Hi guys!

 

Is it possible to replace a parity, specifically parity 2 with a failing drive in the array?

 

I have two 10 TB drives as my parity right now and I got a 16 TB drive on the way to slowly prepare for larger drives, but yesterday one of my 6 TB drives started to puke out read errors and reallocated sectors so I don't trust it to complete a parity sync.

 

I'm aware of the parity swap procedure but I would prefer to not have the server offline for 24+ hours so I'm wondering if I can unassign Parity 2 and assign it to the failed disk and have it rebuilt with a single parity drive until the new disk arrives?

 

Backups are in order so if shit hits the fan, I won't be crying about it. I would just prefer to make the swap as painless as possible.

Edited September 18, 20205 yr by SelfSD
Solved, editing title

So the disk just went offline and will not spin up again at all, so I attempted this and it's rebuilding data now. It was about as easy as I imagined but I was rather worried when an existing drive started showing errors lol.

 

For anyone else looking to do this:

 

1. Unassign the bad disk if it's not offline already.

2. Unassign parity 2.

3. Start array in maintenance mode. This will make Unraid forget parity 2 and you don't have to wait for dockers and VMs to start first.

4. Assign the old parity disk to the offline disk

5. Start the array normally and the rebuild will begin.

 

Also a side note that I was not aware of but I'm glad it happened! I forgot to turn off auto start for my array but with the missing disk Unraid did not start up automatically.

Well that was a scare, but the disk has been successfully rebuilt!

 

The CrashPlan Pro docker decided to fill up all my RAM like it sometimes does, throwing Out Of Memory errors and killing random things. I guess setting it to use max 16 GB out of the 32 I have does nothing when it first starts to go crazy.

I was able to SSH into the server after a few attempts as it kept killing my SSH sessions and managed to open top and luckily see the mdrecovery process run until it disappeared and confirmed that it was done with the HDD indicator on my server.

I wasn't able to run anything else, poweroff and powerdown just closed my session and pressing the power button did nothing, so I force rebooted it.

 

Thankfully Unraid had written to the USB drive that it was done so the disk is now fine, but I will let it run through the parity check again just to be sure, this time with CrashPlan turned off lol.

 

I am honestly quite impressed that nothing went wrong with the rebuild and my confidence in Unraid keeping my data intact is growing more and more with every problem I encounter.