D How to Upgrade to Oracle Grid Infrastructure 11g Release 2

D.1 About Oracle ASM and Oracle Grid Infrastructure Installation and Upgrade

In past releases, Oracle ASM was installed as part of the Oracle Database installation. With Oracle Database 11g Release 2 (11.2), Oracle ASM is installed when you install the Oracle Grid Infrastructure software. Oracle ASM shares an Oracle home with Oracle Clusterware when installed in a cluster such as with Oracle RAC or with Oracle Restart on a standalone server. The directory in which Oracle Clusterware and Oracle ASM is installed is called the Grid home.

If you have an existing Oracle ASM instance, you can either upgrade it at the time that you install Oracle Grid Infrastructure, or you can upgrade it after the installation, using Oracle ASM Configuration Assistant (ASMCA). However, a number of Oracle ASM features are disabled until you upgrade Oracle ASM, and Oracle Clusterware management of Oracle ASM does not function correctly until Oracle ASM is upgraded, because Oracle Clusterware only manages Oracle ASM when it is running in the Oracle Grid Infrastructure home. For this reason, Oracle recommends that if you do not upgrade Oracle ASM at the same time as you upgrade Oracle Clusterware, then you should upgrade Oracle ASM immediately afterward.

You can perform out-of-place upgrades to an Oracle ASM instance using ASMCA. In addition to running ASMCA using the graphic user interface (GUI), you can run ASMCA in non-interactive (silent) mode.

In earlier releases, you could use Database Upgrade Assistant (DBUA) to upgrade either an Oracle Database, or Oracle ASM. With Oracle Database 11g Release 2 (11.2), you can use DBUA to upgrade only an Oracle Database instance. To upgrade Oracle ASM separately, you use ASMCA.

D.2 Restrictions for Oracle Clusterware and Oracle ASM Upgrades

Be aware of the following restrictions and changes for upgrades to Oracle Grid Infrastructure installations, which consists of Oracle Clusterware and Oracle ASM.

• Upgrade Your Existing Installation to the Minimum Required Version

• Apply Required Patches

• Other Requirements and Restrictions

• Changes to the Upgrade Process

D.3 Understanding Out-of-Place and Rolling Upgrades

Oracle Clusterware upgrades can be rolling upgrades, in which a subset of nodes are brought down and upgraded while other nodes remain active. Oracle ASM 11g Release 2 (11.2) upgrades can be rolling upgrades. If you upgrade a subset of nodes, then a software-only installation is performed on the existing cluster nodes that you do not select for upgrade. Rolling upgrades avoid downtime and ensure continuous availability while the software is upgraded to a new version.

During an out-of-place upgrade, the installer installs the newer version in a separate Oracle Clusterware home. Both versions of Oracle Clusterware are on each cluster member node, but only one version is active. By contrast, an in-place upgrade overwrites the software in the current Oracle Clusterware home.

To perform an out-of-place upgrade, you must create new Oracle Grid Infrastructure homes on each node. Then you can perform an out-of-place rolling upgrade, so that some nodes are running Oracle Clusterware from the original Oracle Clusterware home, and other nodes are running Oracle Clusterware from the new Oracle Grid Infrastructure home.

If you have an existing Oracle Clusterware installation, then you upgrade your existing cluster by performing an out-of-place upgrade. An in-place upgrade of Oracle Clusterware 11g Release 2 is not supported.

D.4 Preparing to Upgrade an Existing Oracle Clusterware Installation

If you have an existing Oracle Clusterware installation, then you upgrade your existing cluster by performing an out-of-place upgrade. You cannot perform an in-place upgrade. The following sections describe the tasks you should complete before starting an upgrade:

• Verify System Readiness for Patches and Upgrades

• Gather the Necessary System Information

• Upgrade to the Minimum Required Oracle Clusterware Version

• Unset Environment Variables Used By Oracle Software

D.5 Using CVU to Verify System Readiness for Patches and Upgrades

If you are completing a patch update of Oracle Clusterware or Oracle ASM, then after you download the patch software and before you start to patch or upgrade your software installation, review the Patch Set Release Notes that accompany the patch to determine if your system meets the system requirements for the operating system and the hardware.

Navigate to the staging area for the upgrade, where the runcluvfy command is located, and run the command runcluvfy stage -pre crsinst -upgrade to check the readiness of your Oracle Clusterware installation for upgrades. Running runcluvfy with the -pre crsinst -upgrade flags performs system checks to confirm if the cluster is in a correct state for upgrading from an existing clusterware installation.

The command uses the following syntax, where variable content is indicated by italics:

runcluvfy stage -pre crsinst -upgrade [-n node_list] [-rolling] -src_crshome
src_Gridhome -dest_crshome dest_Gridhome -dest_version dest_version
[-fixup[-fixupdir fixupdirpath]] [-verbose]

The options used in this command are:

Use the Cluster Verification Utility (CVU) to assist you with system checks in preparation for patching or upgrading. You can run CVU before starting the upgrade, however, the installer runs the appropriate CVU checks automatically, and prompts you to fix problems before proceeding with the upgrade.

