12.5 Migrating Standalone Oracle Grid Infrastructure Servers to a Cluster
If you have an Oracle Database installation using Oracle Restart (that is, an Oracle Grid Infrastructure installation for a standalone server), and you want to configure that server as a cluster member node, then complete the following tasks:
-
Inspect the Oracle Restart configuration with
srvctl
using the following syntax, where db_unique_name is the unique name for the database, and lsnrname is the name of the listener:srvctl config database -db
db_unique_name
srvctl config service -db
db_unique_name
srvctl config listener -listener
lsnrname
Write down the configuration information for the server.
-
Stop all of the databases, services, and listeners that you discovered in step 1.
-
If present, unmount all Oracle Automatic Storage Management Cluster File System (Oracle ACFS) file systems.
-
Log in as
root
, and change directory to Grid home/crs/install
. For example:# cd /u01/app/12.2.0/grid/crs/install
-
Unconfigure the Oracle Grid Infrastructure installation for a standalone server (Oracle Restart), using the following command:
# roothas.sh -deconfig -force
-
Prepare the server for Oracle Clusterware configuration, as described in this document. In addition, you can install Oracle Grid Infrastructure for a cluster in the same location as Oracle Restart, or in a different location.
Installing in the Same Location as Oracle Restart
-
Proceed to step 7.
Installing in a Different Location than Oracle Restart
-
Set up Oracle Grid Infrastructure software in the new Grid home software location as described in Installing Only the Oracle Grid Infrastructure Software.
-
Proceed to step 7.
-
-
Set the environment variables as follows:
export oracle_install_asm_UseExistingDG=true or false export oracle_install_asm_DiskGroupName=disk_group_name export oracle_install_asm_DiskDiscoveryString=asm_discovery_string export oracle_install_asm_ConfigureGIMRDataDG=true or false export oracle_install_asm_GIMRDataDGName=disk_group_name
-
As the Oracle Grid Infrastructure installation owner, run the installer.
You can complete the installation either interactively or in the silent mode. If you perform a silent installation, save and stage the response file as described in Recording Response Files.
-
After saving the response file, run the command:
$ Grid_home/gridSetup.sh -silent -responseFile $ORACLE_HOME/GI.rsp
-
Run
root.sh
. -
Mount the Oracle ASM disk group used by Oracle Restart.
-
If you used Oracle ACFS with Oracle Restart, then:
-
Start Oracle ASM Configuration Assistant (ASMCA). Run the
volenable
command to enable all Oracle Restart disk group volumes. -
Mount all Oracle ACFS file systems manually.
-
-
Add back Oracle Clusterware services to the Oracle Clusterware home, using the information you wrote down in step 1, including adding back Oracle ACFS resources. For example:
/u01/app/grid/product/12.2.0/grid/bin/srvctl add filesystem -device /dev/asm/db1 -diskgroup ORestartData -volume db1 -mountpointpath /u01/app/grid/product/12.2.0/db1 -user grid
-
Add the Oracle Database for support by Oracle Grid Infrastructure for a cluster, using the configuration information you recorded in step 1. Use the following command syntax, where db_unique_name is the unique name of the database on the node, and nodename is the name of the node:
srvctl add database -db
db_unique_name
-spfile -pwfile-oraclehome $ORACLE_HOME -node
nodename
-
For example, first verify that the
ORACLE_HOME
environment variable is set to the location of the database home directory. -
Next, to add the database name
mydb
, enter the following command:srvctl add database -db mydb -spfile -pwfile -oraclehome $ORACLE_HOME -node node1
-
Add each service to the database, using the command
srvctl add service
. For example, addmyservice
as follows:srvctl add service -db mydb -service myservice
-
-
Add nodes to your cluster, as required, using the Oracle Grid Infrastructure installer.
See Also:
Oracle Clusterware Administration and Deployment Guide for information about adding nodes to your cluster.