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

5
workload. At any given time, the virtual machine is automatically allocated resources based on that
policy. The virtual machine’s entitlement value as reported by hpvmstatus –r is actually the most
recent allocation set by gWLM. (The entitlement set for the virtual machine in its configuration file for
the Integrity VM software is not changed while the SRD is deployed.) gWLM prevents the virtual
machine from using more than the gWLM-set allocation. When gWLM stops managing an SRD, it sets
the entitlements of the running virtual machines to their Integrity VM enforced minimums (5%) so that
all other stopped virtual machines have enough resources to be restarted.
You can create an SRD for a set of virtual machines that are already running. To start an additional
virtual machine in a deployed SRD, you must establish a gWLM workload and policy for that virtual
machine. It will be prevented from starting until a policy is established.
While managing Integrity VM resources, gWLM prevents you from changing a virtual machine’s
running (dynamic) entitlement (using the Integrity VM Manager or the Integrity VM command line
interface (CLI)); however, you can use the Integrity VM Manager or the CLI to configure and manage
other elements. (For more information on what you can and cannot change, see “Notes Regarding
gWLM Management of Virtual Machine Workloads” on page 5.)
From a central location, gWLM continuously monitors the activity of the virtual machine workloads in
the SRD to determine how resources should be allocated. This information is consolidated and stored
on the central management server (CMS). If your VM Host is deployed in a hard partition with Instant
Capacity resources, gWLM can meet the needs of your virtual machine workloads by automatically
increasing CPU resources as demand rises or transfering CPU resources to other hard partitions where
demand is greater. gWLM can automatically activate or deactivate Temporary Instant Capacity
resources available to a VM Host based on the total resource demands of all the virtual machine
workloads. In either case, gWLM distributes processor cycles based on your policies, and it attempts
to maintain the minimum number of CPU resources to meet the demands of all the virtual machines.
Taking advantage of Instant Capacity or Temporary Instant Capacity resources, gWLM can ensure
that enough cores remain available on the VM Host to satisfy the virtual machine workload having the
greatest number of configured vCPUs (each virtual machine must have a core available for every
configured vCPU); gWLM maintains the required amount of cores even when one or more cores
associated with that virtual machine become unavailable.
Notes Regarding gWLM Management of Virtual Machine Workloads
When using gWLM to manage virtual machine workloads, note the following. For more information,
see the "Getting the Most Out of gWLM" topic in online help (access online help in SIM by selecting
Tools à VSE Management, followed by the tab Shared Resource Domain, and then the question mark
[?] in the top right corner).
• gWLM operates in one of two modes: managed mode or advisory mode. Advisory mode allows
you to see what CPU requests gWLM would make for a workload—without actually affecting
resource allocation. For virtual machines (as well as PSETs and FSS groups), gWLM can operate
only in managed mode.
• When an SRD is deployed, only those virtual machines managed as workloads in the SRD are
allowed to start. This ensures that gWLM has complete control of the VM Host’s pool of resources.
gWLM allocates resources only to those virtual machines managed in the SRD and that are started.
Any virtual machine that is not managed in the SRD is stopped and will not have a gWLM policy
associated with it; therefore, that virtual machine will not be allocated resources needed to run.
• When managing a VM Host as a workload (managing the nPartition that is serving as a VM Host),
you should set the policies in the SRD so that the VM Host always has enough cores to meet at least
the minimum required for its configuration of virtual machines (typically, this minimum is the greatest
number of vCPUs assigned to any virtual machine). For information on this VM Host minimum, see
the HP Integrity Virtual Machines documentation, available from:
http://docs.hp.com/en/vse.html