C:\stage> runcluvfy.bat stage -pre crsinst -upgrade -n node1,node2 -rolling 
-src_crshome C:\app\11.2.0\grid -dest_crshome C:\app\11.2.0.3\grid
 -dest_version 11.2.0.3.0 -fixup -fixupdirpath C:\user\grid\fixup -verbose

D.6 Back Up the Oracle Software Before Upgrades

Before you make any changes to the Oracle software, Oracle recommends that you create a backup of the Oracle software and databases.

D.7 Performing a Rolling Upgrade of Oracle Clusterware

Use the following procedure to upgrade Oracle Clusterware from an earlier release to Oracle Clusterware 11g Release 2. You can also choose to upgrade Oracle ASM during the upgrade of Oracle Clusterware.

1. If there are non-clustered, or single-instance, Oracle databases that use Oracle ASM running on any of the nodes in the cluster, they must be shut down before you start the upgrade. Listeners associated with those databases do not have to be shut down.

   Note:

   Oracle recommends that you leave Oracle Real Application Clusters (Oracle RAC) instances running during the Oracle Clusterware upgrade. During the upgrade process, the database instances on the node being upgraded are stopped and started automatically during the upgrade process.

2. Stop the Enterprise Manager Database Control on each node. Run the following command to stop Database Control:

   C:\> emctl stop dbconsole
   

3. Start the installer, and select the option to upgrade an existing Oracle Clusterware and Oracle ASM installation.

4. On the node selection page, select all nodes.

   Note:

   In contrast with releases prior to Oracle Clusterware 11g Release 2, all upgrades are rolling upgrades, even if you select all nodes for the upgrade. If you are upgrading from Oracle Clusterware 11g release 11.2.0.1 to release 11.2.0.2, then all nodes are selected by default. You cannot select or de-select the nodes.

5. Select installation options as prompted.

6. If prompted to do so, run the scripts as directed.

7. When the Oracle Clusterware upgrade is complete, if an earlier version of Oracle ASM is installed, then the installer starts ASMCA to upgrade Oracle ASM to release 11.2. You can upgrade Oracle ASM at this time, or upgrade it later.

   Oracle recommends that you upgrade Oracle ASM at the same time that you upgrade Oracle Clusterware. Until Oracle ASM is upgraded, Oracle databases that use Oracle ASM cannot be created. Until Oracle ASM is upgraded, the Oracle ASM management tools in the Oracle Grid Infrastructure 11g Release 2 (11.2) home (for example, SRVCTL) will not work.

   Note:

   After the upgrade, if you set the Oracle Cluster Registry (OCR) backup location manually to the earlier release Oracle Clusterware home (CRS home), then you must change the OCR backup location to the Oracle Grid Infrastructure home (Grid home). If you did not set the OCR backup location manually, then this issue does not concern you.

   Because upgrades of Oracle Clusterware are out-of-place upgrades, the CRS home for the earlier release Oracle Clusterware cannot be the location of the OCR backups for Oracle Clusterware 11g Release 2. Backups in the old Oracle Clusterware home could be deleted.

D.8 Performing a Rolling Upgrade of Oracle ASM

After you have completed the Oracle Clusterware 11g Release 2 (11.2) upgrade, if you did not choose to upgrade Oracle ASM when you upgraded Oracle Clusterware, then you can do it separately using ASMCA to perform rolling upgrades.

You can use ASMCA to complete the upgrade of Oracle ASM separately, but you should perform the upgrade as soon as possible after you upgrade Oracle Clusterware, because Oracle ASM management tools such as SRVCTL will not work until Oracle ASM is upgraded.

D.9 Updating Oracle Enterprise Manager Target Parameters

Because Oracle Clusterware 11g Release 2 (11.2) is an out-of-place upgrade, the upgraded software has a new location for the Grid home. The path for the Oracle Clusterware home in some Oracle Enterprise Manager parameter files must be changed. If you do not change the parameter, then you will encounter errors such as "cluster target broken" on Oracle Enterprise Manager Database Control or Grid Control.

Use the following procedure to resolve this issue:

1. Log in to Database Control or Grid Control.

2. Select the Cluster tab.

3. Click Monitoring Configuration.

4. Update the value for Oracle Home with the new path for the Grid home.

D.10 Checking Cluster Health Monitor Repository Size After Upgrading

If you are upgrading from a prior release using IPD/OS to Oracle Grid Infrastructure Release 2 (11.2.0.3), then you should review the Cluster Health Monitor repository size (the CHM repository). Oracle recommends that you review your CHM repository needs, and enlarge the repository size if you want to maintain a larger CHM repository.

By default, the CHM repository size for release 11.2.0.3 and later is a minimum of either 1GB or 3600 seconds (1 hour).

