Edit:
After I rebooted my TrueNAS machine and replaced one of the “corrupted” disks with a new one I got the message, that the pool is now healthy except for the replaced hdd (meaning one of the two drives is no longer faulted, according to TrueNAS) I don’t know if the reboot resetted the zfs information about the faulted drives or not.
1 Like
A lot of trial and error. I originally bought used SAS drives on eBay to use in my HL15. When I started having issues it was the quick answer. Luckily the seller took them back and I bought brand new Seagate Exos drives. When the same thing happened with with new drives, I knew something else was up.
It was ultimately the SMART data that lead to me down the cable route. On the SAS drives, there’s a SMART page about the controllers. I captured the SMART data periodically and I noticed that two values were incrementing: Invalid DWORD count and Loss of DWORD synchronization. A little googling lead me to other forums where cables or HBA firmware ended up being the problem.
On a hunch, I used Seagate utilities to slow down the link speed from 12gbps to 6gpbs. At 6gps, I was no longer able to replicate the problem. At 12gpbs, the problem returned within a few hours. This was enough evidence for 45HomeLab to send me a new set of cables. I swapped them out and haven’t had this issue since.