[Workaround.org] Postfix Crashing
Charlie Ullman
charlie at ckgt.com
Mon Nov 5 21:23:52 CET 2007
Incidentally, I found the following lines in my syslog file:
Nov 4 23:14:01 ckgt /USR/SBIN/CRON[30904]: (root) CMD
(/usr/sbin/sendmail -q)
Nov 4 23:16:01 ckgt /USR/SBIN/CRON[30906]: (bin) CMD ([ -x
/usr/lib/courier/sqwebmail/cleancache.pl ] &&
/usr/lib/courier/sqwebmail/cleancache.pl)
Nov 4 23:17:02 ckgt /USR/SBIN/CRON[30908]: (root) CMD ( cd / &&
run-parts --report /etc/cron.hourly)
Nov 4 23:17:02 ckgt postfix/master[30277]: terminating on signal 15
Those are two cron jobs (sendmail and sqwebmail) that were left over
from when I had sendmail and sqwebmail installed. Could they possibly
have been causing the crash? Anyway, I have removed those two cron jobs.
Maybe that will have fixed it, I don't know.
Charlie Ullman wrote:
> OK, I've double-checked it's not a hardware problem. My server is
> actually a miniserver, so I can't run memtest, but I double-checked with
> my provider, and they said the memory is fine.
>
> They suggested I check dmesg output to see whether OOM Killer is kiling
> postfix, but I just checked it and there isn't any entry corresponding
> with the last time postfix crashed.
>
> I'd really be enormously grateful for any help on this one. I just don't
> know where to look for the source of this problem.
>
>
> Martin Simovic wrote:
>
>> make sure this is not a hardware problem;
>>
>> try tu run memtest overnight on a server, faulty memory could be a reason
>>
>> martin.
>>
>> 2007/11/3, Charlie Ullman <charlie at ckgt.com>:
>>
>>
>>> Charlie Ullman wrote:
>>>
>>>
>>>> Charlie Ullman wrote:
>>>>
>>>>
>>>>> Hello!
>>>>>
>>>>> I have been having problems with postfix crashing quite often (every
>>>>> other day). It's been going on for months. I increased the amount of
>>>>> memory on my server, but that didn't seem to help. Then I added a
>>>>> daily cron job that restarted postfix (and Amavis) every morning. But
>>>>> it still often remained crashed. I have now added an hourly cron job
>>>>> to restart postfix, but this seems like overkill. I really need to
>>>>> find out why postfix is crashing, the standard mail.log file doesn't
>>>>> seem to give very good information:
>>>>>
>>>>> For instance, here is a crash after the daily cron job. Amavis was
>>>>> restarted, and then postfix, but the log just says that postfix crashed:
>>>>>
>>>>> Oct 31 06:30:34 ckgt amavis[9351]: Internal decoder for .tnef
>>>>> Oct 31 06:30:34 ckgt amavis[9351]: Found decoder for .exe at
>>>>> /usr/bin/unrar; /usr/bin/lha; /usr/bin/arj
>>>>> Oct 31 06:30:34 ckgt amavis[9351]: Using internal av scanner code for
>>>>> (primary) ClamAV-clamd
>>>>> Oct 31 06:30:34 ckgt amavis[9351]: Using internal av scanner code for
>>>>> (primary) check-jpeg
>>>>> Oct 31 06:30:34 ckgt amavis[9351]: Found secondary av scanner
>>>>> ClamAV-clamscan at /usr/bin/clamscan
>>>>> Oct 31 06:30:34 ckgt amavis[9351]: Creating db in
>>>>> /var/lib/amavis/db/; BerkeleyDB 0.31, libdb 4.4
>>>>> Oct 31 06:30:34 ckgt postfix/master[3746]: terminating on signal 15
>>>>> Oct 31 06:34:02 ckgt postfix/postqueue[9456]: fatal: Cannot flush
>>>>> mail queue - mail system is down
>>>>>
>>>>> If I can't work out what is wrong with postfix (can anyone give me
>>>>> any pointers as to where I should be looking?), then maybe I can
>>>>> optimise the hourly cron job a bit. Maybe rather than just blindly
>>>>> restarting postfix every time, it could check whether the process is
>>>>> running or not. Something like:
>>>>>
>>>>> if [ ! -d /var/run/postfix.pid ] # if the process file doesn't
>>>>> exist
>>>>> then
>>>>> /etc/init.d/postfix restart
>>>>> fi
>>>>>
>>>>>
>>>>> The only problem with this, of course, is that I don't know where the
>>>>> postfix process file would normally be stored/created (I just gave
>>>>> /var/run/postfix.pid as an example, I know it's not there).
>>>>>
>>>>> Can anyone help me? I'm really at my wit's end on this one...
>>>>>
>>>>> Cheers
>>>>>
>>>>> Charlie
>>>>>
>>>>>
>>>>>
>>>> Oh hang on, it's /var/spool/postfix/pid/master.pid isn't it? Oh well,
>>>> anyway, can anyone help me with the original problem (that postfix
>>>> seems to keep crashing)?
>>>>
>>>> Thanks
>>>>
>>>> Charlie
>>>>
>>>>
>>>>
>>> Hmm, I also should have used -e rather than -d.
>>>
>>> However, I do appear to have got the location of the postfix process
>>> file wrong. Even with postfix stopped /var/spool/postfix/pid/master.pid
>>> still seems to exist.
>>>
>>> Can anyone advise?
>>> _______________________________________________
>>> Workaround-chitchat mailing list
>>> Workaround-chitchat at workaround.org
>>> http://workaround.org/cgi-bin/mailman/listinfo/workaround-chitchat
>>>
>>>
>>>
>>
>>
>
> _______________________________________________
> Workaround-chitchat mailing list
> Workaround-chitchat at workaround.org
> http://workaround.org/cgi-bin/mailman/listinfo/workaround-chitchat
>
>
More information about the Workaround-chitchat
mailing list