but if you are interested you can find a pretty explanatory article here.

Linux VLAN Configuration – HowTo

In today’s network using VLAN is pretty common and luckily enough usually you don’t have to worry about configuring Linux VLAN to tag traffic directly at the “source” but also if you need to configure Linux to support VLAN tagging take heart as the process is pretty easy and straightforward.

First of all you should gather information about your network, specifically you need to know the VLAN tag that Linux should add to outgoing traffic, for the sake of this post we will use interface eth0 and a VLAN with ID 550 so with this out of the way let’s start the Linux VLAN configuration.

Linux VLAN Configuration – Define VLAN interfaces

The first step into configuring Linux VLAN is editing the interface configuration file, in our example eth0, to resemble what depicted below :

 vi /etc/sysconfig/network-scripts/ifcfg-eth0

DEVICE="eth0"
HWADDR="00:19:B9:2F:6E:71"
NM_CONTROLLED="no"
ONBOOT="yes"

Once you have modified the eth0 configuration you need to setup the VLAN configuration file so that the Linux system will be able to send out tagged traffic so create a file named after the VLAN ID like this :

 vi  /etc/sysconfig/network-scripts/ifcfg-vlan450

VLAN="yes"
VLAN_NAME_TYPE="VLAN_PLUS_VID_NO_PAD"
DEVICE="vlan450"
PHYSDEV="eth0"
BOOTPROTO="static"
ONBOOT="yes"
TYPE="Ethernet"
IPADDR="10.10.10.100"
NETMASK="255.255.255.0"

Once you are done editing the file all you have to do is to restart network services to make the changes effective with a command like :

 service network restart

Linux VLAN Configuration – Alternative method

If your are not comfortable editing configuration file or simply feel like trying something different there is an alternate way to configure Linux VLAN interfaces, in this case make you have installed the package vconfig which does not come with the default installation :

 yum install vconfig

Once you have installed the package you can use the command vconfig to configure Linux VLAN let’s see how it works.

  vconfig add eth0 450

The above command will add the necessary configuration so that interface eth0 can send traffic tagged for VLAN 450, if you need to configure Linux VLAN for multiple networks you can use a command like the following :

 vconfig add eth0 10

vconfig add eth0 450

Now that you have configured your system to support VLAN you can assign an IP address to the interface created with the ifconfig command :

 ifconfig eth0.450 10.10.10.100 netmask 255.255.255.0 broadcast 10.10.10.255 up

You can view the information the configured Linux VLAN device with the following command :

 cat /proc/net/vlan/eth0.450

If you need to remove the VLAN configuration you can do so with vconfig command like in :

 vconfig rem eth0.450

Well that’s pretty much everything about the Linux VLAN configuration, I hope you did find the article useful and will take a second to retweet it or leave me a comment.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

RHCSA Exam - Tips and tricks from the field

Lethe — Fri, 02 Mar 2012 19:07:50 +0000

RHCSA Exam Experience – AKA What I did to pass the exam

As a twitter friend requested me to wrrite this post more than once I finally managed to put aside some spare time to write it hoping that it will help all people out there that are preparing for the RHCSA exam or plan to do it and are not sure how to approach it or how difficult it is.

First of all please keep in mind that I will be covering only RHCSA exam in this post as that is the one I’ve taken, one month ago, as RedHat in Switzerland does not offer the combined “2 exams in 1 day” package hence I’ve been forced to split them as I don’t have the time to travel to Italy or Germany to give both exams in the same day. I’ve booked my exam for next April once I’ve been through it (again) I will write about RHCE experience as well.

RHCSA Exam – Preparation tools

When I started writing this post the first question coming to my mind has been “Where do I start?”, well as I imagine a lot of you, like me, are impatient so I thought of starting out with the tools I used to preare for the exam. As with my last RHCE experience the main source of my preparation has beenMichael Jang’s book which I call the “unofficial RHCE certification bible”, as last time the book has proved itself a pleasure to read thanks to the author style of writing and to the fact that it stricly follows Red Hat RHCSA exam requirements checklist.

The above book has not been the only source of preparation but I’ve also used a number of documents from the official RHEL 6 which I always bring with me both for work and learning purposes, in the specific what has proved really useful have been the Deployment Guide and the SELinux guides, as I’ve ssaid the book iself is more than enough to get you prepared for the exam but I’m also convinced that the certification it’s just a nice piece of paper at the end of a learning path so I really like to pickup details that maybe I won’t need in the exam but will prove useful or an advantage in everyday’s work.

