3 Replies Latest reply: Feb 28, 2012 6:27 AM by HartmutStreppel RSS

    Sun Geo Cluster doubt


      Can you tell me when Sun Geo Cluster will make DR up automatically?

      What are all the possibilities for geo cluster need manual intervention to make the primary or secondary site up?
        • 1. Re: Sun Geo Cluster doubt
          I recommend to read the Oracle Solaris Cluster Geographic Edition Overview document:

          In a nutshell, the decision to either perform a switchover or takeover from the primary cluster to the secondary cluster is always a manual decision, based on the policies around your business continuity plan.
          Typically you don't want to have that decision automated, instead a responsible human decides if there is a need to activate the secondary cluster and specifically declare there is a disaster to act upon.

          Now once that decision is made, Geo Edition then automates all the operations required to perform the switchover or takeover once initiated.

          • 2. Re: Sun Geo Cluster doubt
            One more question,

            In my setup, we have two sets of 5+1 geo cluster in primary and secondary site respectively.

            Assume a scenario, In primary cluster, We are having Node1, Node2, Node3, Node4, Node5 and Failover server.

            if Node3 and Failover server fails, ultimately the application resource group should be online at secondary site.

            Is it right?
            • 3. Re: Sun Geo Cluster doubt
              a) there is no automatic failover in an OSC Geographic Edition setup between clusters. There is only automatic failover within a cluster.
              b) Where a failed application/service fails over to depends on the nodelist property. If, in your case, the nodelist contains only node3 and the failover node, and both of these nodes fail, the service will - very probably - fail as well, as there is no other node to host the service.

              OSC Geographic Edition, btw. monitors clusters but not individual nodes or services. So even, if you had automatic failover between the two sites and clusters setup somehow, it would not solve this problem. If you think, the proposed failure scenario is a real one then you would have to extend your nodelist settings.