Hi,

i added 4x 28TB, so 2 of them are set as new parity. Starting the array resulted in a bunch of unmountable disks that contained data from old disk setup.

Big issue because this would mean big data loss.

What are the next steps before i lose 4 disks ?

old disk arrangement:

Diagnostic file

unraid-diagnostics-20250522-1324.zip

new disk arrangement:

With the array running, post the output from

wipefs /dev/md#p1

Replace # with all the disk numbers that are not mounting, and note that the command as written won't wipe anything, despite the name.

root@Unraid:~# wipefs /dev/md#p1

wipefs: error: /dev/md#p1: probing initialization failed: No such file or directory

p1 as disk1 not correct ?

root@Unraid:~# wipefs /dev/md#p1

wipefs: error: /dev/md#p1: probing initialization failed: No such file or directory

root@Unraid:~# wipefs /dev/md1#p1

wipefs: error: /dev/md1#p1: probing initialization failed: No such file or directory

root@Unraid:~# wipefs /dev/md1p1

DEVICE OFFSET TYPE UUID LABEL

md1p1 0x0 xfs 8baf08db-eb09-4d11-8679-28345a1b1cd1

md1p1 0x105efff44000 zfs_member 1436321885595498112 disk1

md1p1 0x105efff84000 zfs_member 1436321885595498112 disk1

root@Unraid:~#

For disk1 is

wipefs /dev/md1p1

disk2

wipefs /dev/md2p1

etc

root@Unraid:~# wipefs /dev/md1p1

DEVICE OFFSET TYPE UUID LABEL

md1p1 0x0 xfs 8baf08db-eb09-4d11-8679-28345a1b1cd1

md1p1 0x105efff44000 zfs_member 1436321885595498112 disk1

md1p1 0x105efff84000 zfs_member 1436321885595498112 disk1

root@Unraid:~# wipefs /dev/md3p1

DEVICE OFFSET TYPE UUID LABEL

md3p1 0x0 xfs 21f17d02-e56d-4f88-ad50-4d010aa90764

md3p1 0x105efff44000 zfs_member 17208599687362965281 disk3

md3p1 0x105efff84000 zfs_member 17208599687362965281 disk3

root@Unraid:~# wipefs /dev/md4p1

DEVICE OFFSET TYPE UUID LABEL

md4p1 0x0 xfs afeb9d46-e58e-4553-95a2-488086cb2bec

md4p1 0x105efff44000 zfs_member 6186790056179969331 disk4

md4p1 0x105efff84000 zfs_member 6186790056179969331 disk4

root@Unraid:~# wipefs /dev/md5p1

DEVICE OFFSET TYPE UUID LABEL

md5p1 0x0 xfs da9fe415-3d8c-40b3-8066-fb3d7f3a3bf6

md5p1 0x105efff44000 zfs_member 1428139698925851645 disk5

md5p1 0x105efff84000 zfs_member 1428139698925851645 disk5

root@Unraid:~#

Just now, Grobalt said:

md1p1 0x105efff44000 zfs_member 1436321885595498112 disk1

md1p1 0x105efff84000 zfs_member 1436321885595498112 disk1

This is the problem, quick fix is to stop the array, click on each disk and change the filesystem from auto to xfs

Permanent fix it to wipe the old zfs signature, you can use:

wipefs -a -t zfs_member /dev/md#p1

This should be perfectly safe, and I've used it multiple times myself and with other users, but always good to make sure backups are up to date, just in case.

This assuming the disks are all using xfs, of course

okay, i first stopped, clicked the 4 disks and changed auto to xfs

after that i tried the permanent fix but that was not possible, the zfs assignment was already gone

array seem to start correct and is building parity

1 hour ago, Grobalt said:

after that i tried the permanent fix but that was not possible, the zfs assignment was already gone

Not sure I follow

i changed filesystem (stopped) to xfs and after that i run the command wipefs -a -t zfs_member /dev/md1p1 without success.

i ran again wipefs /dev/md1p1 which showed no more zfs assignements etc.

so i started the array and all is looking good, parity is created on the new 28db drives.

Edited May 22, 20251 yr by Grobalt

That doesn't make sense, what did the first wipefs command output?

sorry .. i think it was just empty line without error message ...closed the CLI window already

If you only did it for disk1, post the listing output for another disk, e.g. wipefs /dev/md3p1

i did for all, otherwise array would not start correctly, sorry. now i let the parity build on the new 28tb drives and after that replace the 18 with 28 one by one.

but every time i did too - new config - it did not detect xfs on the existing disks .. had to set it manually every time.

Something is missing, but as long as it's working...