View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0001865||SOGo||Packaging (Debian)||public||2012-07-09 09:30||2012-07-25 11:40|
|Target Version||1.3.17||Fixed in Version||1.3.17|
|Summary||0001865: daily log rotation stops sogo due to old process taking too long to exit|
I have recently switched from a manual installation of SOGo to using the provided Debian packages on Debian squeeze.
Since switching, most days, SOGo stops and fails to restart when cron runs logrotate. The logrotate executes "/etc/init.d/sogo restart", but usually the old process still has the listening tcp socket open when the new one tries to start, so it exits after 5 tries to open the socket.
Old (rotated) file:
Jul 09 06:22:21 sogod : |SOGo| starting method 'REPORT' on uri '/SOGo/dav/mdyson/Calendar/personal/'
New log file:
Jul 09 06:25:23 sogod : version 1.3.16 (build @vizzini.inverse.ca 201206071319) -- starting
|Tags||No tags attached.|
Just to confirm we are experiencing the same problem with Debian(squeeze).
Thanks for reporting this, this behavior was fixed long ago in the rpm packages by using copytruncate to avoid restarting sogo, but we were not using the same logrotate script for debian packages.
This is now fixed and will be available in the next release.
Meanwhile, you might want to fix the logrotate file:
|2012-07-09 09:30||maxb||New Issue|
|2012-07-20 04:00||axelr||Note Added: 0004182|
|2012-07-23 08:40||ludovic||Status||new => assigned|
|2012-07-23 08:40||ludovic||Assigned To||=> jraby|
||Note Added: 0004208|
||Status||assigned => resolved|
||Resolution||open => fixed|
|2012-07-23 14:01||ludovic||Fixed in Version||=> 1.3.17|
|2012-07-25 11:40||francis||Target Version||=> 1.3.17|