Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have a Xen domU running Debian Unstable on a Xen 3.4.2 host.

This morning I noticed that various apps have started seg-faulting.

In particular running "aptitude safe-upgrade" causes a segfault in aptitude-curses with the following error:

aptitude[1035]: segfault at 7f1006ed13f8 ip 0000000000544293 sp 00007fff94b37140 error 4 in aptitude-curses[400000+331000]

This segfault is totally reproducible.

Other apps (such as reportbug, and trying to log into KDE) also cause reproducible segfaults.

I have another Xen domU running Debian Unstable on the same Xen host, running the same kernel (2.6.32.2), but running "aptitude safe-upgrade" doesn't cause a segfault. Both domUs seem equivalent, yet only one of them segfaults. Here are some points to note:

  • Both domUs use exactly the same kernel (64-bit)
  • Both have the same binaries for /usr/bin/aptitude-curses and all the shared libs it depends on (I used md5sum to compare files on both systems, and ldd to see which shared libs aptitude depends on)
  • I did e2fsck -f on the domU root volume that has the problems and there were no reported errors
  • Both domUs have the same amount of RAM and VCPUs allocated to them
  • I know that segfaults could point to hardware failure, but given that these segfaults are reproducible and that an equivalent domU has no problems (even if I change the order in which the domUs are created, hoping to force each to occupy a different part of physical RAM), it strongly suggests that hardware is not a problem
  • I also wonder whether some files are corrupt, but as I said aptitude and all its dependent libs (as reported by ldd) seem ok
  • I have rebooted the problem domU many times, and rebooted the host Xen OS once
  • I have tried booting the segfaulting domU in single-user mode (by setting the default level to 1 in /etc/inittab) and "aptitude safe-upgrade" still segfaults.

I don't think this is a Xen problem, but without knowing what's causing this I can't be sure.

I'm totally perplexed by why one virtual machine should keep segfaulting, and another similar VM doesn't.

Any help would be greatly appreciated.

Thanks.

share|improve this question

2 Answers 2

Run the segfaulting program in gdb, with debugging symbols for all the relevant libraries installed, and diagnose the cause of the problem from there.

share|improve this answer

It can be a memory module failure that spoils running apps' memoryy. Try memtest86+ to ensure your RAM modules are okay.

If yes, then probably you should start analyzing libs: /lib , /usr/lib etc. An easy way is md5sum + diff on different working Linux boxes: maybe some of them are really corrupted?

share|improve this answer

Your Answer

 
discard

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.