I've never had errors on a parity check. Just the last one I ran, I had 352 errors. Since I never had errors, is that a lot and where does unraid write these errors to disk? Or should I ask, where is the log for the parity check?

 

p.s. the system has rebooted since the check.

If it was a correcting parity check, it wrote the corrections to the parity drive, to bring it back in sync with the data drives. If it was a non-correcting check, then it notated the errors, but did not change any data. The only logging is in the syslog, so if you rebooted without saving a copy, the record is gone. 1 error is 1 too many, I would get smart reports on all your drives, then run a non-correcting check, then pull smart reports on them all again and compare. There is no practical way to know what file(s) or disk(s) were involved in the errors, but you can make an educated guess based on the circumstances. If the array was interrupted in some way without being shut down properly, the error is most likely on the parity drive, and a correcting check is the appropriate way to recover. Also, if something operated on the /dev/sd* device, it will cause parity errors, and if your files are ok, the correct action is a correcting parity check. If, however, you have a failing data drive, you would be better off to pull the drive and reconstruct it from parity, instead of writing the "corrected" bits to the parity drive. Theoretically parity should never be updated from a bad drive, but I've seen enough anecdotal evidence to think otherwise.