Bug 10246 - Segfault in Samba from time to time
Summary: Segfault in Samba from time to time
Status: RESOLVED WORKSFORME
Alias: None
Product: Samba 3.6
Classification: Unclassified
Component: Domain Control (show other bugs)
Version: 3.6.9
Hardware: x64 Linux
: P5 critical
Target Milestone: ---
Assignee: Guenther Deschner
QA Contact: Samba QA Contact
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-11-04 09:25 UTC by goetz.reinicke
Modified: 2020-12-19 16:08 UTC (History)
1 user (show)

See Also:


Attachments
one 'panic action' output (3.35 KB, text/plain)
2013-11-04 10:03 UTC, goetz.reinicke
no flags Details
recent backtrack / panic action script output (17.98 KB, text/plain)
2013-11-15 09:48 UTC, goetz.reinicke
no flags Details
Backtrack with machine account information (18.16 KB, text/plain)
2013-11-15 10:07 UTC, goetz.reinicke
no flags Details
smblogfile from 3:14h when a crash happened (8.35 KB, text/plain)
2013-11-15 10:15 UTC, goetz.reinicke
no flags Details
panic script backtrack from 3:14h when a crash happened (12.58 KB, text/plain)
2013-11-15 10:16 UTC, goetz.reinicke
no flags Details
panic script backtrack when a crash happened (18.73 KB, text/plain)
2014-02-26 10:24 UTC, goetz.reinicke
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description goetz.reinicke 2013-11-04 09:25:06 UTC
Dear samba team,

we still run a Red Hat EL 6.x samba-3.6.9-151 PDC with domain login, roaming profiles, Windws 7 clients and LDAP back end.

In the last couple of weeks we notice some unregular crashes with abrt reports. But as an university our RH subscription dose not include the full support.

The crash happens when users try to login to the windows clients. (Users who logged in before the crash still can access all shares and don't notice any problem.)

I cant see any changes to the installation or configuration which might have an influence to the chrashes.

So any suggestion or help on debugging that problem is very appreciated.

I can provide the abrt and I enabled "debug" and "panic action". So there might be some information which could help to debug that problem.

WHat information may i provide? What can I check?

Thanks and Regards . Götz Reinicke
Comment 1 Volker Lendecke 2013-11-04 09:38:24 UTC
Installing debuginfo rpms and giving us the stacktrace output from /var/log/messages or log.smbd would be a first step.
Comment 2 goetz.reinicke 2013-11-04 10:00:39 UTC
(In reply to comment #1)
> Installing debuginfo rpms and giving us the stacktrace output from
> /var/log/messages or log.smbd would be a first step.

O.K. I installed the debuginfo rpm but I guess now we have to wait till the next crash?

Will the current collected information already help?

I can add one output so you see what I got.
Comment 3 goetz.reinicke 2013-11-04 10:03:05 UTC
Created attachment 9363 [details]
one 'panic action' output
Comment 4 Volker Lendecke 2013-11-04 11:01:56 UTC
Is there also preceding log.smbd output available?
Comment 5 goetz.reinicke 2013-11-04 12:19:14 UTC
(In reply to comment #4)
> Is there also preceding log.smbd output available?

I'm afraid not. We have log level 2, one combined log file and max log side 20MB that was flooded.

I increased the side to 40MB. But on the other hand; would you suggest to split the log file by client/system and set the size to e.g. 5 MB?

We have about 100 clients and disc space is no problem at the moment.

What log level might be good in that case?
Comment 6 goetz.reinicke 2013-11-15 09:48:03 UTC
Created attachment 9425 [details]
recent backtrack / panic action script output
Comment 7 goetz.reinicke 2013-11-15 10:07:29 UTC
Created attachment 9426 [details]
Backtrack with machine account information

in this backtrack there are some mashine account information recorded. Like:

#18 0x00007ffbbaa4962b in samr_find_machine_account (mem_ctx=0x7ffbbcfc6e20, b=0x7ffbbcfc7760, account_name=0x7ffbbcfc6980 "PC10768$", domain_sid_p=0x7fffbd936408, user_rid_p=0x7fffbd93641c, user_handle=0x7fffbd936480, access_mask=33554432) at rpc_server/netlogon/srv_netlog_nt.c:600


may be that BT gives some hints.
Comment 8 goetz.reinicke 2013-11-15 10:15:52 UTC
Created attachment 9427 [details]
smblogfile from 3:14h when a crash happened
Comment 9 goetz.reinicke 2013-11-15 10:16:17 UTC
Created attachment 9428 [details]
panic script backtrack from 3:14h when a crash happened
Comment 10 goetz.reinicke 2014-02-26 10:23:45 UTC
Hi,

I'v seen that some debug information where missing.... I apologies for that.

Today again a crash happened but now with more debug information.

Any help and suggestion would be great!
Comment 11 goetz.reinicke 2014-02-26 10:24:36 UTC
Created attachment 9726 [details]
panic script backtrack when a crash happened
Comment 12 Björn Jacke 2020-12-19 16:08:56 UTC
please file a new bug report if this is still an issue with modern releases like 4.12 newer