Problem with pve-upgrade-1.9-to-2.0

[stef1777]

Sorry, these 2 test nodes was in cluster when with 1.9. But with 2.0, I don't do a cluster as the network do not support multicast. But yesterday, I tried to setup the cluster with 2.0 just to verify that multicast is not working...

 

[dietmar]

OK, so we can safely remove the cluster configuration? If so, try:

# /etc/init.d/pve-cluster stop
# rm /etc/cluster/cluster.conf
# pmxcfs -l
# rm /etc/pve/cluster.conf
# /etc/init.d/pve-cluster stop
# /etc/init.d/pve-cluster start

 

[stef1777]

Thank, work fine now.

I removed the cluster config as requested and run again import.

sd-33092:~# ./pve-upgrade-1.9-to-2.0 --import
starting import
import file: /etc/pve/user.cfg
import file: /etc/pve/datacenter.cfg
import file: /etc/pve/storage.cfg
import file: /etc/pve/openvz/101.conf
import done

And the openvz container is back.

Ya!

 

[ryansinn]

I ran the upgrade on an up to date 1.9 system that has been migrated from 1.3 all the way to 1.9 over the last few years.

I had to run some commands manually during the installation to get the system to install properly:

Here's how the installation process went:

./pve-upgrade-1.9_to_2.0
dpkg --remove dab

./pve-upgrade-1.9_to_2.0
dpkg --configure procps

./pve-upgrade-1.9_to_2.0

received an email error that "libc6" config failed to restart some services; specifically, apache2.

Other than that, the upgrade went great!

 

[Smanux]

I got the same issue with the upgrade script, pmxcfs kept saying "Unable to get local IP address". I checked the configuration and /etc/hosts does have an entry for the name defined in /etc/hostname.

The catch is that /etc/hostname defines a fully qualified domain name (foo.example.com) and it seems pmxcfs doesn't handle this properly. I added the sub domain "foo" in /etc/hosts and now pmxcfs works fine.