Consider the above documents as an “extra” to your preparation arsenal, specially the SELinux guide, they are not really needed but as I’ve said they are a nice bonus and specially reading the SELinux guide taught me a few nice tricks that I’ve already put at good use.

RHCSA Exam – Lab Setup

This is probably one of the questions that I am asked more often or read around the various forums, so what I did actually use to prepare my exam? It goes alone that RHCSA is based on, surprise surprise, Red Hat Enterprise Linux line of products so my suggestion is to use a RedHat based distribution as you host system (I personally use Fedora 16 as my desktop system but if you feel that’s a bit too much cutting edge for you CentOS makes a great and stable Desktop OS as well) for a very simple reason practice makes perfect and using a RH system on everyday’s life can teach you a few tricks useful in the exam as well, more on this in a second.

RHCSA Exam – Virtualization technology

Ok probably you know that virtualization is a big deal in today’s market but you should also be aware that it is now part of RedHat exams so you need to know the inside out of the default virtualization technology used for RedHat which is KVM so first step you take into the RHCSA preparation should be issuing the following command :

 yum groupinstall "Virtualization"

This will take care of installing all the needed tools to use and manage a virtual environment with KVM. Probably you will be wondering if you could use VMware products or Virtual Box and all in all there is no technical reason why you should not use one of the other available products the main reason why I prefer to use KVM is that it is what RedHat expects you to know and as part of the RHSA Syllabus you should be able to install, clone, configure virtual guests on KVM. You see the point? Free practice. Again if you feel more comfortable using another product feel free to do so, just be sure to know your stuff on KVM as it will probably appear in your RHCSA exam.

RHCSA Exam – How many machines do I need?

Short answer, 1 host machine and at least 1 guest virtual machine unless you want to mess around with your physical system. The slightly long answer is for the sake of RHCSA exam you don’t need a large of systems as the syllabus only require you to configure services and packages on the local system what I personally did was using a virtual machine as my test bed and a second virtual machine to test changes from outside the local host but again one machine is more than enough. As per the configuration of the Virtual Guests here’s more or less what I’ve used :

Server 01 (The second optional test server has the same configuration apart from the second virtual HDD)

RAM 1024MB
Default NAT network
HDD01 12GB
HDD02 2GB
Desktop installation (just for testing some of the GUI tools)
OS Scientific Linux 6.1/CentOS 6.2 (really unimportant the RHCSA exam

I’ve used two HDD, below the partitioning scheme, but partitioned only the first one for installing the OS and /home directories while used the second HDD for testing stuff like LUKS encryption and LVM configuration :

/dev/sda1 1 14 112423+ 8e Linux LVM
/dev/sda2 15 28 112455 8e Linux LVM
/dev/sda3 29 42 112455 8e Linux LVM
/dev/sda4 43 130 706860 83 Linux

RHCSA – Exam Experience FAQs

Ok so with all the in and out of how I did prepare for my exam let’s discuss the exam experience itself. Let’s start with saying that my job is that of a Linux Systems Engineer and I deal daily with complex configuration and systems hence something that is “easy” for me could prove a challenge for you, or vice versa, secondly I’m already certified as RHCE so at some extent I’ve already been exposed to the stressfull exam experience and was somehow prepared but here I will try to cover every and each tip

Are RedHat exams really so difficult?

While the answers can and will vary from person to person my personal view of of the matter is RedHat exams are something totally different from the typical IT Exams people are used to, you’ve been through LPI or MCSE or Cisco exams?, being an hands on exam where you sit for 2.5 hours in a lab configuring a server with high pressure on your shoulder is not an easily endured torture

On the other hand some people, and I am among these, perform much much better when dealing with real world problems and not with the typical multiple choice kind of question which asks you if you should use the -d or the -e switch with this or that command.

So the exam is easy and people are just overrating its difficulty?

No the exam is difficult, but isn’t that the case with every test in life? Around the net I’ve read every kind of horror story about RedHat exams, let me tell you this IGNORE them. RHCSA Exam is tough, you need to study hard for it and know the topics by heart… if you did this you will pass without any problem. All you need to know is in the exam requirements checklist, don’t be worried about “out of syllabus questions” or stuff like that you will not see it in the RHCSA exam.

Are the question tricky?

