4 Replies Latest reply: Apr 25, 2013 5:07 AM by DK2010 RSS

    CRS down but VIP is still up

    martin75
      GI Version: 11.2.0.3
      Platform : RHEL 5.8


      In one of our 2-node clusters , the CRS is down
      $ crsctl check crs
      CRS-4638: Oracle High Availability Services is online
      CRS-4535: Cannot communicate with Cluster Ready Services
      CRS-4529: Cluster Synchronization Services is online
      CRS-4533: Event Manager is online
      SCAN IP and SCAN Listeners were down. But VIP and the Node listener running on VIP is up and the clients can connect to the DB using this VIP and port of Node listener.

      How is this possible when CRS is down ?
        • 1. Re: CRS down but VIP is still up
          FreddieEssex
          Hi take a look at Issue #3 in the following note: Top 5 Grid Infrastructure Startup Issues [ID 1368382.1]

          One of the possible causes is that a specific resource is not online. Presumably some of your resources are still up.

          PS: Why are clients connecting directly via vip when you have SCAN configured?
          • 2. Re: CRS down but VIP is still up
            martin75
            Thanks Freddie.
            The issue got fixed after a cluster shutdown plus Linux reboot.

            My question was , Who actually manages VIP? Since VIP was up even when CRS was down, Now I know that its not CRS
            • 3. Re: CRS down but VIP is still up
              FreddieEssex
              Hi,

              The error message you received may have been due to one of the CRS daemons being down and not all of the CRS stack.

              See the table here for all of the CRS components:

              http://docs.oracle.com/cd/B28359_01/rac.111/b28255/intro.htm#CJHECEDD
              • 4. Re: CRS down but VIP is still up
                DK2010
                Hi,

                As per my point of view

                VIP is managed by Oracle Clusterware(vip,evm,crs,css..), CRS is one of component of Clusterware
                Oracle Clusterware has 2 heartbeats 1)Network heartbeat 2)Disk heartbeat
                If a node does not send a network heartbeat then node is evicted from cluster
                If disk heartbeat (voting disk) is not updated in <I/Otimeout>, then node is evicted from cluster