Restoring backup - to which device?

I need to restore my OMV system partition from a backup, so I've already mounted an Ubuntu live system and the USB stick containing the zst file

I've already started zstdcat backupfile.dd.zst | sudo dd of=/dev/nvme0n1p2 status=progress

However, is this the right device to restore the backup to?

The .sfdisk file contains the following information:

label: gpt

label-id: 0518E42F-9245-4061-A822-F294B764FC98

device: /dev/nvme0n1

unit: sectors

first-lba: 34

last-lba: 976773134

sector-size: 512

/dev/nvme0n1p1 : start= 2048, size= 1048576, type=C12A7328-F81F-11D2-BA4B-00A0C93EC93B, uuid=CE484F9A-DBF8-408E-BE17-BDC8849DF966

/dev/nvme0n1p2 : start= 1050624, size= 973721600, type=0FC63DAF-8483-4772-8E79-3D69D8477DE4, uuid=D492A8D9-8309-4412-91C7-B4BF6A1CF6EC

/dev/nvme0n1p3 : start= 974772224, size= 1998848, type=0657FD6D-A4AB-43C4-84E5-0933C84B4F4F, uuid=F5AB79E3-BDC9-4824-BFE3-2DBAAFB7C7B7

So should I restore to /dev/nvme0n1 instead as the file probably contains all 3 partitions (boot, main, swap)? Or doesn't the dd file contain a straight sector by sector dump of the SSD which I can just send to the original device?

Also I guess that there's no way to speed things up as the partition is very large (512GB)? I will certainly resize it but I guess that there's no way to skip the empty space? At this rate it will take over 5 hours.

EDIT: I've changed the restore to target /dev/nvme0n1 and I've added conv=sparse to try to skip the empty parts

EDIT2: the command completed successfully but the system doesn't boot

EDIT3: trying to follow the restore instructions [How-To] Restore OMV system backup made with openmediavault-backup plugin

I was able to restore the partitions although gparted gives a warning that the backup GPT partitions table is corrupt

However, I think that the zstd -d command will try to extract the full 512GB to the 32GB usb stick which obviously won't work. I'll probably try to add an USB harddisk later

Thanks!

Thanks for your reply.

So the first thing which I tried was the correct command:

zstdcat backupfile.dd.zst | sudo dd of=/dev/nvme0n1p2 status=progress

as /dev/nvme0n1p2 is the OS partition. Sadly, this resulted in an unbootable system. Could this be caused by the conv=sparse parameter?

Because restoring without sparse would take 5 hours I've started to rebuild my NAS but now I've ran into a new problem. My setup uses two harddisks, encrypted with LUKS in a RAID 1 configuration. After installing OMV to my SSD I've successfully installed LUKS and decrypted the harddisks (setup automatic decryption at boot as well).

So I have /dev/sda , unlocked , /dev/mapper/sda-crypt and /dev/sdb, unlocked, /dev/mapper/sdb-crypt

However, I can't create a RAID 1 array using multiple device as both devices won't show up and there aren't any other options available.

I'm not sure which mdadm command I should use now as I don't want to restore all the data.

I've tried the following command:

mdadm --assemble /dev/md0 /dev/mapper/sda-crypt /dev/mapper/sdb-crypt

mdadm: no recogniseable superblock on /dev/mapper/sda-crypt

mdadm: /dev/mapper/sda-crypt has no superblock - assembly aborted

The output is surprising to me as /dev/mapper/sda-crypt is shown as unlocked under encryption

And, if I remember correctly, during the original install I first encrypted both harrdisks with LUKS and created the array afterwards on both decrypted devices.

EDIT: ok, I was able to just mount the existing filesystem under /dev/dm-0! So I guess that I didn't use MD to configure the array? It's confusing to me that I can't see the RAID 1 array there and under encryption only /dev/mapper/data-crypt1 is shown as referenced so it's probably not RAID1.

EDIT2: according to filesystems RAID 1 is active:

Label: none uuid: 82b09deb-bd8f-4ce1-91af-42cfd5824c14

Total devices 2 FS bytes used 8.97TiB

devid 1 size 16.37TiB used 9.31TiB path /dev/mapper/data-crypt1

devid 2 size 16.37TiB used 9.31TiB path /dev/mapper/data-crypt2

Data, RAID1: total=9.30TiB, used=8.96TiB

System, RAID1: total=8.00MiB, used=1.66MiB

Metadata, RAID1: total=12.00GiB, used=9.55GiB

GlobalReserve, single: total=512.00MiB, used=0.00B

# I/O error statistics

[/dev/mapper/data-crypt1].write_io_errs 0

[/dev/mapper/data-crypt1].read_io_errs 0

[/dev/mapper/data-crypt1].flush_io_errs 0

[/dev/mapper/data-crypt1].corruption_errs 0

[/dev/mapper/data-crypt1].generation_errs 0

[/dev/mapper/data-crypt2].write_io_errs 0

[/dev/mapper/data-crypt2].read_io_errs 0

[/dev/mapper/data-crypt2].flush_io_errs 0

[/dev/mapper/data-crypt2].corruption_errs 0

[/dev/mapper/data-crypt2].generation_errs 0

# Scrub status

UUID: 82b09deb-bd8f-4ce1-91af-42cfd5824c14

Scrub device /dev/dm-0 (id 1)

no stats available

Scrub device /dev/dm-1 (id 2)

no stats available

mdadm --detail /dev/md-0 doesn't work so mdadm is probably not the right tool

I've decided to do a rebuild, it's still not clear how the RAID configuration works as I expected it to be a part of the (wiped) SSD. But happy that it works.

Switched to the fsarchive format for backups and I will consider other backup options for the system SSD like clonezilla.

Thanks for your reply BernH. That's a huge difference compared to DD and I've already configured backup to use this format.

I hope that it will take a long time before I'll need to use it.

The RAID 1 issue (not showing in MD after recovery) is still unclear to me. I'll do some further research and make a new post if I can't figure it out.