Potentially a silly question, but if you don't ask you'll never learn...

 

I recently switched my SSD Cache drive to use zfs instead of btfrs and since then I've noticed that my CPU usage will top out at 100% for long periods of time, seemingly at random. I'm running a fairly old but otherwise reliable N54L so it's fairly noticeable and am trying to troubleshoot a separate php issue since upgrading from 6.11.

 

I've used the CLI command htop which shows it's usually a docker run command that's the culprit. Looking at my docker settings I can see the data-root is still set up to use to btrfs vDisk, should I change this either xfs or directory now the cache drive it's on is no longer btrfs?

 

I haven't seen this accounted for in any of the conversion tutorials or docs, so am not sure if there's a best practice.

Using a btrfs vdisk on top of zfs should not cause any issues, could be a specific container causing the CPU usage, but you can try recreating docker image using xfs to see if there's any difference.

I think you're hitting the capacity of the N54L - It's old and under specced for what UNRAID can now throw at it.

I'm not sure exactly when you observed sustained peak CPU activity.. but I saw similar behaviour with my N40L when throwing certain VM's and Containers at it.
After I upgraded the hardware, (HPE ML310e Gen8v2) I haven't seen this occur.. (I'm using xfs for my array, btrfs for my vdisk; and zfs for my cache pool) 

I'm having the same problem as well, can't seem to find the issues, Cpu running high, zfs ram is maxed. Sever starting to get hot. Its does seem if I stop all dockers then it start to act normal. I started plex and it would start to peak with usage. I also have a error from fcp "out of memory errors detected on your server".

conwayserver-diagnostics-20240108-1005.zip

Was this ever figured out?