Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am trying to setup a PXE Boot server with RHEL 6 and am following these directions.

Has anyone done this successfully?? I am having various issues. I used the rsync command to copy over the contents of another machine to use as a source. It seems to mount correctly, and by the errors I get, the new server seems to find it ok. Unfortunately, the errors on the console cannot be saved anywhere (that I know of), so I have to type the important details. Here it goes.

The following things appear to work correctly:

  • DHCP - I am getting an address from my server on boot correctly.
  • TFTP - My files are getting transferred over correctly, and the RedHat install seems to start.

After what appears to be RedHat installion starting up, I start getting errors like the following.

udevd-work[1081]: Error changing netif name 'eth3' to 'eth8' device or resource busy (it does this for each of the ethx adapters)

Setting up Logical Volume Management: /etc/lvm/lvm.conf: stat failed: permission denied

fsck.ext4: No such file or directory /dev/mapper/vg_myhome-lv_root

The superblock could not be read or does not describe a correct ext2. Run e2fsck -b 8193

So many errors here I don't understand.

  1. Why is the PXE process trying to rename by ethX ports? (maybe I don't care).

  2. Permissions error for LVM? Not sure how. I am getting my "image" from an NFS mount from another machine. It seems to find it ok... but I have never specified a user/password on this "raw" box. How can I get a permission denied??

  3. fsch.ext4 error. That file does exist in my mounted image.. so I don't get that.

  4. Superblock? I am using ext4, mainly... why is it complaining about ext2?

Anyway, the process then says it is going to shutdown and give me a chance to enter a root password for maintenance, but I don't know the password, as I have not set up the box yet (seems like a chicken/egg thing). I tried the root from the image and from the PXE server, but that did not work.

Anyone know how to do this?

share|improve this question

migrated from Jun 1 '11 at 14:34

This question came from our site for professional and enthusiast programmers.

Coincidentally, I did find (what I believe is) an error in the documentation I referenced. In section 19.3, step #3 directs to use the dracut service to create initrd. The example given states: dracut initramfs-kernel-version.img vmlinuz-kernel-version. I think the vmlinuz-kernel-version, however should be just kernel-version, as file-not-found appear when putting the full thing. Possibly related to my problems? – John Galt... who Jun 1 '11 at 15:18
Well, this isn't an answer, but I noticed that your remote, diskless system is attempting to run fsck on a logical volume... Which implies a physical volume, ie a HDD... thus the initrd, and probably the /etc/fstab has something directed at a lv... The image should be a squashfs, and the /etc/fstab shouldn't have any reference to dev/mapper/vg_myhome-lv_root... in all seriousness if the system had no hdd, lvm is un-necessary on the image. – darkdragn Jun 2 '11 at 1:17
Thanks Christopher. I wish I understood that better, but I am a bit of a admin newbie. I did notice that the machine I am trying to build this from, is not setup as I am familiar with. When it was setup, a 'mkfs -t ext4 /dev/sdx' was run for each of the physical drives (8 of then), and them mounted in that fashion. So, no direct logical drives, other that was setup during the installation program. I think what I am saying is.... I don't know what I am doing and you might just be on to the answer. I am going to rebuild a test machine, trying it again with something more standard. Thanks – John Galt... who Jun 2 '11 at 11:55
Any time, I know that the package managers are a bit different, however I think this link can give you an idea of what to: . It has the basics, as far as compressing the root partition as a squashfs, and having some kind of casper scripts that will seek the squash fs. If you are starting with an install on a live system, when you make the squashfs you have to make sure you blank the /etc/fstab, and casper should do it's job mounting the squashfs as the root. If you want, you can send me a private message, I have some custom casper scripts. – darkdragn Jun 2 '11 at 21:02

You must log in to answer this question.

Browse other questions tagged .