Proxmox OS (Rollback) Feature?

[Dualized]

Is there a feature for this just in case due to breaking changes?

Proxmox OS (Rollback) Feature​

As currently I performed an Update and now I Can't even login to my Proxmox

Thanks for your time!

 

[dietmar]

No, there is no rollback. You need to find the problem instead.

Is the host network online? Can you load the GUI with the browser?

 

[Dualized]

Yeah, never mind I can login again, it was a hardware issue.
I was just wondering if Proxmox had like an option to run the OS and have a secondary option ( similar to the BIOS ) to roll back, but no worries. I will just save the configuration file.
I might need to look into best practices just in case Proxmox OS takes a dump on recovering the VM's

 

[bbgeek17]

I was just wondering if Proxmox had like an option to run the OS

Your best bet is to research using a filesystem level snapshot: https://docs.oracle.com/cd/E18752_01/html/819-5461/ghzvz.html
It would only make sense for a short period of time, unless you have a lot of free space available for snapshot change.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[killmasta93]

i recommend run zfs snapshot -r rpool@now before updating on proxmox then to rollbackup boot up ubuntu and revert

 

[alexskysilk]

I was just wondering if Proxmox had like an option to run the OS and have a secondary option ( similar to the BIOS ) to roll back, but no worries. I will just save the configuration file.

Proxmox can be used in one of two ways- standalone or in a cluster.

In a cluster environment individual nodes are meant to be taken as cattle, not pets. this means that nodes have no special value in themselves so there really isnt any point to "backing up". worse, you dont want the same node to be ABLE to connect to the cluster as different.

individually, pve nodes are meant to only act as hypervisors, so the only real importance is the payload- you can either just restore your backups onto a fresh new pve install, OR if you want to be more sophisticated you can keep a copy of /etc/pve/qemu-server and /etc/pve/lxc which you can then just copy back onto a new install.

If you intend to use the host in other ways (eg, provide services other than virtualization) then you need to treat your host as any other server and back up the whole thing, and revert to backup.

 

[UdoB]

On very large upgrades (e.g real dist-upgrade with potential damage) I utilized ZFS checkpoints. It will "snapshot" the whole pool = all volumes w/ all snapshots at once.

This is not applicable in all situations and not meant to be used on a daily basis - a "rollback" will roll back all datasets/volumes at once! But it is another option one should know about Of course I evacuated or shut down all VMs first for this approach.

man zpool-checkpoint has the syntax.