Well in my experience no all question were pretty straightforward like “Do configure this service in this and that way” apart from one question on a specific subject but to that end I have to say that the proctor has been really helpful in explaining, without giving me the answer of course , what the test was expecting from me.

Do I need to study all the topics in the Syllabus?

I think I can safely tell that you will not be asked every dark detail of what you see in the exam syllabus but you should know everything as your exam could be different from mine and you are supposed to know that stuff in any case both for the exam and for real life use. In addition to study also make sure you practice everything, knowing of a service or command often is not enough and you have to test it out yourself to make sure you know what to do in case something does not go exactly as written in the book.

Will I be using a Virtual or Physical system during the RHCSA exam?

Both. Let me elaborate a bit, you will be given a workstation with RHEL installed which you will use as the host system on which a virtual guest could be installed/configured and that is where you will be working to complete the tasks required from the exam.

Will I have access to the internet?

Of course not and you will be asked to put away your phone and kindle (mentioning this as the proctor saw me with it while I was waiting to enter the test room and mentioned it during the briefing ehehe)

Can I use the man pages?

Yes of course, my suggestion here is try not to depend on them this is not because RedHat will ever take them away from the RHCSA exam, nobody expect you to know each and every switch for the lvm* commands, but because you are in an exam environment and the time is quick when you’re enjoying yourself.

Does every candidate get the same exam (thanks to Bodanel for this one)?

As reported by one of the blog readers and as it is logical no, the exam is not the same for all candidates in the lab room and between sessions so let me reiterate over a very important, maybe the most important, point know your stuff by heart. Study each subject and practice it, this will make the difference between passing and failing the exam.

How much time does it take to prepare for the RHCSA exam?

That really depends from person to person, personally with my working schedule it took me 4 months which also account for finding a suitable date to sit the lab. It all boils down to your experience and knowledge of Linux

Can I sit the exam without any previous Linux knowledge?

Will is power, the only limits we have are the only ones we impose on ourselves. It will not the be easiest path but doable.

Will the RHCSA exam make a Linux engineer of me “as see in TV™”?

Ok maybe not really an exam related question but I get asked this quite frequently as of lately No the RHCSA exam will not make an engineer of you just for holding RHCSA certification, or RHCE to that end, the exam and the paper are only the starting point the rest is hard work, study and experience.

RHCSA Exam – The closing note

Well I think this is more or less it, I hope to give you a good coverage of the whole RHCSA exam experience and what to expect from it, I’ve omitted any personal thought or experience and it will vary from testing center to testing center and country, I can tell both my exams in Switzerland were top and I don’t expect anything less from RedHat.

RHCSA exam is not an easy exam but the satisfaction you get from passing it is well worth all the efforts and the dedication that it takes to prepare for it and in this post I’ve tried to summarize all the tools and tricks I’ve used to prepare and pass the exam hoping to help you have an easy task in preparing and passing the exam.

If there is any question or “part” that I did not cover please feel free to leave me a comment or send me a message directly I will be more than glad to help you out and in case update the post with your suggestion, specially the FAQs part.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

SELinux Restore Default File Context

Lethe — Tue, 03 Jan 2012 13:39:56 +0000

Q: I’ve accidentally modified the SELinux context for a file or directory and now my application is not working any more how canI restore the default context?

A: As you probably know SELinux is enabled in Enforcing mode by default on RHEL and Fedora systems and while it helps keep your system secure, of course it is not the panacea to all security risks just one the bricks for a better protection, sometimes it can get in the “middle” of normal operations.

In a system protected by SELinux every file and/or directory has a default context assigned to it which is usually defined by the path under which the object is created and is defined under the /etc/selinux/targeted/contexts/files path in the file_contexts file, I encourage you to open up the file with your favorite editor and have a look at it.

If you changed the context of a file or directory with the chcon command and realize that this disrupted the normal operations of the system you can use the following basic command to restore the default SELinux context :

 restorecon -v $FILENAME

That’s is, SELinux will try to match the file path with he default context and will eventually assign it again to the file or directory restoring the original context and hopefully solving the malfunction of the system.

SELinux is a fascinating and really powerful tool which I encourage you to know and use, many administrators simply disable its control, for this reason I will write again about SELinux implementations and its troubleshooting.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

Configure Linux VNC Server

Lethe — Sun, 01 Jan 2012 12:30:19 +0000

While most of the Linux servers I administer and run have no GUI installed sometimes it can be handy to administer a remote Linux workstation or server via a GUI, this is also an exam requirement for the the Red Hat Certified Systems Administration exam so let’s see how a VNC server is installed and configured under Linux.

