Hi all, I just thought I’d report here that for a little while I’ve been having some pretty serious issues, which I’ve gone through a series of tests and so on, finally tracing it back to a fairly unbelievable cause: An partially downloaded file read by SABNZBD at startup hard crashes TrueNAS without ANY logs, reboots and does the whole thing again every time docker starts up.
I hadn’t realised it was doing this all night and this was what caused my most recent post about losing data on ZFS. To fix it I’d have to stop docker at boot, delete the ix-systems folder, set up every docker container again and each time I set up SABNZBD it would crash some time after.
It feels great to have finally solved it but I have to wonder how this is possible. I really don’t think a corrupt file or whatever it was should be able to hard crash a system via a docker image.
I went to log a ticket in Jira just so the ix-system folks are still aware like I used to and just like the last time I went to do that I didn’t have access any more - are we still meant to log tickets this way? I can’t really even view the things that say they’re public as it keeps popping up and saying it logged me out.
I was monitoring SABNZBD at the time and the last log it gave me was: 2026-03-05 19:22:10,403::DEBUG::[filesystem:931] [sabnzbd.nzbstuff.remove_admin] Deleting file /incomplete-downloads/filename
I have deleted that whole folder and not had a crash since.
This is on 25.10.2. I’m sorry, but there are literally no logs, believe me I’ve been hunting for them, not even a crash dump. My guesstimate is some kind of bad kernel bug, but what do I know.
If nothing else, perhaps one day this will help someone.