veeamservice error; core dump generated at same time

[aj_potc]

Case #04723659

I'm running CentOS 7 on a machine that acts both as a repository and as an agent for backing up a database dump file. I recently experienced several backup jobs failing in a row. When I checked my system log, I found the following entry:

Code: Select all

veeamagent: *** Error in `/tmp/VeeamAgentef08b7ca-98a0-49b8-9dd6-b0928398386b.data/veeamagent': double free or corruption (out): 0x00007ff2a00032f0 ***

By chance, I later discovered a core dump file in root's home directory with an identical timestamp.

Based on these findings, I filed a case with support, suspecting that the Veeam Agent for Linux had crashed and thinking that Veeam would want to know about it. I supplied both my logs and the core dump. However, the support agent does not believe this is a Veeam core dump. I find it very odd that this would not be related to Veeam considering the identical timestamps.

Does the above error necessarily indicate that some Veeam component crashed?

Has anyone seen this before, or have any suggestions on how to proceed?

Thank you very much.

[HannesK]

Hello,
existing core-dumps on Linux repositories is something I have heard of usually when there are storage or any other hardware issues. So finding the root cause can be complicated especially if the issues happens only from time to time. Usually it also involves the Linux distribution support and hardware support.

If that was a one-time issue, then I would just stop every investigation and save every-ones time. If you see it regularly, then I recommend to submit your support ID (currently you are in the evaluation support queue) and escalate the case if necessary (talk to manager button)

Best regards,
Hannes

[aj_potc]

Hi Hannes,

Whether it's worth pursuing is an open question, but I found another core dump today. I've submitted it and escalated the case for review.

This time the error in the system log is a bit different:

Code: Select all

veeamagent: *** Error in `/tmp/VeeamAgent8ef79157-d379-45aa-b5b7-47ad5725c439.data/veeamagent': munmap_chunk(): invalid pointer: 0x00007f58540021e0 ***

Thanks.

[PTide]

@aj_potc,

Thank you for connecting the dots. I think the issue is worth researching and it will be investigated.

Cheers

[aj_potc]

Thank you very much, @PTide.

I tried myself using gdb to analyze the core file against the veeamagent binary like this:

Code: Select all

gdb /usr/sbin/veeamagent core.15334

I then ran the command bt full to get a backtrace.

However, it seems that the binary doesn't have debugging symbols included, so the output didn't appear very helpful.

I suppose your engineers can analyze against a version of the binary with those symbols included, right? I know next to nothing about this -- just trying to help how I can.

Thanks again.