Getting started with HP-UX Workload Manager

41
Automatically resizing virtual partitions
HP-UX virtual partitions, which are implemented in software, provide complete software isolation.
Each virtual partition runs its own instance of the HP-UX operating system. If you require such isolation
but want to optimize use of resources, you can use WLM to migrate cores among virtual partitions
based on the needs of the applications in those virtual partitions.
WLM provides example configurations to simplify migration of cores. In addition, a white paper that
addresses this topic, titled “Resizing partitions automatically with HP-UX Workload Manager,” is
available from http://www.hp.com/go/wlm/ on the “Information library” page.
Automatically resizing nPartitions using HP Instant Capacity software
HP nPartitions are implemented in hardware. As with virtual partitions, each nPartition runs its own
copy of HP-UX. If each nPartition on a system has Instant Capacity software installed, WLM can
simulate migration of cores among the nPartitions. The simulation is achieved by deactivating a core
on one nPartition and then activating a core on another nPartition where more cores are needed,
thereby giving the appearance of moving a core without incurring a charge for an additional core.
WLM provides example configurations to simplify this simulated migration of cores. In addition, a
white paper that addresses this topic, titled “Resizing partitions automatically with HP-UX Workload
Manager,” is available from http://www.hp.com/go/wlm/ on the “Information library” page.
Ensuring allocations of CPU resources after an HP Serviceguard failover
One approach for using WLM with HP Serviceguard is to create a single WLM configuration that you
use on each node in the Serviceguard cluster. This one WLM configuration creates WLM workload
groups for each Serviceguard package in the cluster. Then, whenever any package fails over to any
node in the cluster, it has a workload group waiting for it. However, that group has active SLOs only
when its package is active on that node. WLM then allocates resources on the node to groups with
active SLOs, based on the priorities you assigned.
A white paper that addresses this topic, titled “More efficient high availability and resource utilization
through manageability—Integrating HP-UX Clustering Solutions, Instant Capacity, and HP-UX
Workload Manager,” is available from http://www.hp.com/go/wlm/ on the “Information library”
page.
Isolating Oracle database instances
By creating a WLM workload group for an Oracle database instance, you can separate it from other
databases instances and applications on the system. Once separated, the instance can be allocated
CPU resources to ensure its performance meets service-level objectives.
1. Enter the maximum amount of CPU resources (cores) the system will have. The wizard estimates the
maximum value. Confirm or edit the estimate. Be sure to consider changes in the CPU/core count
caused by Instant Capacity and other such adjustments.