Cost-Effective High-Availability Solutions with HP Instant Capacity on HP-UX

Scripts for implementing failover with Serviceguard

While you can implement rights seizure as part of an automatic failover system, you must ensure

resources are seized appropriately and in a manner that will not cause problems when the failures

are corrected. The Instant Capacity software determines that a partition is down based on whether the

ping command is unsuccessful for the partition. However, ping can be unsuccessful even though the

system is functional (for example, if a network connection is interrupted). In this case, rights seizure

might be inappropriate, leaving workloads without the necessary resources.

Note:

These scripts are for illustrative purposes. They provide a starting point for

implementation and assume that all application workloads are under the

control of Serviceguard. The sample scripts should be modified to fit your

specific requirements. In particular, the start-up script uses remsh for a

remote operation, but a more secure implementation would use ssh (secure

shell). Also, the script should be made more robust by checking for errors

and other possible cluster states, and exiting with appropriate return status

values. The scripts are for customer-defined run/halt commands used in the

Serviceguard legacy package format; however, they can easily be

modified to use external_scripts with the Serviceguard modular package

format. These sample scripts are not part of either the iCAP or Serviceguard

products and are not supported by HP.

Scripts for member failover in an nPartition environment

These scripts correspond to the “Example: Automated (Serviceguard) member failover from a partial

outage with nPartitions” sub-section. This example is based on a Serviceguard cluster that includes the

partitions db1 and db2. The failover package is defined to run normally on db1 and with AUTO_RUN

enabled so that failover will occur to db2. There is no standby Group Manager, so no Group

Manager failover. For this example, assume the package workload needs a minimum of six processor

cores to be activated. The gicap_start.sh start-up script is invoked from

customer_defined_run_cmds of the package control script. It will run on the partition where

Serviceguard starts the package (either a normal start-up or a failover start-up) or when the

cmrunpkg command is invoked.

#cat gicap_start.sh

#! /usr/bin/sh

PATH=$PATH:/usr/sbin

HOSTNAME=`/usr/bin/hostname`

case $HOSTNAME in

db1) OTHER_HOST=db2;;

db2) OTHER_HOST=db1;;

esac

# check the OS status of the other node in the cluster

STATUS=`cmviewcl -f line -n $OTHER_HOST | grep ^status= | cut -f 2 -d=`

If [[ $STATUS != “up” ]]

then