[Solved] Failed rebuild on drive update, need some help

[tabish121]

I just tried to upgrade an old 1TB WD EACS drive to a new 3TB WD Red drive.  I pre-cleared the RED drive 5 time and checked the SMART report and all looked fine so I stopped and powered down the array.  I installed the new drive after removing the old one and powered back up.  Everything looked ok so I assigned the new drive into the old one's slot and started the array.  After some time I came back and checked to see how things were going and the rebuild had stopped and the drive was red balled.  From the looks of the log the drive is having some issues even though it tested fine to begin with. 

 

So now the rebuild has failed and I'm stuck.  I would like to return the old drive to the array while I run some more tests on the Red drive and most likely RMA it but I don't know and can't find a procedure that will get me out of this.  I've read that at this point UnRaid is not going to let me insert a smaller drive into the old slot as the new old is already assigned there. 

 

So where I'm at right now is that I have valid parity, I have the old drive with the correct data, and I have another 1TB spare WD EACS drive to play with.  Can anyone suggest a way to get back into a good state while I figure out what the deal is with this RED drive? 

 

Syslog from the failed rebuild is attached.

syslog-2014-03-30.txt.zip

[bkastner]

You could take a screenshot of your drive positions and then run initconfig from telnet which will wipe the drive assignments. You can then manually add each drive back to it's original spot (including your original 1TB drive). I don't know if this will fubar your parity (though since you haven't written anything since swapping drives) you should likely be okay.

 

I am not 100% on the parity validity, but it will at least get all your data back.

 

you can choose to select Parity is valid, or not. If you do, you hope everything is good. Otherwise you can let parity get rebuilt with your 1TB drive and you should be guaranteed covered.

 

Personally I don't use this method to replace a drive (forcing a rebuild). I will usually add the new drive to a new slot and use mc to move files via screened telnet. I can then initconfig and remove the old drive and rebuild parity. It's definitely more time consuming I guess, but I've never had an issue with this method.

[tabish121]

Thanks that sounds like a possible solution.  I'm not really worried about having to rebuild parity in this case, just want to get the drive back without losing anything. 

 

I had considered the option of adding the new drive and doing a copy however the process of shrinking the array once the copy was done seems a bit daunting, at least from what I see here http://lime-technology.com/wiki/index.php/FAQ_remove_drive. The bit about the mover script confused me a bit, but I'm guessing that I don't need that if I were to move the data myself to a new drive first. 

 

In any case I'll do a bit more documentation digging to try and get a bit more educated on this before doing anything.  My system has been running smoothly for so long I've forgotten a lot since I originally set things up and read all the docs, and thing have changed a bit since 5.0 came out. 

 

Thanks for the tips.

[tabish121]

Did a new config and reassigned the drives in their proper location.  Rebuilding parity just to be on the safe side.