[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

[Cluster-devel] cluster.cman.nodename vanish on config reload



I just updated from 3.1.8 to latest STABLE32:

 

I use this cluster.conf:

 

# cat /etc/cluster/cluster.conf

<?xml version="1.0"?>

<cluster config_version="235" name="test">

  <cman keyfile="/var/lib/pve-cluster/corosync.authkey" transport="udpu"/>

  <clusternodes>

    <clusternode name="maui" nodeid="3" votes="1"/>

    <clusternode name="cnode1" nodeid="1" votes="1"/>

  </clusternodes>

  <rm>

    <pvevm autostart="0" vmid="100"/>

  </rm>

</cluster>

 

cman service starts without problems:

 

# /etc/init.d/cman start

Starting cluster:

   Checking if cluster has been disabled at boot... [  OK  ]

   Checking Network Manager... [  OK  ]

   Global setup... [  OK  ]

   Loading kernel modules... [  OK  ]

   Mounting configfs... [  OK  ]

   Starting cman... [  OK  ]

   Waiting for quorum... [  OK  ]

   Starting fenced... [  OK  ]

   Starting dlm_controld... [  OK  ]

Starting GFS2 Control Daemon: gfs_controld.

   Unfencing self... [  OK  ]

   Joining fence domain... [  OK  ]

 

And the corosync objdb contains:

 

# corosync-objctl|grep cluster.cman

cluster.cman.keyfile=/var/lib/pve-cluster/corosync.authkey

cluster.cman.transport=udpu

cluster.cman.nodename=maui

cluster.cman.cluster_id=1678

 

Note: there is a value for ‘nodename’ and ‘cluster_id’

 

Now I simply increase the version inside cluster.conf (on both nodes):

 

# cat /etc/cluster/cluster.conf

<?xml version="1.0"?>

<cluster config_version="236" name="test">

  <cman keyfile="/var/lib/pve-cluster/corosync.authkey" transport="udpu"/>

  <clusternodes>

    <clusternode name="maui" nodeid="3" votes="1"/>

    <clusternode name="cnode1" nodeid="1" votes="1"/>

  </clusternodes>

  <rm>

    <pvevm autostart="0" vmid="100"/>

  </rm>

</cluster>

 

And trigger a reload:

 

# cman_tool version -r –S

cman_tool: Error loading configuration in corosync/cman

 

And the syslog have more details:

 

Jul 10 13:28:25 maui corosync[488675]:   [CMAN  ] cman was unable to determine our node name!

Jul 10 13:28:25 maui corosync[488675]:   [CMAN  ] Can't get updated config version: Successfully read config from /etc/cluster/cluster.conf#012.

Jul 10 13:28:25 maui corosync[488675]:   [CMAN  ] Continuing activity with old configuration

 

Somehow the nodename and cluster_id values are removed from the corosync objdb:

 

# corosync-objctl|grep cluster.cman

cluster.cman.keyfile=/var/lib/pve-cluster/corosync.authkey

cluster.cman.transport=udpu

 

 

Any Idea why that happens?

 

- Dietmar

 


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]