The method of applying an Oracle patch (e.g. Oracle PSR: Patch Set Release) varies depending on database mounting.
If a database does not need to be mounted before and after patch application, be sure to stop RMS before applying a patch.
If a database needs to be mounted before and after patch application, monitoring must be cancelled before applying a patch as follows:
Information
The following are offered from Oracle Database 11g R2 as a method of applying Oracle PSR (Patch Set Release).
In-place upgrade
It is a method of installing PSR on existing ORACLE_HOME by the superscription. It is offered from previous version.
For further details, refer to "3.2.2.1 Patch Application by In-place upgrade"
Out-of-place upgrade
It is a method of installing PSR in new ORACLE_HOME which is different from existing ORACLE_HOME. It is offered from Oracle Database 11g R2.
For further details, refer to "3.2.2.2 Patch Application by Out-of-place upgrade"
This section discusses how to apply PSR in Oracle Database 11g R1 or Oracle Database 11g R2/12c R1 by In-place upgrade.
Start userApplication (Operation node)
Start userApplication that includes the Oracle instance and listener resources if it is stopped.
Maintenance mode (Operation node)
Start maintenance mode for userApplication on operation node.
Example
If you start maintenance mode for all userApplication, execute the following command.
# /opt/SMAW/SMAWRrms/bin/hvutil -M on
Confirm that state of userApplication has become "Maintenance" on all nodes.
Monitoring disruption of Oracle resource (All nodes)
Stop monitoring the Oracle instance and listener resources by executing the "hvoradisable" command on all nodes.
Example
If you stop monitoring all the Oracle resources on a local node, execute the following command.
# /opt/FJSVclora/sbin/hvoradisable -a
Confirm that states of the Oracle instance and listener resources have become "Warning" on all nodes.
Information
In standby operation, applying Oracle PSR (Patch Set Release) on the standby node without stopping monitoring Oracle resources might cause problems because Oracle database is monitored on the standby node. Therefore it is necessary to stop monitoring them on the standby node.
Oracle database stop (if necessary)
# su - <Oracle user> $ sqlplus /nolog SQL> connect / as sysdba SQL> shutdown immediate
Patch application
Apply patches depending on the destination of the Oracle Database software.
See
For further details, refer to the Oracle Database manual.
Oracle database startup (if stopped)
# su - <Oracle user> $ sqlplus /nolog SQL> connect / as sysdba SQL> startup
Restart Oracle resource monitoring (All nodes)
Restart monitoring the Oracle instance and listener resources again using the hvoraenable command on all nodes.
Example
If you restart monitoring all the Oracle resources on a local node, execute the following command.
# /opt/FJSVclora/sbin/hvoraenable -a
Confirm that states of the Oracle instance and listener resources have become the previous state ("Online" or "Offline") from "Warning" on all nodes.
Stop maintenance mode (Operation node)
Stop maintenance mode of userApplication on operation node.
Example
If you stop maintenance mode for all userApplication, execute the following command.
# /opt/SMAW/SMAWRrms/bin/hvutil -M off
Confirm that state of userApplication has become the previous state ("Online", "Standby" or "Offline") from "Maintenance" on all nodes.
This section discusses how to apply PSR by Out-of-place upgrade in Oracle Database 11g R2/12c R1.
Note
Upgrade to local disk if the Oracle Database software is located on local disk.
Upgrade to shared disk if the Oracle Database software is located on shared disk.
Changing the placement (local disk, shared disk) of the Oracle Database software at the upgrade does not support.
Start userApplication (Operation node)
Start userApplication that includes the Oracle instance and listener resources if it is stopped.
Maintenance mode (Operation node)
Start maintenance mode for userApplication on operation node.
Example
If you start maintenance mode for all userApplication, execute the following command.
# /opt/SMAW/SMAWRrms/bin/hvutil -M on
Confirm that state of userApplication has become "Maintenance" on all nodes.
Monitoring disruption of Oracle resource (All nodes)
Stop monitoring the Oracle instance and listener resources by executing the "hvoradisable" command on all nodes.
Example
If you stop monitoring all the Oracle resources on a local node, execute the following command.
# /opt/FJSVclora/sbin/hvoradisable -a
Confirm that states of the Oracle instance and listener resources have become "Warning" on all nodes.
Information
In standby operation, applying Oracle PSR (Patch Set Release) on the standby node without stopping monitoring Oracle resources might cause problems because Oracle database is monitored on the standby node. Therefore it is necessary to stop monitoring them on the standby node.
Stop Oracle database -dependent resources (Operation node)
If there are Oracle database -dependent resources, stop them manually before stopping Oracle database.
Applying PSR, and upgrade of database
Applying PSR (All nodes)
If Oracle Database Software is located in local disk, execute the Out-of-place upgrade on all nodes.
