Backing up and restoring HP Insight Management 7.3 Central Management Server (Windows)

Technical white paper | HP Insight Management 7.3

Appendix I: HA CMS cluster configurations

This appendix provides information for making full and partial backups of your Insight Management CMS. When the CMS is in

an HA cluster configuration, there are some differences that you must consider.

When stopping and starting Insight Management services as part of the backup or restore process, it is important to use

mxsync-svcutil or the Failover Cluster Manager for most of these operations. Most Insight Management services are

configured as cluster resources—they must be stopped and started (or taken offline and put back online) through the

cluster manager. If you stop an Insight Management service using a non-cluster-aware mechanism (such as net stop), a

failover of the Insight Management cluster service might occur.

The following Insight Management services are not configured as cluster resources, and run locally on both servers in the

cluster:

• HP System Management Homepage

• HP Version Control Agent

• Pegasus WMI Mapper Indications Service

The Pegasus WMI Mapper Indications Service is stopped only for restore. It does not need to be stopped for backup.

Notes:

• For HA CMS cluster configurations, you must manually copy the mxsync-svcutil.bat script to the boot disk

of each cluster node.

The file layout for HA CMS cluster configurations is slightly different than what is described in this document. In an

HA CMS cluster configuration, most of the Insight Management program and data files are located on a shared

volume (for example, the V: drive). This layout enables either server in the cluster to host the Insight Management

cluster service. Similarly, the databases reside on a separate shared volume, or are hosted external to the cluster.

However, some files reside on the boot volumes of the two cluster servers, referred to as server-specific files.

These files are listed in Table 19.

For HA CMS cluster configurations, the full backup must include all shared disks, such as the witness disk and the

boot disks of the two cluster nodes. HP recommends that you follow the procedures from Microsoft for backing

up and restoring Microsoft Failover Clusters. The backup should be made using a cluster-aware backup tool in

case the cluster configuration has to be restored from backup.

• For HA CMS cluster configurations, HP recommends that you follow the procedures from Microsoft for doing a full

backup of the cluster nodes and cluster configuration. For details, see the Microsoft documentation at:

http://technet.microsoft.com/en-us/library/cc771973.aspx

• For partial backups, it is

not

sufficient to back up only the files and directories from the server that is currently

hosting the Insight Management cluster service and the Insight Management databases. A set of server-specific

files are duplicated. Each server's boot disks contain a private (not shared) copy; both copies must be backed up

and restored. This set of server-specific files is listed in Table 19.

In the event of a restore, it is important to keep the four datasets in sync:

• Insight Management data directories and files

• Insight Management databases

• Local Insight Management data directories

• Files on the two servers in the cluster