Occasional watchdog timeouts on netcams
I have motion setup to monitor a bunch (18) of network cameras. Occasionally a thread dies with a watchdog timeout. Unfortunately this causes all the viewers watching the motion feeds to stop so this is a significant hassle. Any ideas why this is happening and what can be done to prevent it?
The webcams are various Axis models.
I don't believe this is caused by loss of connectivity. When I see loss of connectivity I get a "Video signal lost - Adding grey image" message which does not cause problems with the viewers.
Errors logged are:
Oct 10 23:52:05 server motion:  Thread 14 - Watchdog timeout, trying to do a graceful restart
Oct 10 23:52:44 server motion:  netcam camera handler: finish set, exiting
Oct 10 23:52:44 server motion:  Motion thread 14 restart
Oct 10 23:52:45 server motion:  Resizing pre_capture buffer to 1 items
Oct 10 23:52:45 server motion:  Resizing pre_capture buffer to 2 items
| Motion version:
| ffmpeg version:
|| ffmpeg, mysql, postgresql
| Server OS:
|| CentOS 5, 2.6.18-308.4.1.el5PAE
- 11 Oct 2012