3 Replies Latest reply: Jun 25, 2012 1:42 PM by imgrid RSS

    Upgrade of Grid Clusterware from 11.2.0.1 to 11.2.0.2 failed on Node 2

    Oceaner
      We have a 2 Node cluster on 11.2.0.1. We went for upgrade to 11.2.0.2.

      On Node 1, rootupgrade.sh ran successfully but it failed on Node2.

      Now,We removed the node 2 from the cluster and "olsnodes -l -s" shows node2 inactive and unpinned.

      OCR shows crs software active version as 11.2.0.1 and cluster active version is 11.2.0.2. Node 2 ( bad node), the command deconfig was already executed. The binaries were not functional.

      On Node1, database is up and running and our applications are up and running as well. But since node 2 is not functional , we need to bring it up.

      OCR shows crs software active version as 11.2.0.1 and cluster active version is 11.2.0.2 To fix this, Is it advisable to run
       crsctl set crs activeversion 11.2.0.2.0 
      on node 1. Will it make any impact on running node. We cannot shutdown this node (Node1) since everything is running on this node. or
      ./rootupgrade.sh -force
      Thanks.
        • 1. Re: Upgrade of Grid Clusterware from 11.2.0.1 to 11.2.0.2 failed on Node 2
          Pavan Kumar
          Hi,

          If I remember correctly, even I faced some issue while upgrading the 112.01 to 11.2.0.2, difficult to remember

          Have you performed rolling upgrade ?
          Have you applied require patches prior to performing the upgrade. ?

          Note - Remember that rootupgrade.sh is restartable


          - Pavan Kumar N
          • 2. Re: Upgrade of Grid Clusterware from 11.2.0.1 to 11.2.0.2 failed on Node 2
            imgrid
            Pavan,

            I am collegue of Navneet and I was performing this task. I was doing this upgrade in rolling upgrade mode only.

            W have acfs mount point on cluster, which was not getting down when rootupgrade.sh was trying to shutdown the crs on node2. We bounced server but failed and reran the rootupgrade.sh but it failed. We did applied the required PSU prior to running rootupgrade.sh.

            Now as Navneet mentioned earlier OLD GRID HOME and NEW GRID HOME both home is not working and has been deconfigured.

            Since the rootupgarde.sh has not been completed on last node, So upgarde is not complete. As result we are seeing different Clusterware softwareversion and different active version.

            Clusterware Version :- crsctl query crs softwareversion =11.2.0.2.0
            Active Version: -crsctl query crs activeversion =11.2.0.1.0

            Now to set active version to clusterware version and complete the upgrade process, we need to run ./rootupgrade.sh -force or crsctl set crs activeversion 11.2.0.2.0 on first node where rootupgarde.sh was successfully completed.

            Now my question is since node1 is already available to users, If I run the above mentioned command, would this command effect running services of node1 and is there any possibilty that databases can go offline ?
            • 3. Re: Upgrade of Grid Clusterware from 11.2.0.1 to 11.2.0.2 failed on Node 2
              imgrid
              OCR shows crs software active version as 11.2.0.1 and cluster active version is 11.2.0.2. Node 2 ( bad node), the command deconfig was already executed. The binaries were not functional.

              I would like to above line. Replacing Node2 (Bad node) to Node1 (good node)

              OCR shows crs software active version as 11.2.0.1 and cluster active version is 11.2.0.2. Node 1 ( Good node), the command deconfig was already executed. The binaries were not functional.

              Edited by: 818838 on Jun 25, 2012 11:41 AM