Oracle Exalogic Elastic Cloud Software
 
Exalogic Elastic Cloud X5-2 Hardware
 

PLA-Support>Eng Systems>Exalogic/OVCA>Oracle Exalogic>MW: Exalogic Core
 

In this Document

Symptoms

Cause

Solution

References

Created from <SR 3-13769119541>

Applies to:

Symptoms

In Exalogic Virtual 2.0.6.2.X Virtual racks, Issue of Compute Node(s) shown as UNKNOWN status in OVM Manager BUI console is seen.

At the time of the issue rediscovery Job of Compute Node from OVM Manager BUI fails with below "ObjectNotFoundException: No such object" Error.

Cause

This issue happens due to OVM Manager Database corruption. 

Solution

To fix the OVM Manager Database corruption, follow below OVM fix-up procedure.

1. Take backup of the Control vServers EC1, PC1 & PC2 prior to proceeding with fixup procedure. This is very important step.
   
   
2. Stop EMOC and PC services

   • Stop PC1 & PC 2 service on PC1 & PC2 Control vServers using below command. Do this task on one Proxy controller at a time.

   /opt/sun/xvmoc/bin/proxyadm stop -w -v

   Wait for a minute and check the status to see if it is stopped

   /opt/sun/xvmoc/bin/proxyadm status

   • Stop the EC1 service on EMOC Control vServer using below command

   /opt/sun/xvmoc/bin/ecadm stop -w -v

   Wait till the EMOC service shutdown complete. This may take 5 minutes

   Check the Status and make sure EC1 service is stopped

   /opt/sun/xvmoc/bin/ecadm status
3. Shutdown the OVMM service and implement the fixup procedure.

   • Stop OVMM service using below command

   # service ovmm stop

   • Run the fixup procedure on OVMM as follows

   # cd /u01/app/oracle/ovm-manager-3/ovm_upgrade/bin 
   # sh ovm_upgrade.sh --dbtype=OracleDB --fixup --dbuser=<user> --dbpass=<password> --dbhost=localhost --dbport=<database-port> --dbsid=elctrldb

   Following are the default user, dbport and dbsid for Exalogic OVM Control vServer.

   user = ovs
   dbport = 1521
   dbsid = elctrldb
   
   If we are using above default values the fix-up command will look as follows:
   
   sh ovm_upgrade.sh --dbtype=OracleDB --fixup --dbuser=ovs --dbpass=welcome1 --dbhost=localhost --dbport=1521 --dbsid=elctrldb
     
   Below is snippet we see when running above fix-up command on OVM DB. As you can see in highlighted line below snippet it is ObjectNotFoundException for CN01 which is marked as UNKNOWN from OVMM BUI, and OVM fixup script is fixing the corruption on CN01 object.
   
   
   Copying xsl files to patch path location
   

   Copying deleted classes files to patch path location
   2018-04-18 17:04:51,709 INFO Oracle OVM Manager Upgrade Processor
   2018-04-18 17:04:51,709 INFO
   2018-04-18 17:04:51,710 INFO Upgrade Initialization Starting
   2018-04-18 17:04:51,892 INFO Oracle Distributed Object Fabric (ODOF): Copyright (C) 2007, 2012 Oracle. All rights reserved.
   2018-04-18 17:04:51,892 INFO ODOF Version: 1.0.0.0
   2018-04-18 17:04:51,892 INFO Initializing...
   2018-04-18 17:04:52,274 INFO Initialization Complete!
   2018-04-18 17:04:52,275 INFO Upgrade Initialization Complete
   2018-04-18 17:04:52,275 INFO Database Fixup starting
   2018-04-18 17:04:54,389 INFO Fixing up Exchange... (196258 objects)
   2018-04-18 17:05:04,381 INFO Fixup has processed 1000 of 196258 objects
   2018-04-18 17:05:04,882 ERROR ObjectNotFoundException: ONF reference to object 582736: class="VirtualMachineProxy", from object 513: class="ServerDbImpl"
   2018-04-18 17:05:04,894 INFO Committing fixup changed object 513: 'eltestcn01.test.com' (class=ServerDbImpl)
   2018-04-18 17:05:08,093 INFO Fixup has processed 2000 of 196258 objects
   2018-04-18 17:05:09,154 INFO Fixup has processed 3000 of 196258 objects
   2018-04-18 17:05:10,108 INFO Fixup has processed 4000 of 196258 objects

   • Start the OVMM service using below command.

   # service ovmm start

   Once OVMM service is started, login to OVM Manager BUI console and check the status of Compute Node.
   
   

4. Proceed with starting the EMOC and PC services as follows

   • start the EC1 service on EMOC Control vServer using below command

   /opt/sun/xvmoc/bin/ecadm start -w -v

   Wait till the EMOC service is started complete. This may take 5 minutes

   Check the Status and make sure EC1 service is started

   /opt/sun/xvmoc/bin/ecadm status

   • start PC1 & PC2 services on PC1 & PC2 Control vServers using below command. Do this task on one Proxy controller at a time.

   /opt/sun/xvmoc/bin/proxyadm start -w -v

   Wait for a minute and check the status to see if it is started

   /opt/sun/xvmoc/bin/proxyadm status

    

If the issue still exists after implementing above fixup procedure, there may be a need to rebuild the OVM database. Please contact Support by opening Service Request for further assistance if above procedure does not fix the issue. 

References

This solution has no attachment