If the location on disk of a parity check error could be reported in the syslog it may help to isolate on which drive the error lies. This may not be required with p+q parity.

I think the first 100 parity sync errors are reported in unRAID 5.0b3, in earlier versions it was the first 20. It should show up in the syslog.

 

/* called on a sync error */
void md_sync_error(mddev_t *mddev, sector_t sector)
{
       mdp_super_t *sb = &mddev->sb;

       sb->sync_errs++;
       if (sb->sync_errs <= 20)
               printk("md: parity incorrect: %llu\n", sector);

}

OK. I didn't realize those were sectors.

I think that those are sectors relative for the partition, but could easily be mistaken. I don't know if the first sector is 0 or 1.

If the location on disk of a parity check error could be reported in the syslog it may help to isolate on which drive the error lies. This may not be required with p+q parity.

 

It has been available since unRAID 4.5.5 and it can help a bit, unfortunately it only allows you to reduce the region on each drive you have to search for errors.  In this post:

 

http://lime-technology.com/forum/index.php?topic=10364.msg98580#msg98580

 

I used the reported locations as a guide to what portion of each drive I needed to look at, then I used the md5sum utility to calculate a digest for that region on each of my drives, I did this multiple times and by doing so was able to isolate the drive that was causing occasional parity errors.  Since replacing that drive I have not had a single parity error.

 

Regards,

 

Stephen