- Joined
- Jul 12, 2022
- Messages
- 3,222
With ECC RAM, iirc, either the single-digit bit flipped is restored or the system goes panik mode so it should be unlikely for that to have been the cause.Here is the thing though: if I have a mirrored vdev irrepairable data corruption shouldn't happen unless both disks failed on the same bit of data or there was some other hardware problem, right? Since both disks' SMART tests are still looking good I think that I'll start out doing memtesting once I got my pool repaired (just now kicked off the scrub).
I hope that it's RAM, because that I can easily swap out (I suppose that ECC can only do so much with bad RAM). And in that case I'll replace ada1 with a fresh disk so that they're less likely to both fail around the same time.
You can memtest your RAM if you suspect hardware is at fault there.
You can get errors on a disk part of a mirror, and they get fixed if the other copy has the good data when you scrub the pool.