Using the Oracle Toolkit in a HP Serviceguard Cluster README Revision: B.06.00, August 2010
link /ORACLE_TEST1 to /home/oracle.
This configuration makes it possible to run several Oracle instances
on one node, facilitating failover/failback of Oracle packages
between nodes in the cluster.
An alternate setup for Oracle 10g/11g single instance database
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
If you are using LVM/VxVM (supported on Oracle 10gR2 and 11g only)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
For nodes that will run only one database instance, the configuration
described above could be simplified by installing Oracle entirely on
the shared volume/file system (/ORACLE_TEST0, in our example). The
installation is simplified, since no symbolic links are required. In
this case, all Oracle binaries and configuration is migrated to the
standby node on failover.
NOTE: This setup is not supported if Oracle 10g Release 1 is
configured with LVM/VxVM. If Oracle 10g Release 1 is configured with
LVM/VxVM then local configuration is recommended. The above
configuration is supported in Oracle 10g Release 2 and Oracle 11g,
but subject to a condition that Oracle's Automatic Storage Management
(ASM) is not configured on that node.
If you are using CFS (supported on Oracle 10gR1, 10gR2 and 11gR1)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In this set-up, entire Oracle can be optionally installed on a CFS
mounted file system (/ORACLE_TEST0, in our example) with all Oracle
binaries and configuration files in CFS mounted file system.
NOTE: An Oracle license must be installed on every node that might
potentially run Oracle in the cluster.
Set up additional database logical volumes, with LVM
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
It is possible to have the database reside on the same volume
group/logical volume as $ORACLE_HOME/dbs, but more commonly, the
database will probably reside on several volume groups and logical
volumes, which all need to be shared among the nodes that are able to
run the Oracle instances. Again, using a naming convention for the
volume group(s) that includes the instance name (${SID_NAME}) could
be used to associate a volume group with a unique instance.
EXAMPLE:
Raw disk access for Oracle data: