Hello!

On Tue, Jul 07, 2009 at 02:22:24PM +0200, InterNetX - Jürgen Gotteswinter wrote:

> i found some discussions about this, and it seems that the nginx perl
> module can cause this on systems with threaded perl. i recompiled the
> rpm without the "--with-perl..." part and reinstalled nginx. a few
> minutes later nginx died again, here are the last lines from error.log
>
> 2009/07/07 14:00:04 [notice] 15087#0: signal 3 (SIGQUIT) received,
> shutting down
> 2009/07/07 14:00:04 [notice] 15088#0: gracefully shutting down
> 2009/07/07 14:00:07 [notice] 15088#0: exiting
> 2009/07/07 14:00:07 [notice] 15088#0: exit
> 2009/07/07 14:00:07 [notice] 15087#0: signal 17 (SIGCHLD) received
> 2009/07/07 14:00:07 [notice] 15087#0: worker process 15088 exited with
> code 0
> 2009/07/07 14:00:07 [notice] 15087#0: exit
>
> any ideas?

Somebody sent QUIT signal to nginx master. From the time (14:00)
it looks like wrongly implemented log rotation (e.g. via restart
or something like). For log rotation use USR1 signal, it was
specially designated for this (and instructs nginx to just reopen
logs without any extra unneeded work).

Maxim Dounin

>
> On 07/07/2009 01:36 PM, Juergen Gotteswinter wrote:
>> Hello,
>>
>> i got a nginx 0.7.61 running on centos 5.3 x86_64 which keeps crashing
>> with
>>
>> panic: MUTEX_LOCK (22) [op.c:352].
>>
>> any ideas? or is this a known problem?
>>
>>
>>
>
> --
> InterNetX GmbH
> Maximilianstrasse 6
> D-93047 Regensburg
>
> Tel. +49 941 59559-480
> Fax +49 941 59559-245
>
> Geschäftsführer/CEO: Thomas Mörz
> Amtsgericht Regensburg, HRB 7142
>
>