I decided to expand my cache pool from a single 7200 RPM 320 GB 2.5" HDD to include 2x 500 GB SSD 2.5" drives.  The necessary preclear process was conducted and I stopped the array to add the to the pool the next day.  When the array came online I saw an error about profiles that indicated a balance routine needed to be run.  Once that was done I felt confident that I could remove the slower 7200 RPM 320 GB HDD from the pool.  I stopped the array and removed all the drives from the pool to resize it from 3 to 2 since the #1 device was the one I needed to be gone.  I then replaced what was #1 with #2 and #2 with what was #3.  I started up the array and was met with some error regarding unable to mount and also that one of the two disks was a new device.  I didn't format anything, but I've been trying to explore data recovery options in the event that's what is needed.  Unfortunately, I took a shortcut on this step and didn't back up as I felt that a run of the balance would have sufficed.

 

I can force the two drives to be the cache pool, but it's the data recovery or restoration I'm interested in.  I did try going back to the original setup of just the 7200 RPM 320 GB HDD and one cache drive configuration, but i see an error on the main page "Unmountable: No file system".  I tried running a btrfs recovery command.

btrfs restore -i /dev/sdg /mnt/user/docker/recoveredCache

but I got an error that made my hopes sink

No valid Btrfs found on /dev/sdg
Could not open root, trying backup super
No valid Btrfs found on /dev/sdg
Could not open root, trying backup super
No valid Btrfs found on /dev/sdg
Could not open root, trying backup super

I removed all the drives from the cache pool at this moment and have the array started.

 

I'm a bit unfamiliar with btrfs and how to recover from my current situation.

It's missing the partition for btrfs restore, see here:

https://forums.lime-technology.com/topic/46802-faq-for-unraid-v6/?do=findComment&comment=543490

It's also in the FAQ how to add and remove devices from the cache pool.

Before using btrfs restore you can try if the pool still mounts using all 3 devices together.

 

 

Many moons later and living without the cache drives back online I ended up pulling the cables on the original drive and left the two replacement drives connected and re-enabled the cache.  UnRaid flagged the pool unmountable, but recognized something about the cache disk #1.  I was able to conduct data recovery with

btrfs restore -v /dev/sdk1 /mnt/user/Recovery

And then rsync from the recovery directory after formatting the drives and getting cache back online.

 

Lessons learned: backup appdata and follow steps for cache drive/pool replacement/resizing regardless of how safe it may seem.

Edited December 10, 20178 yr by Sn3akyP3t3