Forum Discussion

Tutor

Jul 31, 2020

ReadyNAS Pro 6 won't assemble RAID / X-RAID2 after reboot - "Remove unused volumes"

Running ReadynasOS 6.10.3 Expanded from 6x4TB to 6x8TB and everything went well over the 6 drive swaps, just very slow upgrading. Now after 4GB OS root volume filled up (now cleared) and reboote...

itachi2

Tutor

Jul 31, 2020

So, reading up a bit more on BTRFS (I seem to only do this when there's a problem)...

BTRFS has its own implemetation of volume management, and you can add devices to a particular filesystem's pool. So both md126 and md127 are in the pool, and if you mount either it should mount all devices that are part of that filesystem.

I was able to confirm that the pool was okay:

# btrfs fi show
Label: '33eab4f9: root'  uuid: 1d37ef4d-3226-418e-b4d4-23d2cb134fcc
	Total devices 1 FS bytes used 3.51GiB
	devid    1 size 4.00GiB used 4.00GiB path /dev/md0

Label: '33eab4f9: Data'  uuid: 67bc10c3-1308-49f2-ba7e-8058fcdaf81c
	Total devices 2 FS bytes used 17.28TiB
	devid    1 size 14.54TiB used 14.54TiB path /dev/md126
	devid    2 size 14.55TiB used 2.79TiB path /dev/md127

Running "btrfsck /dev/md126" as a readonly check, which I hope also checks the entire FS including md127:

Checking filesystem on /dev/md126
UUID: 67bc10c3-1308-49f2-ba7e-8058fcdaf81c
checking extents
checking free space cache
checking fs roots
root 259 inode 8367164 errors 400, nbytes wrong

The inode error can apparently be fixed by finding and deleting that file:

find / -inum 8367164 -print

So after btrfsck is done I expect I'll be able to mount the FS, but would rather repair the assembly and mount process at boot than copy everything off, recreate my shares and dump everything back on again.

itachi2

Tutor

Aug 08, 2020

So I believe the only thing wrong is the raid assembly then mount, as

rnutil start_raids
mount -a

gives my volume properly mounted on /Data, but no apps, etc because it didn't happen early enough in the boot order. Has anyone documented the boot process of ReadyNASOS? I'm trying to trace where the RAID assembly is failing, check configuration, etc. as the OS assembles md127 but not md126, so btrfs can't mount the filesystem because the pool is incomplete. I've backed up the data, and am prepared to recreate the volume if necessary, but again, this seems like a small configuration adjustment versus three days of copying files. I thought I was fairly familiar with Debian but can't seem to find out where these steps are supposed to happen.

StephenB
Guru - Experienced User
Aug 08, 2020
Does

# rnutil start_raids
# btrfs device scan
# mount /dev/md127 /data

give you the same result as the mount -a?