Install VNC Server

The first step of setting up a VNC Server under Linux is of course that of installing the required packages, for RHEL and its rebuilt systems the process is pretty straightforward and quick and can be achieved with the following command :

 yum install tigervnc-server

Once yum has done installing the required package, which has very few dependencies, you can proceed to the configuration of the various VNC Sessions.

Configure VNC Server

Once VNC server has been installed you need to configure the various sessions and the ports on which the VNC server will listen, the syntax of the needed command, in its simplest form, is as follows :

 vncserver

When launched without any argument the command will configure the required configuration files under the user’s home directory and will create a listener on the default port for VNC which is 5901, this is good as long as you need to configure only one VNC server but how to proceed in case you need to configure multiple VNC sessions for different users? The command to accomplish this is as follows :

 vncserver :2

After asking you to input a session password, the above command will configure a listening VNC Server for the current user on port 5902 as specified by the argument :2 which can really be any other valid number relative to the 590X port.

VNC Server change password

As anticipated the vncserver command will configure a VNC session for any user and will require you to input a password to protect the session, if later on you need to change the password chosen you can easily do that with the following command :

 vncpasswd

Again the system will require you to input the password twice and will configure it for the session.

To test the configuration of the VNC server just configured you can use the default viewer that ships with Linux to open a connection like in the following picture :

That’s all for the VNC Server configuration under Linux, I hope you did find the post useful and will look forward your comments or questions in case you have any.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

Fedora 16 unable to start vsftpd
Powered by Where did they go from here?

Fedora 16 unable to start vsftpd

Lethe — Thu, 01 Dec 2011 06:31:23 +0000

*EDIT* As the Fedora project released an updated vsftpd package which fixes the described problem the article has become obsolete, I leave it in place just as a reference.

The default FTP server for the Fedora, and RHEL like distributions, is vsftpd which can be easily installed with the command :

 yum install vsftpd

Vsftpd is a feature rich FTP server and up to today it served all my needs but the version that ships with Fedora 16 has still an unresolved bug that makes starting and stopping the service awkward and in any case not as described in the official Fedora documentation.

Fedora 16 – Start VSFTP

According to the official documentation once you have the vsftp package installed you should start the service with the command :

 systemctl start vsftpd.service

When launching the command you will see a message like the following :

 Failed to issue method call: Unit vsftpd.service failed to load:
No such file or directory. See system logs and 'systemctl status vsftpd.service'
for details.

While a patch to this problem is already available in the Fedora testing repository if you don’t feel like messing with test updates you can easily solve the problem starting the service with the command :

 systemctl start vsftpd@vsftpd.service

In case you want the service to automatically start you can use the command :

 systemctl enable vsftpd@.service

Until the patch is fully accepted and released in the stable repositories above solution should work without any problem.

I hope you did find the article useful and would appreciate if you could take a second to comment or share it.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

Linux kill multiple process with a single command

Lethe — Mon, 28 Nov 2011 10:55:28 +0000

Sometimes it can be useful to kill multiple processes with a single command, while using a combination of ps and grep can help you hunt down the process you need and to kill it this is not the easiest or fastest approach if you need to kill a large number of processes. Let’s see how to do this.

The first step would be getting a list of the interesting processes we need to kill, in this example I will use the NFS daemon, which can easily be accomplished with a command similar to the following :

 ps -ef  | grep nfs

The command on my system produced the following output :

 ps -ef | grep nfs
root      1873     2  0 08:50 ?        00:00:00 [nfsd4]
root      1874     2  0 08:50 ?        00:00:00 [nfsd4_callbacks]
root      1875     2  0 08:50 ?        00:00:00 [nfsd]
root      1876     2  0 08:50 ?        00:00:00 [nfsd]
root      1877     2  0 08:50 ?        00:00:00 [nfsd]
root      1878     2  0 08:50 ?        00:00:00 [nfsd]
root      1879     2  0 08:50 ?        00:00:00 [nfsd]
root      1880     2  0 08:50 ?        00:00:00 [nfsd]
root      1881     2  0 08:50 ?        00:00:00 [nfsd]
root      1882     2  0 08:50 ?        00:00:00 [nfsd]
root      4835  4554  0 18:43 pts/0    00:00:00 grep nfs

