Warwick

Update on this. @itimpi was correct, I had to first stop the array, then reset the array configuration, preserving all current assignments, unassigned the SSD from the array, reassigned it to the "timemachine" cache pool and then started the array. I checked the "Parity is valid" option at this point because the SSD had not yet been a part of the array fully (when it was a part of the array it was still going through the zeroing process so the parity drive not have had any changes written to it). Performed a parity check to make sure and it came back all clear. Restarted the array, everything stayed where it was. Rebooted and everything stayed where it was. Formatted the SSD to btrfs and rebooted, everything's still where it is so looks to be completely resolved. Will update OP and title.

I had not and this would match up with why it's being added straight back. Thanks, will try and report back.

UPDATE: Should've removed the disk from the array properly using the method detailed here for "Removing a data disk(s)". After following this procedure the SSD drive is no longer a part of the array configuration and is not being pulled back in to the array upon reboot. Unraid version 6.9.2 I've seen topics from several years ago which reference setting the disk as Cache in IE but I feel this is a little different and assume that problem has been fixed since IE is no longer a browser. Installed a new 2TB SSD to use as a destination for Time Machine back ups. Initially wasn't sure of the best way to do this so I added it to the array so that I could create a share and limit it to this disk.. Realised something was up when the preclear would stop itself continuously and after checking online realised that was a mistake, SSDs should only be used as cache drives/unassigned and never in the array. Noted. Removed it from the array, added it to a new cache pool, shared the drive over SMB as a time machine share. Everything was working! Rebooted server today and when it came up the drive was out of the tm cache pool and was being precleared. Immediately stopped the array, removed it as a disk in the array, lowered the number of slots back down (which I had done previously when moving it to its own cache pool), added it back to its cache pool and started the array. Working again (albeit with a drive that had been partially cleared and needs to be reformatted). Tried rebooting again and the drive was added back to the array and the preclear was beginning. Instead of trying to trouble shoot this while I watch the writes on my SSD climb after each reboot I'd rather ask for suggestions as to the best way to tackle this. My gut feeling is that a config for the array is not being updated somewhere and after reboot it's reverting back to that, but I'm not entirely sure where this would be/how it should look. Would also prefer to avoid redoing the entire parity by creating a new config as a troubleshooting step, but if that's the solution then so be it. xev-diagnostics-20211005-2059.zip