Register a VM from shared storage

[Tmanok]

Hi Everyone,

Hopefully this isn't too obvious but after googling and surfing around I couldn't find it this evening. I'm trying to register a VM in a shared NFS storage, it is apart of a cluster but one of my cluster nodes went down and I want to bring one of the VMs that was running on it back to life on a different node. The power supply of a DL320 died on me and it unfortunately lacks redundant power supplies. When I access the web interface I can see the node and the VM ID that I want, and I know for a fact that it is in the shared storage that was attached to both the dead node and the node that I want to run the now offline VM on.

Please let me know what the steps are to "register a VM" similar to how in ESXi you can either create, or register an existing VM using it's VMX and VMDK files.
Thank you very much for your time, FYI I don't want to recreate the VM using the old storage, I just want the VM to come back to life without fussing with it.

Tmanok

 

[apoc]

That (registering vms) works "different" on proxmox. Configuration is in a different place, so independent.
The Data you see on disk only contains the virtual HDDs.
The configuration is stored under /etc/pve/qemu-server.
If you have not backed up those files and can copy/restore them there, afaik there is no other way than rebuilding the vms using the disk files you have on your NFS share.

 

[LnxBil]

I'm trying to register a VM in a shared NFS storage, it is apart of a cluster but one of my cluster nodes went down and I want to bring one of the VMs that was running on it back to life on a different node.

You have a working PVE cluster, one node failed and now you cannot start the VM directly on another node in the same cluster? That is the purpose of a cluster, isn't it? This is definitely possible.

Worst case, just move the VM configuration file to a new node:

mv /etc/pve/nodes/<old-node>/qemu-server/<vmid>.conf /etc/pve/nodes/<new-node>/qemu-server

 

[Tmanok]

Hi Everyone,

O.K. Thank you for the clarification, I hadn't taken a look at the folders recently and my head has been stuck in ESXi land from my work.

The configuration is stored under /etc/pve/qemu-server

Yes of course this makes much more sense, darn how did I forget? I'm unfamiliar with how vCenter / vSphere implements its migration, there must be some sort of database that handles the VMX files so that two "nodes" or "HVs" don't touch the VMX files at the same time... Trouble of course is that which node owns the database server / how would you ensure high-availability (maybe I'm tired again and this is much easier than I think).

I will ressurect the VM then possibly with just the disk, as you can tell this hasn't been critical if it's gone more than one day without operation. The real issue is that I'm neither on-site nor do I really have the time to work on this.
Thanks everyone, I'll see what I can do soon, feel free to suggest anything I may have overlooked about clustering (maybe the automated backups will have a new-ish configuration file?)