Couldn't start virtual machines after Proxmox 4.1 upgrade

Administration
#1

We have a running Proxmox 4.0 setup on 2 Debian Jessie server with a few container machines. Recently we decided to upgrade both of the servers and Proxmox components with:

$ sudo apt-get update
$ sudo apt-get dist-upgrade

After upgrading Proxmox 4.1, we got a strange problem on one of the servers.

Corosync service failed to start and gives following errors but we don’t have a cluster before:

Jan 10 02:35:37 server1 pmxcfs[1065]: [quorum] crit: quorum_initialize failed: 2
Jan 10 02:35:37 server1 pmxcfs[1065]: [confdb] crit: cmap_initialize failed: 2
Jan 10 02:35:37 server1 pmxcfs[1065]: [dcdb] crit: cpg_initialize failed: 2
Jan 10 02:35:37 server1 pmxcfs[1065]: [status] crit: cpg_initialize failed: 2

We see that there is a corosync.conf file under /etc/pve directory and it is not possible to delete it. It comes from lxcfs mount.

Disabling corosync service on start also doesn’t make sense.

Any suggestions?

#2

If you’re sure about you’re not using any clustering feature and you have a corosync.conf file by a mistake, you can remove the corosync configuration file with following steps:

  • First of all, bakcup your current proxmox sqlite config:
$ sudo cp /var/lib/pve-cluster/config.db /root/config.db.backup
  • Stop the pve-cluster service:
$ sudo systemctl stop pve-cluster.service
  • Remove the row with named corosync.conf from sqlite tree table:
$ sudo sqlite3 /var/lib/pve-cluster/config.db
sqlite> DELETE FROM tree WHERE name='corosync.conf';
sqlite> .quit
  • Restart the server
#3

@demirten Thanks you for your answer. It really helped me. When removing a node I got the same error message in proxmox but that has now been solved with your solution. THANKS