today I reinitialised my NAS and by accident all disks were still plugged in, so the partition tables etc were overwritten.
After patting myself on the back for this brilliancy, I started to assess the damage. Some of the data existed only on that NAS, therefore I now need to recover the data from a (legacy) raid 5. Yay.
I hope someone here has some info on the structure and format of legacy volumes, it would be ideal if someone still runs a legacy volume, some of these things can be found out easily if a running config is available
What is the partition table type of the disks in Legacy Volumes? gpt or "classic" MS-DOS/intel? (Readable via
Code: Select all
parted | grep -i partition
What is the partition layout of these disks? (output of
Code: Select all
lsblk
I know there are usually multiple partitions on each disk and I assume multiple md devices span across the disks. Some data is part of a raid1 across all disks, while the actual data holding partition is of the configured raid type, raid5 in my case.
What is the purpose of the other smaller partitions? Are they relevant for the recovery?
Can the "data partition" (after recovery of each single disks partition table) be added directly to an mdadm?
Can the resulting md device be mounted directly or is there any other abstraction layer in between (like the newer volumes)?
Bonus points: What exactly happens during the drive "deletion" as part of reinitialization? Is a part of the disk overwritten with dd?
Which binary / script on the Nas takes care of this operation?