apache graceful restart kills ifolder
Hi,
I have an iFolder 3.6 server on OES2 (not SP1) x86_64 server.
Every 24 hours Apache does a graceful restart - I believe this is so the logs can be rolled.
I have not been able to find how this is triggered (it is NOT in Crontab !) or how it picks what time to do it.
It was happening at 8:30pm, now since a power outage it happens at 2:00pm. Restarting the server now does NOT change when it occurs though....
When Apache does this "graceful" restart it is anything but graceful. A snippet from /var/log/apache2/error_log
>>>>>
[Thu Oct 29 14:00:04 2009] [notice] Apache/2.2.3 (Linux/SUSE) configured -- resuming normal operations
log4net:ERROR [RollingFileAppender] Unable to acquire lock on file /ifdata/simias/log/Simias.log. Sharing violation on path /ifdata/simias/log/Simias.log
log4net:ERROR [RollingFileAppender] Unable to acquire lock on file /ifdata/simias/log/Simias.access.log. Sharing violation on path /ifdata/simias/log/Simias.access.log
log4net:ERROR [RollingFileAppender] Unable to acquire lock on file /ifdata/simias/log/Simias.log. Sharing violation on path /ifdata/simias/log/Simias.log
log4net:ERROR [RollingFileAppender] Unable to acquire lock on file /ifdata/simias/log/Simias.access.log. Sharing violation on path /ifdata/simias/log/Simias.access.log
<<<<<
After this clients drop all connection and cannot reconnect.
I seriously think a "sleep 20" in the right spot would fix this !
To recover, I stop apache, wait, kill all mono processes and restart apache. We are then good until 2pm the next day...
log4net seems to be a Apache component.
I have set iFolder logging to NONE - no change !
Anyone know:
1. How to fix it
2. How to workaround it
3. How this apache restart is triggered ?
4. How to get it to do the restart in non-peak hours !
Cheers
Ian
|
Threaded Mode
