- Retest
- Minor
In my Unraid box I used to have an SSD cache disk. Some months back I removed the cache disk from cache duty, let it be mounted as just a disk directly on the filesystem. Ostensibly at this point I left cache enabled, just with no disks, mover still appears to be scheduled to run. Everything has been fine with the server for all these months. Then a week back I upgraded to 6.8.0 and it started randomly crashing ~an hour after booting, completely unresponsive over the network. I didn't even really notice at first - I have been quite busy and haven't been using my server much.
Router errors suggested to me I had some sort of network issue, after tearing my hair out reconfiguring my network, buying a new NIC (I wanted a decent dual NIC for this server anyway so no biggie) and nothing really improving I started to really focus on Unraid.
It turns out that what has been happening is that ostensibly mover still running is causing an unhandled null pointer deference (somehow) then the CPU is stalling, network goes down as a result, and it never recovers.
I don't know for sure that this issue was introduced by 6.8.0 explicitly, but I'm *fairly* sure I didn't have this issue before - because it would have driven me bananas until I'd resolved it as it did in this case. I never tested any pre-release versions so I also don't know at what point it happened.
Basically the workaround aka fix for me was to set number of cache disks to none (it was still set at 1), and disable cache in settings (I believe it was set enabled in two places, weirdly) and it's been happily running since for over 24 hours.
Some sort of check to see if there actually is a cache disk before executing mover would probably help, or maybe a warning when users have an incomplete cache setup? I don't know for sure but for me at least it was definitely a thing.