Summary:

 

I had replaced the guts of the system with a new motherboard/cpu and added a shiny new 2TB NVME drive.  I added another slot to Cache Pool and added the NVME to slot 2.  Everything synced and was happy.  I had decided to try to remove the old cache pool drive (2.5 SSD) and was able to remove it.  Of course now I was in a weird state of a half working RAID1 pool.  I was doing more maintenance on the server and rebooted a couple times.  After the last reboot the array fails to start with "Wrong Pool State - cache too many wrong or missing devices"

 

Cache2 correctly identifies the NVME 2TB drive and its green color.  Cache1 is unassigned.  I've tried removing cache2 and then I've tried to mount it in order to back up the data and that doesn't work.  At this point I know how to fix it destructively, but i'd like to get appdata off the cache drive.  Is there anyway to recover the data?  I've got old backups on the array that I could use so its not the end of the world.

 

Diags attached.

 

 

Skipdog

unraid-diagnostics-20250122-1401.zip

The pool profile is set to single, try this, click on the first pool device, then "remove pool", back on main add a new pool with the same name and just 1 slot, assign that NVMe device only, start the array, post new diags.

JorgeB,

 

Thank you so much that got my cache drive back up and working.  I'll definitely perform a backup now of appdata.

 

It looks like when I assigned the single drive and it mounted it returned the pool devices to original 2 slot.  Should I invoke mover and get it all back on the array and then wipe the cache and then user mover to reduce cache drive to 1?  Also the 2TB is showing half the capacity.

 

 

New Diags uploaded.

 

Skipdog

unraid-diagnostics-20250122-1520.zip

You want to keep just that device for the pool, correct?

Hi JorgeB,

Yes for now I want to have a pool with a single NVME drive with the original size of 2TB.

 

 

Skipdog

Edited January 23, 20251 yr by Skipdog

With the current status, the easiest way is to use the CLI to remove the missing device, but first you need to balance the pool to single, click on it, balance to single, then post new diags to confirm.

Balance completed!  I see the space issue looks correct now.

 

New diags attached here.  I assume I can do something as you said to collapse the pool to 1?

 

Skip

unraid-diagnostics-20250123-1524.zip

With the array still running, type:

btrfs dev del missing /mnt/cache

Then stop the array and reimport the pool once more:

On 1/22/2025 at 7:14 PM, JorgeB said:

click on the first pool device, then "remove pool", back on main add a new pool with the same name and just 1 slot, assign that NVMe device only, start the array, post new diags.

 

Thank you!  Good to go!

 

Skipdog

This post helped me a lot. I'm on 7.0.1 and was in the same position, trying to replace a small cache pool disk with a larger one. I too added a slot to the pool and added the larger drive and set it to raid 1 and let it btrfs sync. Then i got stuck trying to remove the other drive. This is what worked for me once i was in the sa:

 

- stop array
- remove old SSD from pool
- start array, clicking "yes to remove disk" (looks like this automatically set the pool back to single mode)
- with array running, run command: btrfs dev del missing /mnt/ssd_cache_docker (name of my pool)
- stop array
- click on pool, "remove pool"
- recreate pool with same exact name