Upgrading ProxMox v1.1 to v1.5 doesn't work....

oeginc

Member
Mar 21, 2009
133
0
16
I tried following the upgrade instructions listed here http://pve.proxmox.com/wiki/Downloads, but this step is failing:

wget http://pve.proxmox.com/debian/dists/lenny/pve-upgrade-1.1-to-1.4

Which is obviously the main part... How can I upgrade from v1.1 to v1.4? I am running in a clustered environment as well which just adds to the confusion.

Worst case scenario, have the containers changed? Can I just vzdump everything to one machine, stop the cluster, rebuild the other 3 machines with v1.5 in a cluster, scp the dumps over and vzrestore... then rebuilding the last machine and finally add it to the cluster?

Thanks in advance!
 
Which is obviously the main part... How can I upgrade from v1.1 to v1.4? I am running in a clustered environment as well which just adds to the confusion.

We do not maintain that upgrade path any longer. If you are familiar with 'apt' you can use 'apt-get' to upgrade. Else I suggest you reinstall form CD and migrate (offline) the VMs to the new node.
 
We do not maintain that upgrade path any longer. If you are familiar with 'apt' you can use 'apt-get' to upgrade. Else I suggest you reinstall form CD and migrate (offline) the VMs to the new node.

I was under the impression that going from v1.1 to v1.4 wouldn't work using apt-get because too many things have changed (and hence why the scripts were written to help facilitate those changes).

Ok, so will my migration work that I've outlined above?
 
with the introduction of 1.5 (three kernel branches) the update script does not work anymore. as almost all users already migrated we stopped maintaining the script. (Martin updated the wiki page today)

but if you are familiar with apt, you can still upgrade but a reinstall should be faster and cleaner, but depends on your knowledge and setup.

You can install a 1.1 test system as a KVM guest and test the upgrade process (run Proxmox VE inside Proxmox VE).
 
with the introduction of 1.5 (three kernel branches) the update script does not work anymore. as almost all users already migrated we stopped maintaining the script. (Martin updated the wiki page today)

but if you are familiar with apt, you can still upgrade but a reinstall should be faster and cleaner, but depends on your knowledge and setup.

You can install a 1.1 test system as a KVM guest and test the upgrade process (run Proxmox VE inside Proxmox VE).

I understand that, that's why I asked if the upgrade I outlined above would work... Here it is again in perhaps simpler terms...

1. Disassociate the cluster back into separate servers (how do I do this?)
2. Shutdown all of the virtual machines
3. Use vzdump to backup all of the containers to .tgz
4. Move all .tgz to a remote storage location
5. Reformat all 4 boxes
6. Reinstall ProxMox VE 1.5+
7. Create a new cluster
8. Copy the .tgz back to their respective machines
9. vzrestore each container
10. Done?!?

Thanks for any help you may have on this matter, I am SURE I'm not the only person in this situation and detailed instructions would go a long way.
 
yes, looks good.

ad 1:

  • stop the ClusterSync and ClusterTunnel service on each node (Via web interface)
  • remove the cluster config on all all nodes (rm /etc/pve/cluster.cfg)
  • restart pvedaemon (/etc/init.d/pvedaemon restart)
 
I'm sure many users have not upgraded yet for just as many reasons.

Do I understand right that adding a 1.5 node to the 1.x cluster and migration (without 'online' selected) to that new node works from the older nodes?
 
no, I suggest a backup/restore (vzdump) if you want to reinstall. or a dist-upgrade.