I am running Unraid 5.05 and have used Unraid for a number of years,but I'm not very good with Linux.  I just replaced two smaller drives with larger drives and had no problems (that I can tell). After the second drive was replaced, and the file system rebuilt, I noticed that one of my other disks showed a red ball beside it and had a large number of errors.  Looking at the statistics, it appeared that the drive was not even online.  I stopped the system, checked the wiring and the disk drive is now up and running, but is showing as unformatted.  I have looked at other posts and have rebooted, etc., but it is still showing as unformatted  I considered unassigning that drive and rebooting, but before I did anything else, I wanted to check.  I'm confident the data is still on the disk and don't won't to format, but I'm also concerned if the data on that disk is part of the parity, or has the Parity been updated as if the data on that disk doesn't exist.  Thanks!

Do NOT format the disk or you will lose the data

 

when a disk shows up as unformatted it frequently means that the disk failed to mount rather than it is really unformatted.  This is typically because some sort of file system corruption is present.  Also the fact that the disk is red-balled means that at some point a write to the drive failed and the drive is now being emulated using the remaining drives plus parity.  The reason for the write failure is not clear but could easily be an intermittent issue (e.g. cabling) rather than a problem with the actual disk.  A SMART report for the disk is worth taking to see if it looks OK.

 

The safest thing to do is to take the physical drive offline, and then try and fix the corruption on the logical drive.  Taking the physical drive offline means it is still available with data on it if the suggested recovery action fails.  The procedure that is suggested is

• Stop the array; set the problem drive as 'unassigned'  and then restart the array in maintenance mode.  The problem disk will be flagged as missing and you will be warned the array is now unprotected, but the missing drive will still be present at the logical level.
  
• From telnet/console session run a command of the form
  reiserfsck --check /dev/md?
  where ? corresponds to the problem disk number (e.g. md5).  This command will check the file system on the disk and if corruption is found suggest an action to take.
  
• If a recovery action is suggested I would advise checking back here for advice.  The advice is likely to be that you proceed with the suggested action, but better to be safe than sorry
  
• If the recovery action succeeds, then you be able to stop the array and restart it in normal mode.  Check that the data appears to be intact
  
• At this point you will want to rebuild onto a real physical drive to get back to a protected state. If you have a spare drive of a suitable size then I would recommend using this as it would still leave the physical problem drive unchanged.  If not then you can try rebuilding onto the problem disk - this will proceed, but is slightly riskier as the problem disk has not been checked
  
• If you rebuilt onto a different physical drive then you could try putting the original problem disk through a pre-clear cycle to check that it Is in fact OK (or not).
  

I ran reiserfsck as recommended.  Attached is the output from running that command.  Recommended next steps?

The good news is that the reiserfsck appeared to be able to access the disk fine, and it looks like corruption WAS detected (which is consistent with it showing up as unformatted0.  I would suggest that you take the recommended action.  This will run for some time (it takes even longer than the check you have already run).

 

Some points that are relevant for later:

• Have you removed the physical disk as recommended so that if by any chance the suggested the recovery does not work you still have that available for attempting data recovery?
   
  
• If the recovery succeeds do you have a spare disk to rebuild onto or will you need to re-use the original (suspect) disk?