If Oracle Database Software is located in shared disk, execute the Out-of-place upgrade on operation node.
For information on how to execute the Out-of-place upgrade, refer to the Oracle Database manual.
Note
Select "Install database software only" on the "Select Installation Option" screen.
For details, refer to the Oracle Database manual.
Upgrade of database (Operation node)
Execute dbua (Database Upgrade Assistant) on operation node, and upgrade the database.
Note
Execute "<New ORACLE_HOME>/bin/dbua".
For details, refer to the Oracle Database manual.
Move the SPFILE to the shared disk (Operation node)
Move the SPFILE (server parameter file) to the shared disk when it is newly generated on a local disk after the upgrade of the database is executed. Afterwards, specify SPFILE path to PFILE (the initialization parameter file "<New ORACLE_HOME>/dbs/init<SID>.ora") or make a symbolic link file to SPFILE.
Note
If Oracle Database Software is located in shared disk, this step is not necessary.
See
For details, see "2.2.6 Oracle Database Creation and Setting".
PRIMECLUSTER Wizard for Oracle Settings
Login profile Setting Change (All nodes)
Open Oracle user's login profile with editor (e.g. vi), and change the setting of the ORACLE_BASE and ORACLE_HOME environment variable to new path. Execute this procedure on all nodes.
Example
Before changes
ORACLE_HOME=/u01/app/oracle/product/11.2.0/dbhome_1
After changes
ORACLE_HOME=/u01/app/oracle/product/11.2.0/dbhome_2
See
For details, see "2.2.2 Oracle Database Software Installation and Configuration".
Transfer PFILE to standby nodes (Operation node, Standby node)
Create tar archive file from "<New ORACLE_HOME>/dbs" on operation node.
Example
# su - oracle $ tar cfpvP /tmp/dbs.tar <New ORACLE_HOME>/dbs
After that, transfer the archive file from operation node to standby node by FTP, and expand it.
Example
# su - oracle $ tar xfpvP /tmp/dbs.tar
Note
If Oracle Database Software is located in shared disk, this step is not necessary.
In single-node cluster operation, this step is not necessary.
Do not use the cloracpy command of PRIMECLUSTER Wizard for Oracle on this step.
Set the access privilege of files which are expanded on the standby node same as the files on operating node.
See
For details, see "2.2.6 Oracle Database Creation and Setting".
The allocation of Oracle Net Services configuration file (All nodes)
Copy the Oracle Net Services configuration file from under the <Old ORACLE_HOME> to under the <New ORACLE_HOME>. Execute this procedure on all nodes
If Oracle Database Software is located in local disk, execute this procedure on all nodes.
If Oracle Database Software is located in shared disk, execute this procedure on operation node.
Example
$ cp -p <Old ORACLE_HOME>/network/admin/*.ora <New ORACLE_HOME>/network/admin/
Note
Edit "<New ORACLE_HOME>/network/admin/*.ora" file when there are parameters depending on ORACLE_BASE and ORACLE_HOME in the Oracle Net Services configuration file after copy.
Do not use the cloracpy command of PRIMECLUSTER Wizard for Oracle on this step.
Confirming Oracle Startup (Operation node)
Confirm that Oracle instance and Listener started. If not, start up them manually.
Note
The Oracle instance should start in the state of OPEN.
Confirm the process of the listener is as "<New ORACLE_HOME>/bin/tnslsnr".
If the process of the listener is as "<Old ORACLE_HOME>/bin/tnslsnr", stop it and start as "<New ORACLE_HOME>/bin/tnslsnr".
Start resources dependent on Oracle database (Operation node)
If resources that are dependent on Oracle database had been stopped on step 4, start them manually.
Restart Oracle resource monitoring (All nodes)
Restart monitoring the Oracle instance and listener resources again using the hvoraenable command on all nodes
Example
If you restart monitoring all the Oracle resources on a local node, execute the following command.
# /opt/FJSVclora/sbin/hvoraenable -a
Confirm that states of the Oracle instance and listener resources have become the previous state ("Online" or "Offline") from "Warning" on all nodes.
Stop maintenance mode (Operation node)
Stop maintenance mode of userApplication on operation node.
Example
If you stop maintenance mode for all userApplication, execute the following command.
# /opt/SMAW/SMAWRrms/bin/hvutil -M off
Confirm that state of userApplication has become the previous state ("Online", "Standby" or "Offline") from "Maintenance" on all nodes.
Confirming of switch to the standby node
Confirm the switch of userApplication to the standby node, and the correct start.
Note
In single-node cluster operation, this step is not necessary.
Confirm step 6 when you cannot correctly start userApplication on the standby node.
Restart service
Switch userApplication to the operation node if necessary, and restart service.
