Lots of content out there on this message, but I have what I believe to be a somewhat unique circumstance.

 

I am trying to replace my cache drive (btrfs). My chassis is somewhat dated and all my PCI slots are taken. Giving this a shot, I disconnected the parity drive and connected the new cache drive, which was not yet formatted.  I added the new drive to the cache pool, but I did not see a way to format it.

 

So I shut down the machine, removed the new cache drive and reconnected the parity drive. Changed the cache pool size back to 1. Started the array, and parity re-ran automatically.

 

However the cache drive now shows "Unmountable: Unsupported or no file system"

 

The cache drive contains a few VMs which are not mirrored to the array, so I want to be careful that I don't lose its contents.

Please post the diagnostics.
 

I can't believe I forgot to do that.

 

[Diagnostics removed for privacy purposes]

 

Edited November 5, 20241 yr by Zudnic

Post output from:

 

blkid /dev/sdf1

and

fdisk -l /dev/sdf

 

Thank you..

 

I rebooted, the drive in question is sdg now.

 

root@NAS:~# blkid /dev/sdg1
/dev/sdg1: UUID="3e4c1317-5258-403f-be33-554ba20fbd00" UUID_SUB="58a6c765-6cc6-4d17-ac9b-4ae31f85a31d" BLOCK_SIZE="4096" TYPE="btrfs"

 

root@NAS:~# fdisk -l /dev/sdg
Disk /dev/sdg: 465.76 GiB, 500107862016 bytes, 976773168 sectors
Disk model: Samsung SSD 860 
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x00000000

Device     Boot Start       End   Sectors   Size Id Type
/dev/sdg1          64 976773167 976773104 465.8G 83 Linux

And the output from:

btrfs fi show

 

5 hours ago, JorgeB said:

And the output from:

btrfs fi show

 

root@NAS:~# btrfs fi show
Label: none  uuid: faa9618e-5b63-48f8-bcd6-c35221312204
        Total devices 1 FS bytes used 412.00KiB
        devid    1 size 1.00GiB used 126.38MiB path /dev/loop2

Label: none  uuid: 3e4c1317-5258-403f-be33-554ba20fbd00
        Total devices 2 FS bytes used 424.97GiB
        devid    1 size 465.76GiB used 465.76GiB path /dev/sdg1
        *** Some devices missing

 

Any idea where the missing device is?

Thank you - that's a good question.

 

The first thing I did was connect the new drive via PCI, expand the cache to 2 devices, and add the new one alongside the old one. (In doing so, I disconnected the parity drive because I needed the PCI slot.)

 

The system recognized the new drive in the pool, but did not format it, and nothing was written to it.  So I removed the new drive, reconnected the parity drive, and shrunk the cache back to 1. This resulted in the old cache giving the "Unmountable: Unsupported or no file system" error as shown in my screenshot. Trying to get the system to mount the cache drive, I created a new pool with the old drive only (with a new name, "newcache") and removed the original cache pool.

 

Currently, the new drive is attached via USB and shows in historical unassigned devices.

7 minutes ago, Zudnic said:

Currently, the new drive is attached via USB and shows in historical unassigned devices.

If it only shows up on historical devices it's not currently attached, attached it and post the btrfs fi show output again, but note that some USB bridges are not transparent, ideally it would be connected to a SATA port.

I don't intend to run the cache off USB, I'm going to replace the 500gb drive in sdg. I had the new one in USB to see if I could try to format it. I've removed it and re-ran the commands:

root@NAS:~# blkid /dev/sdg1
/dev/sdg1: UUID="3e4c1317-5258-403f-be33-554ba20fbd00" UUID_SUB="58a6c765-6cc6-4d17-ac9b-4ae31f85a31d" BLOCK_SIZE="4096" TYPE="btrfs"
root@NAS:~# fdisk -l /dev/sdg
Disk /dev/sdg: 465.76 GiB, 500107862016 bytes, 976773168 sectors
Disk model: Samsung SSD 860 
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x00000000

Device     Boot Start       End   Sectors   Size Id Type
/dev/sdg1          64 976773167 976773104 465.8G 83 Linux
root@NAS:~# btrfs fi show
Label: none  uuid: faa9618e-5b63-48f8-bcd6-c35221312204
        Total devices 1 FS bytes used 412.00KiB
        devid    1 size 1.00GiB used 126.38MiB path /dev/loop2

Label: none  uuid: 3e4c1317-5258-403f-be33-554ba20fbd00
        Total devices 2 FS bytes used 424.97GiB
        devid    1 size 465.76GiB used 465.76GiB path /dev/sdg1
        *** Some devices missing

 

The other SSD is still not connected.

Correct, I want to fix this one first since I need its data.

But this pool is missing a device, and it could be that one.

So I disconnected the parity drive and added the new drive in its PCI slot. Unraid recognized the new drive as part of the pool, and I can now read the contents of the original disk.

 

Thank you!

 

It seems that the drive itself remembers if it has been in a pool, because the current pool is new and only ever contained the old drive. The drive must have somehow retained something about being part of a cache when it was in the old cache with the new drive.

 

I can see the new disk filling, presumably it's mirroring the old drive. Last question - can I simply remove the old drive once the clone is complete?

 

Thanks again!

Post new diags to see the pool layout.

So this is resolved. I found that the PCI ribbon had another connector intended for an optical drive, so I hooked that up and the system recognized both drives. I added both to the cache and now I have a pool.

 

Unfortunately we never did get to the root cause, so apologies to those of you who find this in the future.

 

@JorgeB -- Thank you again for your help.