HP Integrity Essentials Global Workload Manager: Workload Management for HP Integrity Virtual Machines

11
example. Note that size/allocation values “follow” the consumption; in other words, gWLM maintains
an allocation and size level that meets current consumption. Thus, the current policy for workload
tara102 is performing well. For more information on interpreting gWLM reports, see the online help.
Practical Use Scenarios
This section provides several scenarios demonstrating how you might use gWLM to manage Integrity
VM virtual machines as workloads:
• Scenario 1: you have a VM Host with a set of virtual machines that you want managed by gWLM
• Scenario 2: you have a complex that includes a VM Host in one hard partition and a set of virtual
partitions in another, and you want gWLM to make Instant Capacity (iCAP) resources available to
the virtual machines as needed
• Scenario 3: you have a complex that includes a VM Host in one hard partition and a set of virtual
partitions in another, and you want gWLM to make Temporary Instant Capacity (TiCAP) resources
available to the virtual machines as needed
• Scenario 4: you have a complex that includes a VM Host with SMP virtual machines and want to
make sure the SMP virtual machines maintain their minimum number of cores even when cores
are needed elsewhere
Scenario 1: Putting gWLM in Control of Virtual Machines
Assume you have four virtual machines configured for your Integrity VM, as shown in Figure 2. Two
of the virtual machines (VM1 and VM2) run mission-critical applications. The other two virtual
machines (VM3 and VM4) run less critical applications. For normal operations, VM1 and VM2
typically require a smaller amount of cores than VM3, but they need more resources to meet peak
demands. The entitlements and vCPU assignments are described on the right (where one vCPU is
assigned for each core). The virtual machines shown in Figure 2 are running at their entitlement levels.
Figure 2 – Integrity VM Initial Configuration
The Integrity VM entitlements ensure that the virtual machines receive, at minimum, the specified
percentage of CPU resources. However, Integrity VM currently provides no mechanism to limit the
resources consumed by a given virtual machine, other than allowing you to change the configured
number of vCPUs. Thus, in this scenario, it is possible that a virtual machine with a less-critical
application, such as VM3 or VM4, could consume more resources than its entitlement, preventing the
more important applications running on VM1 and VM2 from receiving extra resources as needed.
With gWLM, you can prioritize the virtual machine workloads to ensure that the higher-priority
workloads always receive the extra resources they need prior to the lower-priority workloads. You can
also assign a limit to how much certain workloads can consume.
Figure 3 shows the same virtual machines managed by gWLM. In this scenario, the highest-priority
workload (VM1) has grown and requires more resources. VM1 must receive extra resources to meet
its needs before VM2 is allocated extra resources. Assume you also want to make sure VM1 does not