Last night I upgraded a 4.7 system to a 5.0 stable version, and having not run a parity check in a while did one. Now, one of the drives is showing as a red ball. Up to now, I have not had any problems in over a year. I do not necessarily attribute this to 5.0, but at this point do not know what to do next. Any help would be great.

 

Thanks,

 

Lev

syslog.zip

Did you bump a cable while removing your flash drive for the upgrade?

 

Not that I am aware off. However, I did have to manipulate the case. It's large - Norco 4220, so conceivably it could have dislodged something.  Should I reseat all the cables? If yes, how should I proceed?

Don't forget to reseat all drives.

Once the drives are reseated and the connections are checked, how will I know that the drive is working? Will I have to unassign / reassign it?

Once the drives are reseated and the connections are checked, how will I know that the drive is working? Will I have to unassign / reassign it?

 

Most likely.

Yes, once a drive is disabled, the process to force a rebuild on itself is to Stop the array; unassign the drive;  Start the array;  Stop the array; re-assign the drive to the same position; and then Start the array.

 

I reseated the drive. Unassigned / reassigned and rebuilt the array. At that time the drive "appeared" OK. Then, I ran a parity check. The drive failed again with some 750 errors. I did not capture syslog at that time and just swapped the drive. Now rebuilding it on a new drive. Once rebuilt I'll rerun parity check. 

Last night I upgraded a 4.7 system to a 5.0 stable version, and having not run a parity check in a while did one. Now, one of the drives is showing as a red ball. Up to now, I have not had any problems in over a year. I do not necessarily attribute this to 5.0, but at this point do not know what to do next. Any help would be great.

 

Thanks,

 

Lev

 

r.e. "... having not run a parity check in a while did one ..."  ==>  Did you not do this BEFORE the v5 upgrade?   

That thought has been going through my head. No, I did not run parity check before upgrading.

Well, it's obviously too late to change that ... so just let the rebuild finish and then do a parity check and hopefully all will be well.

 

If you have backups, however, I'd compare the files on that disk to your backups just to be safe.    Or if you have checksums [MD5s, SHA1, etc.] run a verification on those.

 

Well, the new drive restored and parity check ran fine. 'guess that the old drive was bad, but I am surprised that it never showed itself as bad prior to running parity check. Is that something that is a normal behavior? Will unRaid declare a drive bad only during parity check? If I do not perform one for a long time I could potentially be sitting on a few of them.

Well, the new drive restored and parity check ran fine. 'guess that the old drive was bad, but I am surprised that it never showed itself as bad prior to running parity check. Is that something that is a normal behavior? Will unRaid declare a drive bad only during parity check? If I do not perform one for a long time I could potentially be sitting on a few of them.

Unraid will only fail a drive if a write to that drive fails. If a read fails, unraid will spin up all the drives, calculate what should be there, and write the calculated value back to the drive. If that write succeeds, unraid increments the error counter for that drive and moves on. If the write fails, the drive is immediately red balled, and all further operations to that disk are virtual, and only update the parity drive. If a second drive fails before you can rebuild the first, you will lose all the data on both failed drives. It is really important to find silently failing drives before you get two at once. Regular parity checks will force all drives to be read, and hopefully catch errors sooner rather than later.

 

TLDR; unraid will only fail a drive if the bad spot is used. Unread drives can silently fail, and tend to be caught during a parity check.

That is a VERY good thing to know, and not at all verbose. Checking to make sure that the drives are good cannot be overstated. I think that I will make it a priority to perform a monthly parity check. After all, that is the reason why a lot of us went with unRAID - security.

 

Thank you,

 

Lev