Hi,
We are currently testing proxmox VE 5.4-13 to become our new open-source virtualization solution.
Everything is working like a charm, all the features we miss in our current virtualization solution are available and functioning pretty good in proxmox.
Pretty awesome. We are planning to get a subscription!
We are currently using Ceph RBD images for our VM's and after some research we are able to use these within proxmox:
Our Ceph cluster is completely seperate.
Within our virtualization solution, we had to make the RBD disks and resources seperately and after that merge them together as a server (VM).
Therefor we used different and exotic names in our RBD images. These RBD names did not work in proxmox..
The RBD disk webmail_rbd wasn't recognized in proxmox.. so since we had not much experience in proxmox, we read something about proxmox name conventions.
Just as a precaution, we copied the rbd webmail_rbd to vm-100-disk-0.
Proxmox recognized it immediately so we could inject the disk and the VM started to boot.. Pretty cool..
So...we just renamed the original (test) RBD vm with:
rbd mv poolname/webmail_rbd poolname/vm-100-webmail-disk-0
We really wanted some in formation in the image, because vm-100-disk-0 would be just too indistinct.
But is there a way to run our proxmox VM's in a way, the VM's will run their disk device(s) as /dev/vdX instead of /dev/sdX ?
Our VM's in our current virt. solution are running their RBD disk with disk interface virtio and therefor the disks inside the VM's are presenting a /dev/vdX
We tried all the options in proxmox, but the only proper way the VM's can boot are with:
scsi0: poolname:vm-100-webmail-disk-0,size=30G
scsihw: virtio-scsi-single
scsi0: poolname:vm-100-webmail-disk-0,size=30G
scsihw: virtio-scsi-pci
The VM's will boot, but the "a start job is running for de-vda1.device" is not what we want in every VM.
How can we fix this? If there is no way in proxmox to run the VM's with /dev/vdX then I guess we have to change all the 125 Linux VM's manually.. :-/
Kind regards..
We are currently testing proxmox VE 5.4-13 to become our new open-source virtualization solution.
Everything is working like a charm, all the features we miss in our current virtualization solution are available and functioning pretty good in proxmox.
Pretty awesome. We are planning to get a subscription!
We are currently using Ceph RBD images for our VM's and after some research we are able to use these within proxmox:
Our Ceph cluster is completely seperate.
Within our virtualization solution, we had to make the RBD disks and resources seperately and after that merge them together as a server (VM).
Therefor we used different and exotic names in our RBD images. These RBD names did not work in proxmox..
The RBD disk webmail_rbd wasn't recognized in proxmox.. so since we had not much experience in proxmox, we read something about proxmox name conventions.
Just as a precaution, we copied the rbd webmail_rbd to vm-100-disk-0.
Proxmox recognized it immediately so we could inject the disk and the VM started to boot.. Pretty cool..
So...we just renamed the original (test) RBD vm with:
rbd mv poolname/webmail_rbd poolname/vm-100-webmail-disk-0
We really wanted some in formation in the image, because vm-100-disk-0 would be just too indistinct.
But is there a way to run our proxmox VM's in a way, the VM's will run their disk device(s) as /dev/vdX instead of /dev/sdX ?
Our VM's in our current virt. solution are running their RBD disk with disk interface virtio and therefor the disks inside the VM's are presenting a /dev/vdX
We tried all the options in proxmox, but the only proper way the VM's can boot are with:
scsi0: poolname:vm-100-webmail-disk-0,size=30G
scsihw: virtio-scsi-single
scsi0: poolname:vm-100-webmail-disk-0,size=30G
scsihw: virtio-scsi-pci
The VM's will boot, but the "a start job is running for de-vda1.device" is not what we want in every VM.
How can we fix this? If there is no way in proxmox to run the VM's with /dev/vdX then I guess we have to change all the 125 Linux VM's manually.. :-/
Kind regards..
Last edited:
