Yes I check the drive letter to serialnumber after every boot and then
match the order to the original.

testdisk is showing so much wrong data too that I don't believe this
is useful to run.

Is there a way to identify the raid partitions from the separate
devices and then restore those partition information?
I don't really know what to do next now.

Topi

On Sat, 2 Mar 2024 at 14:03, [off-list ref] wrote:

On 2/3/24 20:55, Topi Viljanen wrote:

quoted

Hi,

syslog entries before any problems:

Feb 18 19:52:53 NAS-server kernel: [    5.755150] md/raid:md0: device
sde operational as raid disk 5
Feb 18 19:52:53 NAS-server kernel: [    5.755158] md/raid:md0: device
sdb operational as raid disk 0
Feb 18 19:52:53 NAS-server kernel: [    5.755160] md/raid:md0: device
sdf operational as raid disk 3
Feb 18 19:52:53 NAS-server kernel: [    5.755162] md/raid:md0: device
sdg operational as raid disk 1
Feb 18 19:52:53 NAS-server kernel: [    5.755164] md/raid:md0: device
sdd operational as raid disk 2
Feb 18 19:52:53 NAS-server kernel: [    5.755166] md/raid:md0: device
sdc operational as raid disk 4
Feb 18 19:52:53 NAS-server kernel: [    5.757272] md/raid:md0: raid
level 6 active with 6 out of 6 devices, algorithm 2
[...]
Feb 18 19:52:54 NAS-server smartd[1188]: Device: /dev/sdc [SAT],
ST4000DM004-2CV104, S/N:WFN14WVT, WWN:5-000c50-0bee83a77, FW:0001,
4.00 TB
Feb 18 19:52:54 NAS-server smartd[1188]: Device: /dev/sdd [SAT],
ST4000DM000-1F2168, S/N:Z305RJAR, WWN:5-000c50-087ccf402, FW:CC54,
4.00 TB
Feb 18 19:52:54 NAS-server smartd[1188]: Device: /dev/sdb [SAT],
ST4000DM000-1F2168, S/N:Z305RFDP, WWN:5-000c50-087cd03a5, FW:CC54,
4.00 TB
Feb 18 19:52:55 NAS-server smartd[1188]: Device: /dev/sde [SAT], WDC
WD40EZRX-00SPEB0, S/N:WD-WCC4E7TULA48, WWN:5-0014ee-2b5d93dd9,
FW:80.00A80, 4.00 TB
Feb 18 19:52:55 NAS-server smartd[1188]: Device: /dev/sdf [SAT],
ST4000DM000-1F2168, S/N:Z305RVDW, WWN:5-000c50-087cc5e98, FW:CC54,
4.00 TB
Feb 18 19:52:55 NAS-server smartd[1188]: Device: /dev/sdg [SAT],
ST4000DM000-1F2168, S/N:Z305RTB8, WWN:5-000c50-087cc9d0b, FW:CC54,
4.00 TB

Isn't that quite clear how the order should be? I have added those
disks to the create array in that order 0,1,2,3,4,5
Should I try something else?

What is meant by '0,1,2,3,4,5'? You mean some order of /dev/sd{x,y,z,...}.
Now the disk names may be different from your old setup. This is normal.
You need to use the Serial Numbers to have the correct order.
So in this case I see
        0 sdb Z305RFDP now sdx
        1 sdg Z305RTB8 now sdy
        etc.

Or maybe you already did this, so ignore this note.

quoted

I was wondering if the problem is unused space? The only disk having
it partition table have this:
Unused Space : before=254896 sectors, after=7856 sectors

So if creating a new array starts from the wrong place? Can that be
somehow inspected?

Topi

On Sat, 2 Mar 2024 at 11:46, Pascal Hambourg [off-list ref] wrote:

quoted

On 02/03/2024 at 09:50, Topi Viljanen wrote:

quoted

I was able to get the order of the devices from old syslog file
(smartd) and then created the array again:

(...)

quoted

Running fsck caused so many errors that the mounted ext4 was empty.
I reset the overlay array and now I'm running analyze with testdisk.

Then I am afraid that the order was wrong and testdisk will not be able
to retrieve much.

--
Eyal at Home (eyal@eyal.emu.id.au)