On 08/31/2009 04:30 PM, Jakov Sosic wrote:
Hi.
I have a problem that one of my 3 nodes is sent to RMA to hardware
dealer. I want to know is it possible to disable that one node, because
clurgmgrd is not starting any services because cman is unable to fence
the missing node...
Aug 31 22:02:14 node03 fenced[5082]: fence "node02" failed
Aug 31 22:02:19 node03 fenced[5082]: fencing node "node02"
Aug 31 22:04:38 node03 fenced[5082]: agent "fence_ipmilan" reports:
Rebooting machine @ IPMI:x.y.z.w...ipmilan: Failed to connect
after 30 seconds Failed
The best course of action to take would be to remove that missing node
from your cluster configuration using conga, system-config-cluster, or
by hand editing /etc/cluster/cluster.conf. As long as it exists in the
configuration then the other nodes will expect it to join the cluster,
and they will attempt to fence it when they try to join the cluster and
see it is not present. When you remove it from the configuration by
hand, you will also want to set two_node="1 and expected_votes="1" in
the cman tag like so
<cman two_node="1" expected_votes="1"/>
If you are using one of the graphical tools these options should be
added for you. They ensure that if one of your remaining two nodes
dies, the one that is left can maintain quorum by itself.
After making these changes and propagating the new configuration to both
nodes, you will want to restart the cluster service on both nodes or
just reboot both nodes.
-John
--
Linux-cluster mailing list
Linux-cluster@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/linux-cluster
