cron process deletes videos prematurely
Up to the latest commit (4304fa9a47dfee7d369ce3aacf4074aafe8e2e49
) in master branch, the cron process run by Shinobi is automatically deleting videos after approximately one hour even if it retention time is set globally and per camera to be several days, and the max storage space is a very large number. I have deduced this by running Shinobi with the cron process stopped for about 18 hours and I now see a full list of videos recorded for each camera and the amount of disk spaced use continues to increase (as one would expect if you're recording video without deletion to disk).
My configuration is: Shinobi installed from this git repo on an lxc container (Ubuntu 17.10 for container and host OS) with about 3TB of external storage space located on a zfs pool. I ran the ubuntu.sh installer script after cloning the git repo and configured Shinobi to use sqlite as the database backend, as well as let it install node.js and the dependencies.
Here are a couple threads that report this issue on reddit:
- https://old.reddit.com/r/ShinobiCCTV/comments/9tl4ga/recorded_videos_keep_disappearing/
- https://www.reddit.com/r/ShinobiCCTV/comments/9fm6jd/shinobi_deletes_detection_videos_after_around_an/
It isn't ideal to have to run Shinobi with cron disabled because I believe this would mean old videos will never be automatically deleted. It will be up to a human to login and prune old videos which is a waste of time. If there is an auto delete process, it should work and it should work reliably. Obviously there is some aspect of the configuration that isn't being honored by the Shinobi cron process as far as the way old videos are deleted off the system.