After configuring two-node Cluster (one on NW 6.5 sp7 and the other on NW 6.5 sp6) I started to test its behaviour during failover and failback procedures. Then, I noticed that the node with the higher IP address (sp6) does not failover to the other one (sp7) but rather puts into comatose state.

Looking at the KB I've found TID 10092849 explaining exactly the same situation and recommending to use CRM.NLM from April 6, 2004 or later. However, my nodes have higher versions of that module (1.80.04 from September 12, 2007 on sp7 and 1.80.03 from October 3, 2006 on sp6).

So, why doesn't the node with the higher IP address failover to the other one in our two-node cluster with the newer versions of CRM.NLM?

Sinisa