Proxmox HA moved a VM to wrong host and unable to move back

[Benji99]

Hi folks,

I'm experiencing a strange issue and hoping to get some help.

I have a cluster of 6 Proxmox Servers (all running 8.2.4)
4 of them have access to an iSCSI disk that is hosting the storage for VMs
I have HA enabled for some VMs onl the 4 hosts with iSCSI storage.

It appears that the HA moved one of these VMs to a host that doesn't have access to the iSCSI storage. Obviously it wasn't successful and the VM won't start.
I'm now also unable to move it back to one of the functioning hosts. When I try to Migrate I get an error message that storage iscsi lvm is not available on node (500). The error makes sense as iSCSI is indeed not available.

This VM has it's primary disk on iSCSI along with a CloudInit Drive.

Any tips on successfully moving this VM back to a host that has access to the iSCSI disk?
Please let me know if you need more detailed logs...

Thanks

 

[sw-omit]

1. If you look at the iscsi within proxmox, do you still see the disks there?
2. What I would personally at least attempt would be to disconnect the drives, migrate the VM without the drives, and then attach them again. And if you can disconnect but can't migrate, temporarily assign them to a new VM, or just create a new VM and then migrate the drives to that VM.

 

[alexskysilk]

Any tips on successfully moving this VM back to a host that has access to the iSCSI disk?

as long as the vm is turned off, you can just drop to shell and

mv /etc/pve/nodes/currentnode/qemu-server/vmid.conf /etc/pve/nodes/destinationnode/qemu-server

 

[Benji99]

as long as the vm is turned off, you can just drop to shell and

mv /etc/pve/nodes/currentnode/qemu-server/vmid.conf /etc/pve/nodes/destinationnode/qemu-server

Thanks for the reply, will try that, does it matter which node to run this on? On the node that currently has the VM?

 

[sw-omit]

Since the /etc/pve is shared along the cluster, no, every server will have the same files there and they sync between the servers right away.

 

[Benji99]

Since the /etc/pve is shared along the cluster, no, every server will have the same files there and they sync between the servers right away.

Just did and fixed my issue. the VM started successfuilly on the right host. Thanks for the simple fix