To enlarge the CHM repository, use the following command syntax, where RETENTION_TIME is the size of CHM repository in number of seconds:

oclumon manage -repos resize RETENTION_TIME

The value for RETENTION_TIME must be more than 3600 (one hour) and less than 259200 (three days). If you enlarge the CHM repository size, then you must ensure that there is local space available for the repository size you select on each node of the cluster. If there is not sufficient space available, then you can move the repository to shared storage.

For example, to set the repository size to four hours:

C:\> oclumon manage -repos resize 14400

D.11 Troubleshooting Upgrade Issues

This section describes some commonly encountered problems during the upgrade process.

• OUI-35078: Error in updating env on nodes nodename

D.12 Downgrading Oracle Clusterware After an Upgrade

After a successful or a failed upgrade to Oracle Clusterware 11g Release 2 (11.2), you might want to restore Oracle Clusterware to the previous version.

To restore Oracle Clusterware 11g Release 2 (11.2) to an earlier version of Oracle Clusterware, use the following procedure. In the following procedure, the local node is the first node on which the rootupgrade script was run. The remote nodes are all other nodes that were upgraded.

1. Use the downgrade procedure for the release to which you want to downgrade.

   Downgrading to Oracle Clusterware 11g Release 1 (11.1) or Oracle Clusterware 10g Release 2 (10.2):

   There is currently no supported downgrade procedure. Instead, you must reinstall Oracle Clusterware for the target release and then reapply all required patches.

   Downgrading to releases 11.2.0.1, 11.2.0.2, or 11.2.0.3:

   Use the command syntax rootcrs.pl -downgrade -oldcrshome oldGridHomePath -version oldGridversion, where oldGridhomepath is the path to the previous release Oracle Grid Infrastructure home, and oldGridversion is the release to which you want to downgrade. For example:

   rootcrs.pl -downgrade -oldcrshome C:\app\grid -version 11.2.0.1.0
   

   Note:

   Ensure that Oracle Clusterware version is specified in the correct format. For example, 11.1.0.1.0.

   If you want to stop a partial or failed 11g Release 2 (11.2) installation and restore the previous release of Oracle Clusterware, then use the -force flag with this command.

2. After the rootcrs.pl -downgrade script has completed on all remote nodes, on the local node use the command Grid_home\perl\bin\perl\rootcrs.pl -downgrade -lastnode -oldcrshome oldGridhome -version oldGridversion [-force], where oldGridhome is the home of the earlier Oracle Clusterware installation, and oldGridversion is the release number of the earlier Oracle Clusterware installation.

   For example:

   C:\app\11.2.0.3\grid\perl\bin\perl> rootcrs.pl -downgrade -lastnode 
   -oldcrshome C:\app\grid -version 11.2.0.1.0
   

   This script downgrades the OCR. If you want to stop a partial or failed 11g Release 2 (11.2) installation and restore the previous release Oracle Clusterware, then use the -force flag with this command.

3. Run the as prompted by the rootcrs.pl -downgrade command.

4. Log in as the Grid Infrastructure installation owner and run the following commands, where C:\app\11.2.0\grid is the location of the new Grid home (11.2):

   setup.exe -nowait -waitforcompletion -ignoreSysPrereqs
    -updateNodeList -silent CRS=false ORACLE_HOME=C:\app\11.2.0\grid
   

   Append -cfs to the end of the above command is the Oracle Grid home is a shared home.

5. If the newer version of Oracle Grid Infrastructure is 11.2.0.2 or later

   1. If you are downgrading from Oracle Grid Infrastructure Release 2 (11.2.0.4) to an earlier release of 11.2, after execution of the upgrade scripts is completed on all cluster nodes, run the following command from the 11.2.0.4 Grid home:

      acfsroot uninstall
      

      After the command completes, from the earlier release Grid home, run the following command as an Administrator user:

      acfsroot install
      

   2. On each node, start Oracle Clusterware from the earlier release Oracle Clusterware home using the command crsctl start crs. For example, where the earlier release home is C:\app\grid, use the following command on each node:

      C:\app\grid\bin> crsctl start crs
      

6. Run the gridconfig.bat script as prompted by the rootcrs.pl -downgrade command from the earlier release Grid home in sequence on each member node of the cluster. After you complete this task, downgrade is completed.

   Running gridconfig.bat from the earlier release Oracle Clusterware installation home restarts the Oracle Clusterware stack, starts up all the resources previously registered with Oracle Clusterware in the older version, and configures the old initialization scripts to run the earlier release Oracle Clusterware stack.

7. As the Grid infrastructure installation owner, run the following command, where C:\app\grid represents the home directory of the earlier Oracle Clusterware installation, for example:

   setup.exe -nowait -waitforcompletion -ignoreSysPrereqs
    -updateNodeList -silent CRS=true ORACLE_HOME=C:\app\grid
   

   Append -cfs to the end of the above command is the Oracle Clusterware home is a shared home.