I have been using two 2TB Intel nvme drives as my cache pool.  One of the drives started reporting read and write errors, so I removed it, put in a 2TB XPG 2TB NVME.  XPG_GAMMIX_S11_Pro_2J3720088745 - 2.05 TB

 

Unfortunately, it reported like 3 writes and then apparently disappeared and stopped reading/writing.  In the meanwhile, I did a surface read/write test in HD Sentinel on the supposed bad 2TB drive and it was fine, no issues/errors.

 

So I just went to put back the original Intel 2TB in unraid server, so its the original dual Intel nvme drives, but getting:

I'm guessing that the issue is that the XPG is reporting 2.05TB and the Intel drives just 2.0TB and so I need to use a larger drive to get the array to start?

 

 

Am I able to go to 1 slot using the Intel cache drive with data on it, start the array without formatting the cache drive, stop and re-add the Intel drive back to mirror cache pool?  All without losing data? It doesn't look like I can select 1 slot.

 

Another option would be to screenshot the current array, make a new config, set the array as it was, and tell it to start the array without formatting?

 

Thanks in advance!

 

 

Oops, didn't include a diagnostics. My bad.

I installed the XPG 2TB drive back in and have a "Warning btrfs too many profiles" issue. The array still starts and runs from the Intel 2TB nvme, but based on searches, I'm wondering if I can go to Cache->Convert to Raid1 mode->Balance?

unraid-diagnostics-20250216-1923.zip

Post the output from

btrfs fi show

 

1 hour ago, JorgeB said:

Post the output from

btrfs fi show

 

 

 

# btrfs fi show
Label: none  uuid: 9e7e00da-48c5-4dae-9371-25e83b91f33c
        Total devices 2 FS bytes used 1.09TiB
        devid    3 size 0 used 0 path  MISSING
        devid    4 size 1.86TiB used 1.38TiB path /dev/nvme2n1p1

Label: none  uuid: e6df60ce-29d8-483b-868d-789b591f608e
        Total devices 1 FS bytes used 72.97GiB
        devid    1 size 100.00GiB used 82.07GiB path /dev/loop2

 

Pool is currently only using one device, reimport the pool with that single device, it will be imported degraded, you can then replace the missing device:

 

on main click on the first device for that pool and then "remove pool"
back on main, create a new pool with the same name and 1 slot
assign the pool device (nvme2n1), leave the filesystem set to auto
start the array to import the pool

 

Then stop the array and replace the missing device with the new one, just assign it and start the array

2 minutes ago, JorgeB said:

Pool is currently only using one device, reimport the pool with that single device, it will be imported degraded, you can then replace the missing device:

 

on main click on the first device for that pool and then "remove pool"
back on main, create a new pool with the same name and 1 slot
assign the pool device (nvme2n1), leave the filesystem set to auto
start the array to import the pool

 

Then stop the array and replace the missing device with the new one, just assign it and start the array

 

In the first step, "on main click on the first device for that pool and then "remove pool"", "REMOVE POOL" is greyed out and cannot be selected. Do I need to stop array first?

Yes

16 hours ago, JorgeB said:

Yes

So I did as you last stated, and after selecting 1 slot, (nvme2n1) device and start the array, I get:

 

And the following errors:

 

I shutdown, put the original 2TB Intel NVME back in and started up. The array was stopped and expecting cache slot 1 to be assigned. I assigned it the Intel nvme (taken offline and tested and cleared), and I get:

 

 

So it looks like I'm back where I started.

 

After removing the pool, maybe something else needs to be done to make it forget the old profiles before adding the single cache drive back?

 

Thanks!

 

Post new diags.

26 minutes ago, JorgeB said:

Post new diags.

Thank you.

unraid-diagnostics-20250218-0015.zip

Pool has part of the data using single profile, so it won't accept the new device, with the array running type:

 

btrfs balance start -dconvert=raid1,soft -mconvert=raid1,soft /mnt/cache

 

When that's done, stop the array, start the array, post new diags, you still may need to reimport the pool once more.

On 2/18/2025 at 2:08 AM, JorgeB said:

Pool has part of the data using single profile, so it won't accept the new device, with the array running type:

 

btrfs balance start -dconvert=raid1,soft -mconvert=raid1,soft /mnt/cache

 

When that's done, stop the array, start the array, post new diags, you still may need to reimport the pool once more.

Sorry, got busy and only getting time to try now:
 

# btrfs balance start -dconvert=raid1,soft -mconvert=raid1,soft /mnt/cache
ERROR: error during balancing '/mnt/cache': Invalid argument
There may be more info in syslog - try dmesg | tail

 

Sorry, since there's a missing device, that won't work, type this instead:

 

btrfs balance start -f -dconvert=single,soft -mconvert=single,soft /mnt/cache

 

Post new diags when done, with the array still running 

On 3/3/2025 at 1:32 AM, JorgeB said:

Sorry, since there's a missing device, that won't work, type this instead:

 

btrfs balance start -f -dconvert=single,soft -mconvert=single,soft /mnt/cache

 

Post new diags when done, with the array still running 

 

 

 

# btrfs balance start -f -dconvert=single,soft -mconvert=single,soft /mnt/cache
Done, had to relocate 1197 out of 1439 chunks

Thank you, again. Note, it has a 2TB drive, so that 4.1TB is odd for only showing 1 drive.

unraid-diagnostics-20250305-1351.zip

With the array still running, type:

 

btrfs dev del missing /mnt/cache

 

Then stop the array, and reimport the poool with just that device:

 

on main click on the first device for that pool and then "remove pool"
back on main, create a new pool with the same name and 1 slot
assign the pool device, leave the filesystem set to auto
start the array to import the pool and post new diags

On 3/5/2025 at 11:49 PM, JorgeB said:

With the array still running, type:

 

btrfs dev del missing /mnt/cache

 

Then stop the array, and reimport the poool with just that device:

 

on main click on the first device for that pool and then "remove pool"
back on main, create a new pool with the same name and 1 slot
assign the pool device, leave the filesystem set to auto
start the array to import the pool and post new diags

 

That seemed to go ok. 

 

Thank you again!

unraid-diagnostics-20250309-2049.zip

Looks OK, you should now be able to add the other device.

14 hours ago, JorgeB said:

Looks OK, you should now be able to add the other device.

All good, thanks very much for walking me through all the steps. Much appreciated!