Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

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

So I have a CentOS webserver running on a 2950. I assumed for sure that it was 64 bit, but according to uname it seems to be i686:

 [root@www ~]# uname -a
 Linux 2.6.18-92.1.10.el5PAE #1 SMP Tue Aug 5 08:14:05 EDT 2008 i686 i686 i386 GNU/Linux

So, anyway, I clonezilled it, and dropped it on a badass R900 (16 cores, 64GB). Unfortunately, when I go to boot up the new server, I just get a million of these errors:

Unknown interrupt or fault at EIP 000000060 c04011ed 000011ec

So I tried booting it up without acpi, but still no love. Do I need to change the boot kernel? How the hell would I go about doing that?

I'll take any bone that you all are willing to throw my way!

share|improve this question
Uh, you haven't updated the kernel on that machine since Aug 5th 2008? – MDMarra May 4 '12 at 17:12
how about virtualizing the server? VMware converter would do it all for you if you install free esxi on to your new R900. There will still be enough space (memory) for a bunch of other servers. – Dmitri Chubarov May 4 '12 at 17:15
up vote 7 down vote accepted

This is a 32-bit OS. You're trying to move to a dissimilar architecture with a very old OS/kernel. You should really rebuild onto the new hardware and migrate your data. There's no way this would work without major surgery.

share|improve this answer
there is a way, but it's a pain in the (.) to follow :) – dyasny May 4 '12 at 17:54

Get the list packages that are installed on the system and install the corresponding packages on the new machine. Tweak your new machine's configuration as needed.

Afterwards, you can copy over every file that is not identified as owned by the package management system as well as manually deciding which configuration files are related.

For avoiding the problems you're experiencing right now, it is good practice to keep a separate view of your programs, your data, and your configuration files and to migrate fairly regularly if you can afford to, even if just to a test server that is wiped down afterwards. You may find that tools such as Puppet are helpful for this.

As somebody else suggested, you can get by in the immediate term by using virtualization software. There are many packages that allow conversion of a hardware system to a VM guest system. In the long term, however, you'll need to adjust and become capable a proper migration.

Also, I feel the need to echo MDMarra's observation that not updating your kernel for almost four years is a bit scary.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.