View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0003026||SOGo||Backend General||public||2014-12-15 08:39||2015-01-10 05:19|
|Reporter||Sigurd Holter||Assigned To||ludovic|
|Platform||[Server] Linux||OS||RHEL/CentOS||OS Version||7|
|Fixed in Version||2.2.14|
|Summary||0003026: WOWatchDogRequestTimeout stopped working with 2.2.11a|
After upgrading to 2.2.11a WOWatchDogRequestTimeout = 120 is not honored any more, and processes are killed after 10 minutes.
|Tags||No tags attached.|
That code has NOT changed since March 2011!
I've just tried with the following values:
and Apache proxy configuration:
ProxyPass /Microsoft-Server-ActiveSync http://127.0.0.1:20000/SOGo/Microsoft-Server-ActiveSync retry=0 connectiontimeout=5 timeout=3600
and it works as expected. The daemon was running well after 10 mins.
I get repeated occurences (38 times between 0330 and 1030 today) of this sequence :
Dec 16 09:50:22 sogod : |SOGo| starting method 'OPTIONS' on uri '/SOGo/Microsoft-Server-ActiveSync?User=benjamin&DeviceId=B4389A49654394ACF8174149614DDA60&DeviceType=WP8'
I have in sogo.conf :
WOWatchDogRequestTimeout = 120;
I have also added to the startup script (to see if it helps) :
DAEMON_OPTS="-WOWorkersCount $PREFORK -WOPidFile $PIDFILE -WOLogFile $LOGFILE -WOWatchDogRequestTimeout 120"
ps aux shows :
sogo 18737 0.0 0.0 429748 16364 ? Ss des.14 0:01 /usr/sbin/sogod -WOWorkersCount 30 -WOPidFile /var/run/sogo/sogo.pid -WOLogFile /var/log/sogo/sogo.log -WOWatchDogRequestTimeout 120
... and most importantly of all - the setup and sogo.conf file have not changed for a month or more.
greping sogo.log shows NO "has been hanging for 11 minutes" etc., it always exits after 10 minutes.
What else can be causing this ?
The child existed with a segfault - signal 11. It was NOT terminated by the partent.
Attach using gdb to a "hanging child' and produce a stacktrace when that happens:
backtrace.log.tar.gz (989,279 bytes)
Backtrace uploaded. Note that crash occured after about 20 minutes this time.
Same crash occurs after 10 minutes when same user uses Outlook 2013 with ActiveSync - same account/server etc.
Could you be running out of RAM?
I thought it was a memory issue, too, at first and set SxVMemLimit = 8192
But no, doesn't seem that way. There's still 4+ GB free when SOGOd process claims about 2 GB. And no OOM messages either.
It's the same server/setup I've been using since August, and it runs fine in production for other users on Thunderbird etc.
Try with 2.2.13 with the new SOGoMaximumSyncResponseSize configuration parameter.
Sorry, I see I should have responded to this issue instead of posting a new bug at 0003051.
But anyways, the issue persists in 2.2.13 with SOGoMaximumSyncResponseSize=10240.
Backtraces have been uploaded to 0003051.
|2014-12-15 08:39||Sigurd Holter||New Issue|
|2014-12-16 11:38||ludovic||Note Added: 0007846|
|2014-12-16 11:38||ludovic||Status||new => closed|
|2014-12-16 11:38||ludovic||Assigned To||=> ludovic|
|2014-12-16 11:38||ludovic||Resolution||open => unable to reproduce|
|2014-12-16 12:34||Sigurd Holter||Note Added: 0007850|
|2014-12-16 12:34||Sigurd Holter||Status||closed => feedback|
|2014-12-16 12:34||Sigurd Holter||Resolution||unable to reproduce => reopened|
|2014-12-16 13:22||ludovic||Note Added: 0007851|
|2014-12-18 05:42||Sigurd Holter||File Added: backtrace.log.tar.gz|
|2014-12-18 05:44||Sigurd Holter||Note Added: 0007875|
|2014-12-18 05:44||Sigurd Holter||Status||feedback => assigned|
|2014-12-18 05:46||Sigurd Holter||Note Added: 0007876|
|2014-12-18 05:46||Sigurd Holter||Note Edited: 0007876||View Revisions|
|2014-12-18 08:40||ludovic||Note Added: 0007878|
|2014-12-18 08:55||Sigurd Holter||Note Added: 0007879|
|2014-12-30 12:45||ludovic||Note Added: 0007946|
|2015-01-05 06:50||Sigurd Holter||Note Added: 0007988|
|2015-01-09 06:22||Christian Mack||Relationship added||related to 0003051|
|2015-01-10 05:19||ludovic||Status||assigned => resolved|
|2015-01-10 05:19||ludovic||Fixed in Version||=> 2.2.14|
|2015-01-10 05:19||ludovic||Resolution||reopened => fixed|