The docker service suddenly died and failed to restart.

Can't delete the vdisk file:

rm: cannot remove '/mnt/user/system/docker/docker.img': Read-only file system

 

I checked the syslog and found this

Apr 24 00:16:26 HP-GEN7 kernel: BTRFS warning (device nvme0n1p1): nvme0n1p1 checksum verify failed on 237434634240 wanted 0x20b19b65 found 0xaec091ec level 0
Apr 24 00:16:26 HP-GEN7 kernel: BTRFS: error (device nvme0n1p1) in __btrfs_free_extent:3095: errno=-5 IO failure

 

I thought it's the cache disk's issue at the first place, but then I noticed that all disks became read-only and the mover can't move my data from cache to other disks.

 

How should I backup the data on the cache disk and restore the system?

hp-gen7-diagnostics-20220424-0025.zip

Edited April 24, 20224 yr by johnmick

I'll leave it to others to answer, but only the cache is read-only (and by reference, the docker.img file).  Mover can't properly move because it can't remove the files from the cache drive. 

 

 I will say however that on a single device pool if you have no intention of making it multi-device you're better off with XFS than BTRFS as while BTRFS does work, it is dependent upon an absolute rock-solid system and XFS is more forgiving.

1 hour ago, Squid said:

I'll leave it to others to answer, but only the cache is read-only (and by reference, the docker.img file).  Mover can't properly move because it can't remove the files from the cache drive. 

 

 I will say however that on a single device pool if you have no intention of making it multi-device you're better off with XFS than BTRFS as while BTRFS does work, it is dependent upon an absolute rock-solid system and XFS is more forgiving.

Yes, you're right, I've tested again and only the cache is read-only.

Can I set all cache settings as "No" and manually copy the cache files back to other disks? Will unraid recognize those files correctly?

Btrfs is detecting a lot of data corruption, you should run memtest.