Closed member7me closed 1 year ago
Screenshot from other, but similar server with similar configs. It's test server, where ejabberd wasn't updated, but installed as fresh from Debian 12 distro.
journalctl -u ejabberd.service
example of rebootings
Did you follow the steps in https://docs.ejabberd.im/admin/upgrade/ after the upgrade? Eg. If using SQL etc
Did you follow the steps in https://docs.ejabberd.im/admin/upgrade/ after the upgrade? Eg. If using SQL etc
I don't use sql, just mnesia.
You probably can fix this temporary by removing WatchdogSec=30
from ejabberd.service, looks like for some reason systemd doesn't get "pings" from ejabberd.
I noticed, this bug exists only when ejabberd is under load with active connected members on board. If to upload same environment on other server where there is no active connected members, there will be no errors and service crash.
Environment
Bug description
Hello. Similar bug to https://github.com/processone/ejabberd/issues/4048 and https://github.com/processone/ejabberd/issues/4054 Ejabberd 21.01 worked fine. After updating Debian to version 12 (bookworm) ejabberd also updated to 23.01. After that service started to crash. Fully randomly. Since 5-60 minutes after starting. Firstly crashes ejabberdctl, i.e.
ejabberdctl status
stops answering. All this time service still working, web admin also working.ejabberdctl live
doesn't start. In few minutes after that full ejabber service crashes.Service ejabberd status
gone into inactive status. Webadmin crashes also. Whole service is restarting. All this happens without ending at intervals of 5-60 minutes, randomly. Config file didn't change.error.log
andejabberd.log
are empty, I mean - only with standart warning messages. No errors. At all. Also there is no CPU or RAM high usage.Only downgrading to the old version Ejabberd 21.01 solves the problem.