I'm about to reconfigure my array, currently:

2TB(p) + 2TB + 2TB + 1TB

replacing a slow-ish 2TB drive with a faster one, and then using the slow 2TB to replace the 1TB drive, so I initiated a Parity Check to make sure was all OK before I started messing about.

 

After about 20 mins, there was a short power cut which brought the server down.  I connected up the UPS that had been sitting idle and connected the UNRAID server to it (my optician says my hindsight is 20/20! )

 

On power up, UNRAID seemed to continue the Parity Check where it left off - I was expecting it to start again.  Is this expected behaviour?

 

I'm planning to use the "Replace a failed disk" procedure to swap in the faster 2TB drive, then the "Replace a single disk with a bigger one" to replace the 1TB with the slow 2TB one.  Does this seem reasonable? 

Will UNRAID get confused by seeing a drive that used to be in another slot?  In which case, how do I stop this happening?

 

This is running UNRAID 6.1.19, but it started out as Version 5 so the drives are all formatted as ReiserFS.  Is there any advantage to changing the format of the new drives to something else?  (is it even possible?)  If so, what would be better? 

 

Cheers,

 

Howard

 

Parity checks do not resume, but it will start a check automatically after an unclean shutdown. ReiserFS has some advantages, but as you move forward with bigger drives in the future, you're probably going to see more and more issues with ReiserFS drives. It's outdated and not maintained. To that end, I would recommend the following.

 

1. Finish your current parity check, and obtain the diagnostics zip file after it has completed.

2. Post the diagnostic file here, and ask someone to look it over, paying special attention to the smart reports for your existing drives.

3. Preclear the new 2TB drive, post results here for evaluation.

4. Assuming no drive issues come to light, set a new configuration with ALL the drives, so the new drive is in the parity slot, and your old parity drive is in slot 1.

5. Set the format for your old parity drive to XFS, and start the array. Make sure the rest of your drives are still readable, and select the option to format your old parity drive.

6. Copy the contents of your old drive 1 (now in slot 2)  to the newly XFS formatted drive one. Either use a copy command that does verification during the copy like rsync with the -c option, or verify the copy after the fact.

7. Stop the array, change the desired format of the disk you just copied off of to XFS. Start the array, make sure all drives are readable except the one you just changed, then format that drive.

8. Repeat the copy process from the next drive down onto the newly XFS formatted drive.

9. Continue down the chain until you are left with only the 1TB drive as ReiserFS and its contents are safely copied to one of the 2TB XFS drives.

10. Set a new config omitting the 1TB, making sure all the other drives are assigned where they were. Start the array and recalculate parity once again. Do a parity check after the initial build is complete.

 

If someone has a better (safer, faster) way to accomplish this, please suggest it.