Hi, we have been running a two-node pacemaker/corosync cluster for the past half year quite successfully. Now we extended it to a three-node cluster with quorum and are still happy. But here is a question:
We have a group of resources ("ClusterGroup") that contains a number of
services (tftp, dhcp-server, bind, backuppc and the fs of the shared drbd) to
provide the basic services in the network. This group has to run on one f the
first nodes as only these have the drbd-resources containing the needed
filesystems. I added an anti-location constraint to keep the cluster from
running the drbd stuff on the third node. I also added anti-location constraint
to prevent the clustergroup from running on the third node.
But I still get error-messages ($OCF_ERR_NOT_INSTALLED) when the cluster tries
to monitor the resources on the third node. I don't want that. I want the
cluster to either not even run monitor (for these resources) on the third node
or to ignore monitor-errors from these resources on the third node.
How to achieve this? Do I have to add anti-location rules for all the
resources in that group individually?
Thanks for your input,
Arnold
PS: anti-location == location with -inf
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Pacemaker mailing list: [email protected] http://oss.clusterlabs.org/mailman/listinfo/pacemaker Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
