A little background first:
I run Proxmox on just one single node/host with local onboard NVME storage, because that covers my current needs just fine, and I backup all important VM's to an SMB share on a NAS in a different location with gigabit connection. I also have a 2nd identical machine on hand, just to be able to restore and be up and running asap in case of HW failure.
I'm not using PBS because I like to keep my setup simple, I already own the NAS, I do not have a spare computer with the storage capacity for backups, and I do not want to use PBS in an unsupported way.
The task:
I recently migrated my environment to the 2nd machine, when I wanted to add 2TB HDD storage and upgrade from PVE 8.x to PVE 9.1 at the same time, planning to avoid any pesky in-place-upgrades if I could. So I ran a final backup on all VM's, mounted the backup storage on the new freshly installed node and restored the VM's (no clustering on the nodes). And hey, it worked fine... until I looked at the backup view on each VM.
The problem:
The naming scheme of the backup files are: vzdump-qemu-<VMID>-yyyy_mm_dd-hh_mm_ss.vma.zst
There is no node name in the backup filename by default, and I can't find a way to turn it on either.
I did not restore my VM's in the same order that I created them, as I did not think it would matter. But it did!
Now all of a sudden, my DNS-server lists the backup files of my file-server and so on, and I think it is because the file-server now has the same ID that the DNS-server previously had - on the previous node.
None of the VM's have the same ID as they had before and no 2 VM's have switched clean. It's like the chair game where everyone sits down in a random free chair (ID), so backup files are now a mess! At least until my retention policy deletes the old ones and creates new backups.
But that is another problem. The retention policy will now delete the wrong backups! Backups of VM's that only has one single backup without a backup plan will now get deleted, and old backups of VM's with backup plans will require manual intervention as they are now listed in VM's with no backup plan.
All of this, because the node name is not in the filename... Is it possible to add, please?
I run Proxmox on just one single node/host with local onboard NVME storage, because that covers my current needs just fine, and I backup all important VM's to an SMB share on a NAS in a different location with gigabit connection. I also have a 2nd identical machine on hand, just to be able to restore and be up and running asap in case of HW failure.
I'm not using PBS because I like to keep my setup simple, I already own the NAS, I do not have a spare computer with the storage capacity for backups, and I do not want to use PBS in an unsupported way.
The task:
I recently migrated my environment to the 2nd machine, when I wanted to add 2TB HDD storage and upgrade from PVE 8.x to PVE 9.1 at the same time, planning to avoid any pesky in-place-upgrades if I could. So I ran a final backup on all VM's, mounted the backup storage on the new freshly installed node and restored the VM's (no clustering on the nodes). And hey, it worked fine... until I looked at the backup view on each VM.
The problem:
The naming scheme of the backup files are: vzdump-qemu-<VMID>-yyyy_mm_dd-hh_mm_ss.vma.zst
There is no node name in the backup filename by default, and I can't find a way to turn it on either.
I did not restore my VM's in the same order that I created them, as I did not think it would matter. But it did!
Now all of a sudden, my DNS-server lists the backup files of my file-server and so on, and I think it is because the file-server now has the same ID that the DNS-server previously had - on the previous node.
None of the VM's have the same ID as they had before and no 2 VM's have switched clean. It's like the chair game where everyone sits down in a random free chair (ID), so backup files are now a mess! At least until my retention policy deletes the old ones and creates new backups.
But that is another problem. The retention policy will now delete the wrong backups! Backups of VM's that only has one single backup without a backup plan will now get deleted, and old backups of VM's with backup plans will require manual intervention as they are now listed in VM's with no backup plan.
All of this, because the node name is not in the filename... Is it possible to add, please?