Currently, I have a VM (essentially a NAS) that has a number of physical disks directly passed through to it. How can I move these disks to a new VM (within the same node)? Can/should I first move the disks to a storage pool (e.g., LVM) and then attach the disks to the new VM from there?
Move passed-through disks to new VM?
- Thread starter dubdub
- Start date
Do you mean this disk passthrough?: https://pve.proxmox.com/wiki/Passthrough_Physical_Disk_to_Virtual_Machine_(VM)
In that case you could just stop that old and new new VM. Then open the VMs config files, remove the lines in the old VMs config file and add them to the new VMs config file.
In that case you could just stop that old and new new VM. Then open the VMs config files, remove the lines in the old VMs config file and add them to the new VMs config file.
Yes, my disks are passed through as per that link. I'll have a look at the config files, thanks.
As an aside, I passed through the disks when first setting up the VM because it seemed like the right approach. Having read a little more since, I wonder what the best practice is - are there advantages in putting the disks in a storage pool (e.g., LVM)?
As an aside, I passed through the disks when first setting up the VM because it seemed like the right approach. Having read a little more since, I wonder what the best practice is - are there advantages in putting the disks in a storage pool (e.g., LVM)?
What do you excatly mean? Creating a LVM inside your VM or create a LVM on your host and then using virtual disks (LVs) for that VM?
In both cases its virtualized as the method described in the link isn't directly passing through real physical hardware.
Always depends on your needs and usecase. Especially concering backups/snapshots and being able to access those data on a server failure from another host.
In both cases its virtualized as the method described in the link isn't directly passing through real physical hardware.
Always depends on your needs and usecase. Especially concering backups/snapshots and being able to access those data on a server failure from another host.
The usage of a reasonable HBA [in IT-mode] (or raid-controller if you want to use hardware raid) and PCIe-passthrough it to the VM.
Thanks for the replies. For clarification, I'm passing the disks (ext4) from pve to an openmediavault VM, where they are pooled using mergerfs. Disks are a mix of consumer grades, plugged directly into the motherboard (i.e., SATA). So no HBA - yet anyway. I only got into virtualisation about six months ago, so still at the beginning of the learning curve.
And thanks for this. Couldn't have been easier.