How to restore Timeshift backup to a new volume?

General system Newbie
1

I had an EndeavourOS btrfs installation, but my fs got corrupted, so I needed to recreate it.

I created the system backup using Timeshift in rsync mode, and /home backup via rsync -av manually, reinstalled EOS on this drive and restored both backups.

After that I could not boot because it could not find the new volumes by UUID (hibernation device 'UUID=...' not found, /new_root: can't find UUID ...). So I updated them in /etc/fstab and /etc/default/grub, chrooted and grub-mkconfig -o /boot/grub/grub.cfg, but still getting the same errors, even though they seem to show the current UUIDs.

Photo - Google Photos — Mozilla Firefox 2022-12-01 16.15.20
Photo - Google Photos — Mozilla Firefox 2022-12-01 16.15.201644×330 66 KB 

lsblk -f
NAME        FSTYPE      FSVER            LABEL       UUID                                 
nvme1n1                                                                                                  
├─nvme1n1p1 vfat        FAT32                        42EC-3490                                           
├─nvme1n1p2 btrfs                                    ee578df1-6236-4eea-b6e5-62579060b349                
└─nvme1n1p3 swap        1                swap        4d31088c-9a64-43de-8864-9fa9ef0f7aea
2

You might need to regenerate the initrams after making that change.

arch-chroot in and run mkinitcpio -P

3

I think I did this too already. :thinking:

How is it possible that it can’t find UUIDs that exist? At least in lsblk from the LiveCD. (can’t check it in the emergency shell, my keyboard doesn’t work there)

4

The bootloader is able to load the root volume but the initram can’t.

My best guess would be something wrong with the initramfs. Have you tried booting the fallback entry?

5

Oh, mkinitcpio -P fixed it, I guess I did it somehow wrong before.

Now it boots and the system things are restored fine, but my home is missing. Looks like @home is not mounted during @ mounting.

When I was restoring it, I simply mounted the disk via the file manager in the EOS LiveCD and rsync to ......../@home. Is it wrong?
I checked via the LiveCD again, and indeed the /home of the mounted disk is empty, while my files are present in @home.

But if I mount via

sudo mount -o subvol=@ /dev/nvme1n1p2 /mnt
sudo mount -o subvol=@home /dev/nvme1n1p2 /mnt/home

then everything is correct in /mnt/home. So I thought the same is supposed to happen when booting.

I don’t see any mounting errors in journal.

fstab:

UUID=ee578df1-6236-4eea-b6e5-62579060b349	/	btrfs	subvol=/@,defaults,noatime,autodefrag,compress=zstd	0	0
UUID=4d31088c-9a64-43de-8864-9fa9ef0f7aea	swap	swap	defaults,noatime	0	0
6

That isn’t automatic.

All your subvolumes should have separate entries in /etc/fstab

7

hm, but why it worked before? This fstab was working fine before the reinstall.

I thought I heard that nested subvolumes like this are mounted automatically. :thinking:

8

I don’t think your /etc/fstab looked like that. Do you still have the original from before you started making changes?

@home isn’t a nested subvolume.

A nested subvolume would be something like @/home

9

hm, weird, looks like they are indeed present in the timeshift backup, I doubt that I deleted all that. :thinking:

So the correct fstab is

UUID=42EC-3490                            /boot/efi      vfat    umask=0077 0 2
UUID=ee578df1-6236-4eea-b6e5-62579060b349 /              btrfs   subvol=/@,defaults,noatime,autodefrag,compress=zstd 0 0
UUID=ee578df1-6236-4eea-b6e5-62579060b349 /home          btrfs   subvol=/@home,defaults,noatime,autodefrag,compress=zstd 0 0
UUID=ee578df1-6236-4eea-b6e5-62579060b349 /var/cache     btrfs   subvol=/@cache,defaults,noatime,autodefrag,compress=zstd 0 0
UUID=ee578df1-6236-4eea-b6e5-62579060b349 /var/log       btrfs   subvol=/@log,defaults,noatime,autodefrag,compress=zstd 0 0
UUID=4d31088c-9a64-43de-8864-9fa9ef0f7aea swap           swap    defaults,noatime 0 0

now everything is restored :tada:

1 Like