Am 03.06.2024 um 11:04 schrieb Christian Mack (christian.m...@uni-konstanz.de):
Hello

Am 02.06.24 um 22:32 schrieb Paul Leiber (p...@onlineschubla.de):
Am 29.05.2024 um 21:52 schrieb Paul Leiber:
[...]
For finding the cause of the above behavior, I am often changing configuration, which requires me to restart SOGo. However, when restarting SOGo with "systemctl restart sogo", sogo is not functional anymore and the following log entries come up:

May 29 21:41:02 sogod [2058]: [WARN] <0x0x557824d99d10[WOWatchDog]> listening socket: attempt 1 failed May 29 21:41:03 sogod [2058]: [WARN] <0x0x557824d99d10[WOWatchDog]> listening socket: attempt 2 failed May 29 21:41:04 sogod [2058]: [WARN] <0x0x557824d99d10[WOWatchDog]> listening socket: attempt 3 failed May 29 21:41:05 sogod [2058]: [WARN] <0x0x557824d99d10[WOWatchDog]> listening socket: attempt 4 failed May 29 21:41:06 sogod [2058]: [WARN] <0x0x557824d99d10[WOWatchDog]> listening socket: attempt 5 failed May 29 21:41:07 sogod [2058]: [ERROR] <0x0x557824d99d10[WOWatchDog]> unable to listen on specified port, check that no other process is already using it

A reboot leads to a functional sogo system (besides the issues mentioned above), but takes quite some time. In the past, I didn't observe this behavior.


FYI: With the current nightly, restarting sogo with systemd is working again.


I do not think that this has to do with the nightly build.
I only get those error messages, if stopping sogo didn't stop all workers, before it starts sogo again.
Then the sockets are still in use by an outdated worker.
This especially happens with long running workers, used for e.g. ActiveSync. Therefore I stop sogo, then test with ps command if all workers are really stopped, before starting it again.

Yes, I was wrong, yesterday I unfortunately observed this issue again with the same nightly. I can also confirm that the cause is that not all SOGo processes are stopped when stopping SOGo via systemctl. However, even using kill -9 for killing all worker processes I could find with ps didn't help. I must have overlooked something.

Best regards,

Paul

Reply via email to