I’m getting ready to put ApPak2 into our PPRD environment. And the problem has come up again. The 3rd node in the cluster problem. Yeah- recall in Vista 3 one installation script installed three nodes. In Vista 4, a separate installation script is used for any nodes over two. We are only running two nodes in Development or Preproduction. Our Operations and Engineering department has maxed out their VM capacity. No more storage. No more Nodes in Development and Preproduction environments until September or later.
A cluster is supposed to provide high availablity, fault tolerance, a measure of confidence that if one application instance on a single box goes haywire, others will rise to the occasion. But these Vista/BEA Weblogic clusters, installed separately from the initial cluster, with spotty session failover at best, and no JMS failover whatsoever … just complicate my life without doing anything they’re supposed to.
In our case, with no way to review, practice or test installing a third node in the cluster, we will be running our cluster through the peak fall period on two nodes.
Pass me those pecan turtle clusters please…