Using HP Global Workload Manager with Serviceguard

6
running in vPar 6 until needed by a package failover. The system used to host the secondary servers
can be sized to handle only a small number of concurrent failovers at full performance or can provide
complete redundancy with the primary servers. It can also use Temporary Instant Capacity to provide
full redundancy—while avoiding the full cost until the resources are actually needed.
Figure 3. Failover to server using vPars
Configuring HP Serviceguard
Integrating gWLM with Serviceguard requires no additional Serviceguard configuration and can even
simplify your failover scripts. With gWLM managing resource allocation among the packages, you
can be assured that your high-priority packages will receive the resources they need regardless of
other lower-priority work that may be present on the secondary servers. For example, if your failover
scripts were customized to shut down low-priority packages when high-priority packages fail over,
they would no longer have to do this.
Configuring HP Global Workload Manager
Depending upon your Serviceguard cluster configuration, there is significant flexibility on how you
can configure gWLM to meet your resource allocation goals. This section assumes that you will be
deploying gWLM in the context of an already defined Serviceguard cluster. It assumes that you are
already familiar with gWLM concepts as described elsewhere in this paper. The following sections
provide guidance on defining workloads and grouping them into shared resource domains enabling
you to capitalize on both the HA of Serviceguard and the resource management of gWLM.
g