Had been having write issues (red balls) and have reconstructed disks a few times in the past few weeks after they red-balled.

 

Upgraded from 5.0-rcsomethingorother to 6.1.3 just now.  Array came up and immediately went into parity check (I didn't do anything).  On the dashboard the parity drive has a red thumbs down and is saying it has 560 reallocated sectors.

 

First, should I stop the parity check?

 

Second, why did my parity drive suddenly throw SMART errors after an upgrade?

 

Third, is the parity drive toast?  I ask because unraid says its overall SMART health "passed."

 

Thanks.

5.0 never had smart monitoring. If you werent monitoring it yourself you'd never know about them.

 

6.x has smart monitoring and notification of said items.

 

In other words, you had these issues before the upgrade but never checked for them.

Understood.  Some research leads me to believe that if that number climbs consistently, I should worry, but if it remains the same for many reboot cycles, I should be OK. 

 

I have a spare drive on the shelf... Should I swap it out and attempt to have the mfg warranty the existing parity drive, or am I reasonably safe?

As I assume you know, reallocated sectors aren't necessarily bad ... modern drives are in fact designed to reallocate defective sectors.    What's BAD is "pending" sectors ... these are sectors that are known to be bad, but can't be reallocated because the drive can't read the data from them to reallocate it.  Even worse, when a good read is finally accomplished; or new data is written to that sector and it works okay, the SMART system often simply resets the status and does not reallocate the sector -- resulting in sectors that can toggle back-and-forth between being pending or not.    These SHOULD simply be reallocated so they won't be used anymore, but for some reason that's not the way it always works.

 

Nevertheless, 560 is a fairly large number of reallocations ... I'd be inclined to simply replace the parity drive and relegate your current drive to backup status.    Much array data in most UnRAID systems is fairly static, so drives that you may not want to use for a lot of frequently changing data work just fine as backups ... once the data's backed up and verified, it's not likely to ever be accessed again.

 

I replace all drives that have > 10 reallocated sectors and move them to my backup pool.

 

 

As for why it started a parity check, if during your upgrade you restored a copy of config/super.dat that was taken while the array was started, then unRAID thought that you had not shutdown properly.