Hoping for some assistance here.  In my setup (6.12.14), I have a two-drive zfs cache pool, both are SSD's.  These drives house the appdata (plex, arr's, etc) and I back that up to a zfs backup drive (just to be safe). Every couple of months I will get a warning about the cache pool and one of the drives is in a degraded state.  I replace the drive and 2-3 months later, the other one will go.  

 

Has anyone encountered this kind of issue and if so, was there anything you did to fix it?  I can't imagine that having to replace an ssd every couple of months is a normal thing!

 

Diagnostics are attached.

casatower-diagnostics-20250118-0542.zip

Looks more like a power/connection issue, I would recommend swapping the SSDs to the onboard SATA ports so that TRIM is supported, use new cables, then retest.

On 1/19/2025 at 1:57 AM, JorgeB said:

Looks more like a power/connection issue, I would recommend swapping the SSDs to the onboard SATA ports so that TRIM is supported, use new cables, then retest.

Interestingly enough, after I posted this, I also noticed that for some reason I had a 'data' folder going to the cache pool (cache to array) which was essentially just my timemachine backups...a good reason to be up to nearly 13M writes, I suppose.

 

I had planned on moving everything back to the onboard SATA ports.  The HBA was because my previous board didn't have enough room.  Now, with 6 ports plus two m.2, the HBA isn't needed.  I will say though, that the cache pool drives were the only ones to ever be affected.

 

I've read that ZFS pools have an option for trim, is this where I should be turning it on or through the system settings?

Auto trim is enabled by default for zfs pools, but still recommend doing the regular trim once a week or so. Settings - Scheduler