Capture Daemon Errors Since Upgrade...

[jsylvia007]

Howdy all! I'm running Ubuntu 18.04 LTS (fully updated) and recently (yesterday) upgraded to ZM 1.34. The upgrade went perfectly fine, and after some tuning with zmNinja and Home Assistant, everything is working as before with one exception...

All of my IP Cameras are Q-SEE 4K cameras. I have a "2-monitor" setup for each camera. One is using a lower resolution (640x360) for MODECT and then the 4k resolution for NODECT, but linked to the monitor.

Since the upgrade, I'm seeing the following errors rather frequently, which I think might be causing issues with capture...

Code: Select all

Date/Time 		Component	PID	Level	Message														File			Line
2020-02-10 15:38:22	zmdc		2291	ERR	'zma -m 11' exited abnormally, exit status 255									zmdc.pl	
2020-02-10 15:38:22	zma_m11		17465	ERR	Got empty memory map file size 0, is the zmc process for this monitor running?					zm_monitor.cpp		574
2020-02-10 15:38:22	zmdc		2291	ERR	'zma -m 11' exited abnormally, exit status 255									zmdc.pl	
2020-02-10 15:38:22	zma_m11		17458	ERR	Shared data not initialised by capture daemon for monitor Back Yard - Deck					zm_monitor.cpp		498
2020-02-10 15:38:21	zmwatch		2514	WAR	Restarting capture daemon for Back Yard - Deck, time since last capture 10 seconds (1581367101-1581367091)	zmwatch.pl

The above is specifically for Monitor ID 11, but sometimes it's for 9, or 13, or one of the other 4K feeds. It doesn't appear to affect the low resolution feeds.

Can someone point me in the direction on where I need to look to stop the capture from crashing?

I am running this in a VM on ESXi, but the VM has 8-cores, and 64GB memory. I have a total of 14 monitors (across 7 cameras).

A typical monitor is configured as an FFMPEG source, and the camera is configured to pass x264. The cameras do support x265, should I change the cameras over to this?

Any help is appreciated, I wasn't seeing nearly so many drop-outs on 1.32.

Thanks!

[bbunge]

While I do not run Zoneminder on a VM I did an upgrade on an old production box the other day from 1.30.4 (Bionic/LAMP). Right after the upgrade the ZM Console was "messy" until I rebooted. Then I got into the console and set the timezone, something new since 1.32.
You have plenty of RAM it seems but make sure the docker is allocating enough of that to shared memory.

[jsylvia007]

Yea, same here, I've rebooted a couple times though... My shm is at 74%, and this box is dedicated to ZM.

[dardar]

I'm seeing something similar (on a physical Xubuntu box (18.04.1) running zm 1.34.0), that I also think may be a new issue since the upgrade. I have two Amcrest IP cameras with similar setups to yours (although I don't use the low res feed). I have been using VLC to monitor dropped video frames on my camera streams, and my problematic camera (where I get a lot of "Restarting capture daemon..." messages from zmwatch) drops many more frames than the camera that doesn't have any problems (e.g., lost 46 frames out of 6000 displayed just now, compared to 1 or 2 frames for the good camera). Do you see dropped frames from the cameras having problems in your system?

Not sure whether my problem is a WiFi or other networking issue or something with the camera itself, although nothing has changed on the hardware or network side since the ZM upgrade. Is there a way of buffering streams in the monitor configuration that would compensate for lost video frames and keep the daemon from restarting? Is that even likely to be causing this problem? BTW, my /dev/shm is 4%, so I don't think I have a memory problem. Plus, it's only impacting one camera in my system.