High Availability for Oracle ASM using HP Serviceguard Solutions, September 2010

The toolkit package parameter KILL_ASM_FOREGROUNDS determines whether the scripts should kill

the ASM foregrounds if the file descriptors are open on the dismounted disk group volumes. The ASM

PMON process will recover any foreground processes that were killed. This parameter can be set

either to yes or no only. Default value is “yes”.

The scripts will check to see if any ASM foreground processes of the form oracle<ASM_SID> have file

descriptors open on the dismounted disk group volumes of the database instance. If this parameter is

set to “yes”, and the ASM foregrounds have file descriptors open after dismount of the disk group,

they will be killed using a SIGTERM signal by the scripts. Before killing the ASM foreground

processes, the scripts will check every 3 seconds for a maximum of 12 seconds to see if the

foreground processes have closed the file descriptors. If the file descriptors are closed within this 12

second timeout, then volume groups will be deactivated without waiting for the completion of this

timeout. If this parameter is set to “no”, the ASM foregrounds will not be killed. However, when this

parameter is set to “no”, and if ASM foregrounds have file descriptors open, volume group

deactivation and hence the database instance package halt will fail.

Installing, configuring, and troubleshooting

Oracle ASM is part of the Oracle database server installation. No additional software from HP is

required to operate ASM in Serviceguard environment.

Oracle 11gR2 onwards, Oracle Clusterware needs to be installed to use Oracle ASM.

Oracle ASM and ASM disk groups may be configured at the time of creating a database or as a

separate step. Ensure that the LVM volume groups and raw disks have been prepared and activated

prior to ASM disk group configuration or reconfiguration, as discussed in this document. Use the

names of raw LVs contained in the LVM volume groups when Oracle commands and utilities require

the specification of the names of raw HP-UX devices for storing ASM disk groups. Ensure that each

database instance uses different disk groups.

It is important to note that Oracle 11gR2 onwards, by default, the spfile of the ASM instance is stored

on the ASM disk group. This is true only for the ASM instance and not for the database instance that

uses the ASM disk group. For an ASM instance, the ECM Toolkit cannot verify the existence of the

spfile on the ASM disk group. So, it is mandatory that a pfile is created from the spfile on the local

disk, at a location that ECMT can access. The value of the PFILE attribute of the ASM instance

package must point to the pfile that is created on the local disk.

Assume that the Serviceguard cluster, ASM instance, and one or more database instances are

already installed and configured.

• Halt the ASM and database instances.

• Configure the ASM MNP using the ECMT Oracle scripts provided by HP following the instructions

in the README file in the scripts bundle.

• Start the ASM instance on all the configured nodes by invoking cmrunpkg on the ASM MNP.

• Next, configure the database failover package using the ECMT Oracle scripts provided by HP

following the instructions in the README file in the scripts bundle.

• Start the database instance on the node by invoking cmrunpkg on the database package.

For troubleshooting, it may be useful to look at subdirectories for database and ASM instances under

$ORACLE_BASE/admin/, where log and trace files for these instances are deposited. Also look at

HP-UX syslog and Serviceguard package logs.