I’m running trueNAS core. I’m running a pool on a RAIDZ2 6 drive configuration.
I just replaced an inferior 6 port SATA card with an 8 port SAS HBA in IT mode. That went well. I had to replace 2 drives and the resilver took about 5 hours each. I then did a scrub and reset the error count. I also deleted or restored a dozen files from an earlier zpool status -v. The only errors remaining in a filesystem nested under another dataset (a timeserver).
I decided to destroy/delete the errant filesystem. Using the GUI, the system would spin the magic blue circle, then hang, and reboot. After several tries at the child filesystem I tried the parent dataset. It also caused the reboot, after deleting all the sibling filesystems, leaving the errant child intact. One file CLI attempt resulted in the same reboot.
As a side note, the pool shows up as ONLINE (unhealthy). I ran a SMART short test on 3 of the drives with a ‘SUCCESS’ result.
I’m considering trying to go nuclear on the pool but I’m not sure of the steps involved (probably export with data wipe?), recreating the pool & datasets, and using import from a USB connected backup drive (NTFS).
Any recommendations?