When network traffic is blocked for one out of two nodes in a RHEL 6 pacemaker cluster with fence_scsi, both nodes show resources as "Started" and the fenced node shows both of them still online

Solution Unverified - Updated 2014-05-27T18:41:44+00:00 -

Issue

After a node is fenced by the fence_scsi agent, the cluster status becomes strange.
When network traffic is disrupted between nodes by adding iptables rules to simulate an outage, both nodes show themselves as owning the resources, and one node still shows both nodes Online while the other shows one Offline
Is it dangerous to see both nodes as having started the same resources and having different membership views after network traffic between them fails?

Red Hat Enterprise Linux (RHEL) 6 with the High Availability Add On
pacemaker
Nodes configured to use fence_scsi stonith devices
Some sort of network issue is occuring that results in a split between the nodes

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.