Hi I've came across an odd behavior, which might be considered as inconsistent. As we know, pacemaker doesn't allow deleting a resource that's running, but this doesn't produce same behavior every time.
Let's take a VM with a default stop timeout (90 seconds). First behavior -------------- # crm resource stop vm ; crm primitive delete vm This will result with: ERROR: resource vm is running, can't delete it Eventually, vm resource will stop and delete can be 'reexecuted'. It will be removed and everybody's happy. Ideally, IMHO, delete would be queued. Second behavior --------------- # crm cib new test # CIB_shadow="test" crm resource stop vm # CIB_shadow="test" crm primitive delete vm # crm cib commit test In this case shadow CIB is created and committed with the same changes as in first case. Pacemaker will accept both requests and it will try deleting the vm while stopping is in progress. This will ignore stop timeout for vm (90 seconds) and declare stopping as timed out after 20 seconds. Depending on cluster configuration, this might result in node being fenced. Shouldn't this be consistent? Best regards _______________________________________________ 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
