v1.37.50-latest Continuous Events still "broken" about zoneminder HOT 8 CLOSED

I believe you have crashing. The events with no Duration are an indicator that the zmc process crashed before finalising the event. Please check yours logs.

from zoneminder.

Oh, yes you are absolutely right. OOM kills.
So, but this is also new with 1.37.50...
zoneminder zmdc[10878]: INF ['zmc -m 9' crashed, signal 8]

Well, this explains a lot.

Jan 25 09:20:12 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 09:57:22 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 10:06:54 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 11:02:18 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 12:14:53 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 13:26:39 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 14:42:39 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 15:55:22 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 17:07:00 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.
Jan 25 18:18:00 zoneminder systemd[1]: zoneminder.service: A process of this unit has been killed by the OOM killer.

I enabled Debug Logging for zmc will try to look into it

from zoneminder.

I just downgraded to 1.37.49~20240116.22-jammy and as far as I can tell everything works flawlessly again, even the monitors that just would not create an event started working instantaneously.

I will add my remaining monitors to the new zm instance and report back tomorrow.

from zoneminder.

1.37.49~20240116.22-jammy works perfectly.

Guess when I switched back to .49 ;)

Can I help with some log files?
I can create a second instance with 1.37.50 for troubleshooting purposes, if that helps.

from zoneminder.

from zoneminder.

from zoneminder.

Just as a minor addition, I just upgraded to latest commit (0e88d76) today. ZM now has a very slow mem leak that wasn't there in .49 (Don't recall which commit specifically). I am usually hovering around 4-7GB RAM for days on end, after a reboot and 5.5 hours of commit 0e88d76, I am at 14.5GB and climbing slowly about 1-3 GB an hour.

Edit: 7.25 hrs and now at 17.25GB

Edit 2: At 7.5 hrs and using htop, I saw it was certain monitors eating memory. I changed those monitors max image buffer size from 0 (unlimited) to a set number and it has slowed/stopped the mem leak. I'm back down to 5.1 GB

Edit 3: This morning I checked memory and it is back up to 14GB. On one of the monitors, the estimated max RAM use was around 151 MB, this morning it was taking up 12.5% (roughly 4 GB) of 32 GB. IDK if that means the max image buffer size isnt being enforced or possibly memory needs to be freed from the proc? Restarting hte high mem use monitors frees the memory up and then it slowly builds.

Fixed in 5285c6e

from zoneminder.

Seems stable as of v1.37.51.
Thank you for your hard work @connortechnology !

from zoneminder.