View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0005256||SOGo||ActiveSync||public||2021-02-06 06:54||2021-07-26 06:48|
|Platform||[Server] Linux||OS||Ubuntu||OS Version||16.04 LTS|
|Summary||0005256: Sogo EAS running out of memory despite SxVMemLimit|
We run sogo EAS for around 60 devices balanced between two machines with 4vCPUs and 16GB RAM and 8GB swap. The relevant configuration is this:
WOPort = "0.0.0.0:20000";
We constantly get OOM errors from sogo processes eating all available RAM:
I understand EAS is resource intensive but this happens with only few (4-6) devices connected and it looks like sogo doesn't follow SxVMemLimit to terminate its processes once they have reached the limit.
Below I've attached some context from the sogo logs as well.
|Steps To Reproduce|
We use a multitude of AS clients and it's difficult for me to pinpoint the exact scenario but given enough time sogo will end up with processes that are never relinquished and keep growing infidelity.
Feb 06 11:11:42 sogod : [WARN] <0x0x55d268406790[WOWatchDogChild]> pid 64745 has been hanging in the same request for 9 minutes
|Tags||No tags attached.|
Same problem here.
Also sogod ignore SOGoCacheCleanupInterval, because in such sutuation it NOT clean cache and memory leaks continues.
Additional info to 100% reproduce the bug.
So, finally way to reproduce the bug:
Do you need any additional info?
|2021-02-06 06:54||tomushkin||New Issue|
|2021-07-06 00:55||akimov||Note Added: 0015351|
|2021-07-07 22:42||akimov||Note Added: 0015353|
|2021-07-08 04:05||akimov||Note Added: 0015354|
|2021-07-08 04:30||akimov||Issue cloned: 0005353|
|2021-07-08 04:30||akimov||Relationship added||related to 0005353|
|2021-07-26 06:48||akimov||Note Added: 0015401|