Glossary

A  B  C  D  E  F  G  H  I  J  K  L  M  N  O  P  Q  R  S  T  U  V  W  X  Y  Z

	A

allocation

The amount of a resource, such as CPU, that gWLM sets aside for a workload's compartment after arbitrating resource requests from the policies for all the workloads.

In managed mode, gWLM makes an allocation available to a workload's compartment. In advisory mode, however, gWLM simply reports what the allocation would be--without actually affecting resource allocations on a system.

application

The collection of processes that are within a single compartment. The compartment can be an nPartition (npar), a virtual partition (vpar), a processor set (pset), or a Fair Share Scheduler (fss) group. gWLM manages a workload by adjusting the system resource allocations for its compartment.

associate

Associating, or applying, a policy to a workload instructs gWLM to manage that workload according to the policy.

	C

cabinet

The physical enclosure that contains cells and/or I/O chassis in a complex. A cabinet also includes power and cooling hardware.

clipping

The limiting of a policy's resource request.

There are several types of clipping:

• Compartment clipping

  A workload's compartment may already be at its maximum size (as set using a vpar command for example), with policy requests trying to increase it beyond its configured maximum.

• Policy clipping

  A workload is already getting the maximum CPU allocation allowed based on its policy; however, the request would be higher if the policy maximum were higher.

• Priority clipping

  This type of clipping happens when all the workloads' compartments at higher priority levels or at the same priority level are allocated their requested resources, not leaving enough resources for the compartments remaining at the same or lower priority levels.

compartment

An npar, vpar, pset, or fss group with its resource allocation being managed by gWLM.

Multiple compartments--all of the same type--are grouped to form a shared resource domain, or SRD. The compartments all share the resources within the SRD. Each compartment holds a workload and can be in only one SRD. gWLM manages each workload by adjusting the resource allocation for its compartment.

compartment consumption

The amount of a resource being consumed by the processes in a compartment.

For example, with the CPU resource, compartment consumption is the amount of CPU being used by the compartment's processes.

CompartmentMax

The maximum amount of a resource that a compartment can have. This value is the maximum resource allocation allowed by the underlying compartment. However, gWLM may reduce this number at times because an SRD has a large number of compartments and each compartment must receive a minimum portion of the resources.

See PolicyMax

CompartmentMin

The minimum amount of a resource that a compartment can have. This value is the minimum resource allocation required by the underlying compartment.

See PolicyMin

compartment utilization

The compartment consumption divided by the compartment's size.

For example, if a compartment's consumption is two CPUs and its size is four CPUs, the compartment utilization is 50%.

complex

A single hardware configuration that can support multiple instances of an operating system by means of nPartitions. A complex includes one or more cabinets that are cabled together and all of the hardware resources that they contain.

convergence rate

Each workload behaves differently when allocated CPU: Small changes in allocated CPU may have a significant effect on one workload's performance while those same small changes produce no effect at all in another workload's performance.

The convergence rate allows you to indicate how sensitive a workload is to changes in CPU allocation. The default rate is 1.0. Larger values produce larger changes in the allocation, resulting in a faster convergence on the policy's target. Similarly, smaller values slow down convergence on the target.

custom policy

A policy for managing a workload's compartment. This type of policy allows you to provide your own metric. gWLM then manages an associated workload, adjusting allocation as needed, based on how the value of its metric compares to a target you specify. (You update values for the metric using the gwlmsend command on the HP-UX or Linux instance where the workload is running.)

You can set a priority for custom policies to ensure gWLM attempts to satisfy the policies in a particular order.

	D

deploy

Enable gWLM control of an SRD.

Deploying an SRD in managed mode enables gWLM control of resource allocation within the SRD. For example, in an SRD based on a vpar that has psets for compartments, deploying an SRD in managed mode allows gWLM to actively migrate CPUs between psets.

When deploying an SRD in advisory mode, gWLM simply reports what the allocation would be--without actually affecting resource allocations on a system.

discovery

gWLM can examine systems you specify, automatically identifying which compartment types (npars, vpars, psets, and fss groups) are present on the systems. You then form SRDs based on the discovered compartment types.

	E

Effective PolicyMax

See PolicyMax

Effective PolicyMin

See PolicyMin

	F

fixed policy

A policy for managing a workload's compartment. This type of policy guarantees that a workload's compartment has a fixed (constant) amount of CPU resources.

Fixed policies do not have a priority you can set. gWLM satisfies these policies before attempting to satisfy any other type of policies.

fss group

A group of processes that has its CPU allocation managed by the Fair Share Scheduler that is available with HP-UX. A benefit of fss groups is their granularity: You can allocate fractions of CPU resources, rather than only whole CPUs, to the group of processes.

