proper reinstall procedure? pls help

[stooovie]

I had catastrophic failure of the PM (8.2.2) host (boot drive failed). I do have backups. I even have a host backup but it turns out

a) the restore with this script is ONLY possible with the same PVE version the backup has been created with (no idea which was it created with)
b) I have no idea how to actually restore the LXC and VM backups
c) My LXCs and VMs are on a separate NVMe drive, so they should be largely unaffected. I see it only in node/Disks/LVM-Thin but how do I actually connect/assign/use it. There's also an another ZFS drive that I don't see in PVE UI at all.

Please help. Thanks!

EDIT: the backed-up host files are from 8.1.0. I can install that but will the backups from newer versions work in it?

 

[stooovie]

When I try to restore the host backup, it runs for a bit, then I get "Connection closed: 1006" and subsequent attempts at connecting to the PM GUI returns HTTP Error 501 (???).

 

[aaron]

What is the storage config from the backup? (/etc/pve/storage.cfg) That would be a first guide on how the disk on which the backups are stored is configured. Depending on this, we would know how to make it available again so that the guest backups can be restored.

 

[stooovie]

Right! I'm getting somewhere.

/etc/pve/storage.cfg from the backup is:

dir: local
    path /var/lib/vz
    content iso,vztmpl,snippets,backup
    shared 0


lvmthin: storage
    thinpool storage
    vgname storage
    content images,rootdir
    nodes pve


cifs: Public
    path /mnt/pve/Public
    server 192.168.0.216
    share Public
    content vztmpl,iso
    username stooovie


zfspool: zfs
    pool zfs
    content rootdir,images
    mountpoint /zfs
    nodes pve


dir: backup
    path /mnt/pve/backup
    content iso,images,rootdir,snippets,vztmpl,backup
    is_mountpoint 1
    nodes pve
    shared 0

I just copied it from the backup to the corresponding folder via SFTP and now I have storage. I see I can also copy over files from

/etc/pve/nodes/pve/lxc

and I have my LXCs Same with /etc/pve/nodes/pve/qemu-server for VMs. Cool!

My host backup is three months old so there are some mismatches in numbers but I think I'll be able to get it back to work now. Thank you!

 

[aaron]

and I have my LXCs Same with /etc/pve/nodes/pve/qemu-server for VMs. Cool! i think I'll be able to get it back to work now. Thank you!

well, those are just the config files. It is possible that it works because their disk images were not stored on the disk on which you installed Proxmox VE to. Probably the ZFS pool or the network share?

Otherwise, the procedure would have been to mount the filesystem of the disk on which you have the backups and either create a new directory storage at the mount point, or re-use the old storage.cfg and make sure it is mounted in /mnt/pve/backup. You could have then restored the guest backups.

 

[stooovie]

Most of the LXCs and VMs were on a separate NVMe, as I stated in the OP.

I have five drives:

1) a boot SSD drive
2) NVMe that holds most VMs and LXCs
3) a smaller SSD formated with ZFS that holds some LXCs
4) a USB drive for backups
5) a SMB mount for backups

For the ZFS volume, I used the

zpool import -f

command with the numeric id as instructed by the GUI and it remounted fine. I write this here for posterity.