Restore failing with error boot: value does not match the regex pattern

[noel.hoskins]

Can anyone help shed some light on the error I have please.

We have a 3 node cluster in production that has an active subscription and a 2 node cluster without a subscription that we use for R&D. I have created a VM in the R&D cluster, tested it, backed it up and I am trying to now restore it to the production cluster. The back job completes without error and the restore job completes without error. However the VM fails to find a bootable hardrive on the first boot and then, when stopped, any of the changes I try to make to the hardware or options fail with the error "boot: value does not match the regex pattern". If I detach the hard drive I can then navigate through the configurations in the GUI without the error.

One difference is the VM was built in R&D node on storage with a name of "vm-data" and the production nodes all have storage names as "local-lvm" but I select / change that as part of the restore job and the restore is good.

Also the 2 clusters are slightly different versions. production cluster = pve6.2-12 and the R&D cluster = pve6.4-6

Is this a version mismatch, a storage name mismatch, cluster config / subscription mismatch or something different?

Thanks

 

[Stoiko Ivanov]

Also the 2 clusters are slightly different versions. production cluster = pve6.2-12 and the R&D cluster = pve6.4-6

I would suggest to upgrade both clusters to the same minor version
also consider upgrading to PVE 7.1 (6.4 will be EOL in 4 months)

I hope this helps!

 

[noel.hoskins]

Thanks Stoiko, I wil schedule this in, though I am assuming some downtime at some point to get an upgrade complete to 7.1 right?

Also can you confirm that this is the cause of the failure? I would be frustrating to wait untill some downtime to be scheduled in, perform the upgrade only to find that it's related to storage naming or something else.

 

[Stoiko Ivanov]

Thanks Stoiko, I wil schedule this in, though I am assuming some downtime at some point to get an upgrade complete to 7.1 right?

yes - I would definitely plan for a potential downtime and ensure that working backups are present!

In most cases you might keep all your guests running by migrating them to another node before running the upgrade - see the upgrade page in the wiki:
https://pve.proxmox.com/wiki/Upgrade_from_6.x_to_7.0

 

[noel.hoskins]

Back on this as we now are starting the process of planning and upgrading so slightly off topic now.......

If I migrate the VMs to a different node and cluster currently running 6.2, to free nodes of active VMs, then one node at a time follow the process to go from version 6.2 to 6.4 first then on to 7.1, will I break my cluster running say node 3 as 7.1 and nodes 1 and 2 still at 6.2? and also will I still be able to migrate VMs from node to node in the cluster whilst they are all on different versions?

I don't want to be stuck where I have nodes 1 and / or node 2 heavily loaded with VMs and nowhere to offload them because I have comitted node 3 to a higher version.

Thanks

 

[Stoiko Ivanov]

In general we do not recommend to mix versions in a cluster.
for the time of the upgrade (which should be kept relatively short) running the latest 6.4 and the new 7 in parallel should work
* I would really recommend to first perform the upgrade to the latest 6.4 for all nodes - and then start the upgrade to 7
* migrating to a node with a newer version should always work - for migrations to older versions this cannot be guaranteed in general

I hope this helps!