I went for it anyhow.

 

Stole the 2TB drive from my main computer.  Put it in slot 8.  unRaid is currently rebuilding disk8.  Hopefully it works.  If it does work then I can replace the parity with a 3TB and be able to use the replacement drives I bought for this purpose.  Hopefully before the other that is about to fail does so.

 

Nothing like having a drive fail during an upgrade. 

 

Update.

 

The rebuild of disk8 where I used the 2TB from my main computer went fine.  I then replaced the parity drive with one of the 3TB drives I bought.  Now that I'm using 5.0.5 it was seen and the rebuild of that went good.  I precleared the other new 3TB and that went good.  I've now got that in slot 3 (the other 500GB that was almost dead) and its currently rebuilding.  I'm also at the same time have the 2TB drive (the original parity drive) doing a preclear.  This drive will go back in slot8 giving me back my original 2TB drive so I can put it back in my main computer where it came from.

 

After this is done I will consider the whole thing complete.  At that point I'll try and do preclears on the 3 bad drives just to see what happens.

 

 

Overall here is what happened. 

- Originally Disk 6 died.  I replaced it and it froze during the replacement/rebuild.  Another reboot and it rebuilt successfully.

- During this build of disk 6, disk 8 had major issues and was unmountable.  Its a good thing unRaid does not need the filesystem mounted in order to rebuid a drive using parity.

- Disk 3 also gave reports of errors and looked like it might be close to death. 

- bought replacement drives my tower could not see due to it being an old version.

- Drive 8 red balled for good just as I updated the unRaid version.

- Used a stolen temp drive from another computer as my bought replacements were to big.

- replaced parity with one of the new bought replacement drives.

 

- Now rebuilding drive 3 with the other bought replacement.

- yet to go.  Recover the temp drive.  do some testing on the bad drives.

 

I have had some errors on drives 3 and 8 throughout this process so there may be a file or two that might be corrupt.  But I believe I still have all my data intact at this point.

 

Its something to be said that unRaid can go though this and come out without losing data.  But I admit.  I got lucky.  If drive 8 had died a reboot before it did,  I would have lost at least a drive full of data.  I've also learned more about smart reports and what a preclear is during all of this.

 

 

 

Question.

Is there a tool that lets me force a read/write of all sectors like the preclear does that can be run on a running in use system to check disk heath and force a fix of the smart reports "Current Pending Sector Count" by forcing a write and a remap if needed.  Does a standard parity check do this?  Or does that just read sectors?  If a section is not read, does it use parity data to attempt to write back to it and therefore force any remaps if needed.

 

  Or does that just read sectors?  If a section is not read, does it use parity data to attempt to write back to it and therefore force any remaps if needed.

That's it exactly. If a drive gives a read error, unraid reads the rest of the drives, calculates the needed data, and writes it back to the drive. If the write succeeds, the drive error column is incremented, and life moves on. If the write fails, the drive is red balled, and all further writes to the drive are calculated into the rest of the data drives and the parity drive.

Any drive with a pending sector should be rebuilt.

 

http://lime-technology.com/wiki/index.php/Troubleshooting#Resolving_a_Pending_Sector

Everything has finally finished.  All suspect drives are now replaced and rebuilds done.

The original disk 6 red balled and was swapped and rebuilt.  This drive was not precleared but the new one built without any errors.

Drive 8 then red balled so Disk 8 was swapped with my stolen HD from my main windows box.  Not precleared but built.  Had a couple errors.  At the time Disk 3 was also throwing errors so I just figured it was related.

Parity was swapped with a precleared (no errors during preclear.) drive and rebuilt.  No errors on the parity but again a couple errors on disk 3.  Don't remember offhand if disk 8 had errors.

Disk 3 was swapped with a precleared (no errors during preclear) drive and rebuilt.  Disk 3 rebuilt and had no errors.  Disk 8 showed a couple errors.

Disk 8 was swapped with a precleared (no errors during preclear) drive and rebuilt again.  This is where I put back the orinal parity drive and its back in the array now as disk 8.  Drive 8 rebuilt and no errors showed anywhere.

 

At this point the array is fully repaired.  3 new drives are in place as replacements for the disks that had issues.  No errors on any of these drives.  The odd swapping order is due to when the drives died and the fact that there is a unRaid version update in there to take the newer bigger drives.

 

So now I have my original HD that took from my main machine that showed a couple errors when it did its duty as a temp in slot 8.  I had done a smart report on it after it had a couple errors duing one of the rebuilds and knew it had 3 pending sectors.  I've just finished a preclear on it to see if that would fix it.  I believe it has but I'm not sure what the results mean.

 

The preclear script says it had 3 pendings before it stared and 0 pending after the zeroing part.  However the re-allcated number did not go from 0 to 3 as I expected.  It stayed at 0.  I ran a short smart test again after the preclear finished and it shows "Offline_Uncorrectable" at 3.

 

Both the preclear results and the short test smart report are in the attached txt.

 

I'm aware that I rebuilt twice with drives that showed errors in the error column.  But I had no choice.  I suppose this means there might be few files that are corrupt.  I do not have a way to test them other then coming across them in the future.  However.....  Having 2 drives redball and one more about to go.  Oh and during one rebuild I also had a drive saying "unformated" the entire time.  Doing all the above.  And it seems I have not lost any data.  I'm very very happy.  unRaid once again rocks.

 

 

 

In regards to this last drive.

So is this drive fixed?  Can I put it back in my main windows machine I took it from?  I guess my question is whats "offline_uncorrectable" mean in the smart report.  I was expecting the 3 pendings to turn into 3 reallocated but instead they tuned into offline_uncorrectable.  Is this the proper result?

 

after_preclear.txt

The reports looks ok. Not all pending sectors get reallocated.

Thanks for the info.  I've marked this as solved.  I'll come back just for completeness sake once I attempt to zero these 3 bad drives.

 

I would just like to thank everyone who posted.  Thank you very much.

 

Cal.