Designing Disaster Recovery Clusters using Metroclusters and Continentalclusters, Reprinted October 2011 (5900-1881)

For additional examples of setting up and running DR rehearsal in different environments, see the
Disaster Recovery Rehearsal in Continentalclusters whitepaper available at: http://www.hp.com/
go/hpux-serviceguard-docs.
Data Replication Rehearsal in a Sample Environment
This section describes how to set up and run data replication (DR) rehearsal with the example of
a single instance Oracle application with Continentalcluster with Continuous Access P9000 or XP
integration.
For additional examples of setting up and running DR rehearsal in different environments, see
Disaster Recovery Rehearsal in Continentalclusters white paper at http://www.hp.com/go/
hpux-serviceguard-docs.
This section discusses the following topics:
Setup Environment
Rehearsing Failure for a Single Instance Application
Setup Environment
This section describes the setup required to rehearse failure of the recovery group billing_recgp
which is configured with Cluster1/Cluster2 as the recovery pair where Cluster1 is the primary
cluster and Cluster2 is the recovery cluster. It is assumed that Continentalclusters is configured with
the recovery groups and that the primary packages are up at the primary cluster.
To set up Continentalclusters for a DR rehearsal process, you need to make the following changes
in your environment:
Set up the continentalcluster state directory on the monitor package non-replicated shared
disk.
Configure the Continentalclusters monitor package script to mount FS /opt/cmconcl/
statedir on the non-replicated shared disk.
You need to make these changes only once to prepare your environment for DR rehearsal. For this
example, it is assumed that you have already made these configuration changes in your environment.
Device Group Configuration Changes
To prevent failure of primary packages during rehearsal (where replication is suspended), you
must disable the domino mode for device groups in an SRDF environment. However, in a Continuous
Access P9000 or XP environment, you need not disable the Fence=data setting.
Rehearsal Package Configuration
To rehearse failure for the recovery group billing_recgp, you must configure a non-metrocluster
type package. This package must be different from the recovery package with a separate package
directory with its own package configuration and control file. Configure the rehearsal package
with the volume group and the file system mount point that you have configured for the recovery
package. In addition, to prevent split brain scenarios, it is recommended that you configure the
rehearsal package with an IP address that is different from the one that is configured for the recovery
package. Since Serviceguard does not allow two different packages to be configured with the
same SERVICE_NAME, the SERVICE_NAME parameter and the SERVICE_CMD parameter for the
rehearsal package has to be different.
It is recommended that you prepare the rehearsal package configuration by cloning the recovery
package configuration and then changing the relevant configuration required for rehearsal. Also,
the Metrocluster environment file must not be present in the rehearsal package directory on any
node.
Distribute the rehearsal control scripts on all the nodes of the recovery cluster and apply the rehearsal
package configuration on the recovery cluster.
104 Designing Continentalclusters