VMmanager: Automatic troubleshooting

From ISPWiki
Jump to: navigation, search
Hierarchy: VMmanager KVM ->OTHER
VMmanager Cloud ->OTHER

Automatic troubleshooting

VMmanager finds and automatically fixes many common problems with cluster nodes, and tracks their current statuses. A problem node won't be available on the list of virtual machines that can be added. Once VMmanager detects that the problem has been resolved, the cluster node will become available.

For more information please read the article Problem notifications.

The following issues are currently registered:

ssh connection failed

VMmanager detected that ssh request to an existing node failed or VMmanager authentication failed.

Pausible causes of this error:

  • The private key or its location has been changed (the sshPrivatKey and sshPublicKey parameters in the configuration file)
  • The VMmanager public key is not present in the /root/.ssh/authorized_keys file on the cluster node.
  • Connection to the cluster node failed (check the network's settings)

VMmanager failed to connect to libvirt on the cluster node

  • libvirt is not running on the cluster node (the service is suspended or deleted)
  • The connection socket is located in a non-standard directory (normally the socket is located in /var/run/libvirt/libvirt-sock if you run CentOS or in /run/libvirt/libvirt-sock if you run Debian)
  • ssh connection failed (see above)

Failed to access the NFS-storage with OS templates

  • The NFS-storage with OS templates is located on the master node. Check the export rules.
root@msk-vkvm-master-190:~# exportfs 
/nfsshare     	10.7.2.10
/nfsshare     	10.7.2.11
/nfsshare     	10.7.2.12
/nfsshare     	10.7.2.13

The node must be on the export list:

  • Check the iptables rules. Ports for NFS must be open.
  • Make sure the information for /nfsshare on the cluster nodes in /etc/fstab is correct. The IP address of the master node must be specified.
  • Try remounting /nfsshare manually.