Ok we can see the full list of processes bound to NFS together with other information we actually don’t need, what is really interesting for us is PID of the process which is exposed in the second column of the output so let’s call awk to help us in the task :

 ps -ef | grep nfs | awk '{print $2}'

The above command will list all processes on the system bout to NFS but this time instead of printing everything it will print on screen only the content of the second output column which again on my system looks like :

 ps -ef | grep nfs | awk '{print $2}'
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
4871

Ok we’re almost there, now let’s put all together and let’s kill all the processes with the command :

  kill `ps -ef | grep nfs | awk '{print $2}'`

That’s it all the processes that results from our expression will be killed with the single command, that is even easy to script when you often need to kill processes on a server (I keep a copy of it on one of my lab machines where I often run a script that launches 100+ instances of the same program and need to kill them all).

I hope you will find the article interesting and useful and would appreciate if you could take the time to retweet or share it.

Cheers Lethe.

Incoming search terms:

Readers who viewed this page, also viewed:

Linux change the speed and duplex settings of an Ethernet card | linux change duplex settings

Lethe — Tue, 01 Nov 2011 18:27:02 +0000

Q: I need to change the speed and duplex settings for my network card, is this possible?

A: Under Linux you can use the built-in tool ethtool to display and change the duplex settings for your network card under Linux, depending on the NIC installed on your system you could try to use the mii-tool as well. In my experience ethtool works for most of the modern network cards so I will describe only the use of this tool

Install ethtool

Depending on your system ethtool could be already present and so you will be able to change duplex settings right away, in case it is not present you can easily install with the following command:

 yum install ethtool

Or if you are using Debian or Ubuntu with the command :

 apt-get install ehtool

Get information of duplex settings for network card

Once you have installed the tools of the trade is time to show test they are functioning, the syntax of the command is pretty straightforward all you have to do for a basic test is launch the command ethtool followed by the name of the interface for which we want to get information here’s output for my eth0 interface :

 Settings for eth0:
Supported ports: [ TP MII ]
Supported link modes:   10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supports auto-negotiation: Yes
Advertised link modes:  10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised pause frame use: Symmetric Receive-only
Advertised auto-negotiation: Yes
Speed: 10Mb/s
Duplex: Half
Port: MII
PHYAD: 0
Transceiver: internal
Auto-negotiation: on
Supports Wake-on: pumbg
Wake-on: g
Current message level: 0x00000033 (51)
drv probe ifdown ifup
Link detected: no

In the above output you can see all important information about the network card, it is disconnected so your output could be slightly different, assuming that you need to configure you card to operate on full duplex at a speed of 100 Mbps you would use the following command :

 ethtool -s eth0 speed 100 duplex full

ethtool -s eth0 speed 100 duplex half

ethtool -s eth0 speed 10 duplex full

As you can see above I gave multiple examples on how to use ethtool to configure various duplex setting for the network card, be aware that changes made this way will not survive a reboot and so are a good way to test different settings.

Make Duplex Settings change permanent

Once you have configured your network card to the correct duplex settings you will want to make the changes permanent, to do so you will need to change the configuration file for your network card in our example the file will be /etc/sysconfig/network-scripts/ifcfg-eth0 on a Red Hat based system and add the following text :

 ETHTOOL_OPTS="speed 100 duplex full autoneg off"

service networking restart

The above will set the speed of the network card to 100 Mbps with full duplex enabled disabling auto negotiation, the last command is needed so that Linux will pick up the duplex changes, of course you could also reboot the server but I prefer to leave that to the Windows world

Under Debian/Ubuntu the needed steps are slightly different, in this case you will need to create a bash script and put something like the following code :

 #!/bin/sh
 ETHTOOL="/usr/sbin/ethtool"
 DEV="eth0"
 SPEED="100 duplex full"
 case "$1" in
 start)
 echo -n "Setting eth0 speed 100 duplex full...";
 $ETHTOOL -s $DEV speed $SPEED;
 echo " done.";;
 stop)
 ;;
 esac
 exit 0

Once saved the script make it executable and copy it under the /etc/init.d/ directory and finally add it to the startup sequence with the command :

 update-rc.d  defaults

You can try to restart the system to check the results of the new configuration.

That’s all you need to change duplex settings under Linux, ethtool is invaluable when testing and troubleshooting network related problems and I hope this post will help you as well.

In case you face any problem feel free to write me or leave a command and I will be more than glad to help you out.

Cheers Lethe.