1 Reply Latest reply: Feb 23, 2012 2:30 AM by T.F.-Oracle RSS

    2 node Sun Cluster 3.2, resource groups not failing over.

    user13431458
      Hello,

      I am currently running two v490s connected to a 6540 Sun Storagetek array. After attempting to install the latest OS patches the cluster seems nearly destroyed. I backed out the patches and right now only one node can process the resource groups properly. The other node will appear to take over the Veritas disk groups but will not mount them automatically. I have been working on this for over a month and have learned alot and fixed alot of other issues that came up, but the cluster is just not working properly. Here is some output.

      bash-3.00# clresourcegroup switch -n coins01 DataWatch-rg
      clresourcegroup: (C776397) Request failed because node coins01 is not a potential primary for resource group DataWatch-rg. Ensure that when a zone is intended, it is explicitly specified by using the node:zonename format.

      bash-3.00# clresourcegroup switch -z zcoins01 -n coins01 DataWatch-rg
      clresourcegroup: (C298182) Cannot use node coins01:zcoins01 because it is not currently in the cluster membership.
      clresourcegroup: (C916474) Request failed because none of the specified nodes are usable.
      bash-3.00# clresource status
      === Cluster Resources ===

      Resource Name Node Name State Status Message
      ------------- --------- ----- --------------
      ftp-rs coins01:zftp01 Offline Offline
      coins02:zftp01 Offline Offline - LogicalHostname offline.

      xprcoins coins01:zcoins01 Offline Offline
      coins02:zcoins01 Offline Offline - LogicalHostname offline.

      xprcoins-rs coins01:zcoins01 Offline Offline
      coins02:zcoins01 Offline Offline - LogicalHostname offline.

      DataWatch-hasp-rs coins01:zcoins01 Offline Offline
      coins02:zcoins01 Offline Offline

      BDSarchive-res coins01:zcoins01 Offline Offline
      coins02:zcoins01 Offline Offline

      I am really at a loss here. Any help appreciated.

      Thanks
        • 1. Re: 2 node Sun Cluster 3.2, resource groups not failing over.
          T.F.-Oracle
          My advice is to open a service call, provided you have a service contract with Oracle. There is much more information required to understand that specific configuration and to analyse the various log files. This is beyond what can be done in this forum.

          From your description I can guess that you want to failover a resource group between non-global zones. And it looks like the zone coins01:zcoins01 is reported to not be in cluster membership.

          Obviously node coins01 needs to be a cluster member. If it is reported as online and has joined the cluster, then you need to verify if the zone zcoins01 is really properly up and running.
          Specifically you need to verify that it reached the multi-user milestone and all cluster related SMF services are running correctly (ie. verify "svcs -x" in the non-global zone).

          You mention Veritas diskgroups. Note that VxVM diskgroups are handled in the global cluster level (ie. in the global zone). The VxVM diskgroup is not imported for a non-global zone. However, with SUNW.HAStoragePlus you can ensure that file systems on top of VxVM diskgroups can be mounted into a non-global zone. But again, more information would be required to see how you configued things and why they don't work as you expect it.

          Regards
          Thorsten