	G

Global Workload Manager (gWLM)

gWLM is a tool that allows you to centrally define resource-sharing policies that you can use across multiple HP servers. These policies increase system utilization and facilitate controlled sharing of system resources. gWLM's monitoring abilities provide both real-time and historical monitoring of the resource allocation.

gWLM

See Global Workload Manager

	H

hardware partition

See nPartition

host

A system or partition running an instance of the operating system.

hostname

The name of a system or partition running an OS instance.

	M

measured

The current value of a metric being used in a policy.

mode

gWLM offers two modes: advisory and managed. Advisory mode allows you to see what requests gWLM would make for a workload's compartment--without actually affecting resource allocation. You can use this mode when creating and fine-tuning your policies. Once you are comfortable with your policies, use managed mode to have gWLM automatically adjust the resource allocations for your defined compartments.

You can only set the mode on the SRD level. All compartments within an SRD operate in the same mode, either advisory or managed.

	N

nPartition

A partition of an HP server, consisting of a group of cells (containing CPUs and memory) and I/O chassis (containing I/O devices). Each nPartition, or npar, operates independently of other nPartitions and can run a single instance of an operating system or can be further divided into virtual partitions (on systems that support the vpars software).

nPartitions, or npars, can be used as compartments managed by gWLM as long as several requirements are met.

	O

OS

Operating system

OwnBorrow policy

A policy for managing a workload's compartment. This type of policy allows you to set the:

• Amount of CPU resources a compartment owns

• Minimum amount of CPU resources a compartment must have (after lending resources to other workloads)

• Maximum amount of CPU resources a compartment can have (after borrowing resources from other workloads)

A compartment is guaranteed its owned CPU resources when needed. When a workload is not busy, gWLM may lend CPU resources from that workload's compartment down to its minimum to help workloads that are busy. (If a compartment has lent out CPU resources and that compartment's workload becomes busy, the compartment re-acquires those lent-out CPU resources immediately.) Also, a compartment with a busy workload can borrow CPU resources up to its maximum, assuming resources are available for borrowing.

OwnBorrow policies do not have a priority you can set.

owned

A workload's compartment is guaranteed its owned CPU resources when needed. The compartment can lend its owned resources to other workloads if configured to do so.

	P

partition

A subset of the processor, memory, and other hardware resources of a server, on which a single instance of an operating system can be run. Partitioning allows a single server to run multiple instances of the operating system, each instance running in its own partition, with isolation from other partitions.

policy

A policy is a collection of settings that instruct gWLM how to manage a workload's resources. For example, a policy can indicate the amount of CPU resources a workload owns (and is guaranteed when needed) as well as how much of those resources the workload can lend to other workloads.

A single policy can be associated with multiple workloads.

PolicyMax

When discussing a given policy: The maximum amount of a resource, such as number of CPUs, for a compartment as specified in that policy's definition.

In graphs, the Effective PolicyMax is shown. This value is the smaller of PolicyMax and CompartmentMax (the maximum amount of a resource that a compartment can have).

See CompartmentMax

PolicyMin

When discussing a given policy: The minimum amount of a resource, such as number of CPUs, for a compartment as specified in that policy's definition.

In graphs, the Effective PolicyMin is shown. This value is the larger of PolicyMin and CompartmentMin (the minimum amount of a resource that a compartment can have).

See CompartmentMin

policy pass/fail

A policy can fail to meet its target. A failure can be due to clipping of the policy's resource requests.

See clipping

priority

The importance of a policy relative to other policies. The highest priority is 1. Lower priorities are 2, 3, and so on--up to and including 1000.

gWLM uses priorities to determine the order in which to allocate resources when the sum of the resource requests exceeds the resources available in the SRD.

Fixed policies do not have priorities and are allocated resources before priorities are evaluated.

If all resource requests have been met and resources are still available, the weight assigned to each policy--rather than its priority--determines how the excess resources are distributed.

processor set (pset)

A processor set is a collection of processors (CPUs) grouped together for the exclusive access by applications assigned to that processor set. Each application runs only on processors in its assigned processor set. (On Linux, gWLM simulates psets using CPU affinity masks.)

	R

request

The amount of a system resource that a policy asks gWLM to give to the policy's compartment. (Parameters you specify in defining a policy may restrict its request.) Each policy makes a request. gWLM then arbitrates the requests from all the policies to determine what it actually gives, or allocates, to the compartments. (For example, with virtual partitions, the amounts allocated may be restricted by the minimum and maximum number of CPUs you set for the compartments when using vpar commands.)

	S

shared resource domain (SRD)

A collection of compartments--all of the same type--that share system resources. The compartments can be nPartitions (npars), virtual partitions (vpars), processor sets (psets), or fair share scheduler (fss) groups. A server containing npars can be an SRD--as long as npar requirements are met. A server or an nPartition divided into vpars can be an SRD for its vpar compartments. Similarly, a server, an npar, or a vpar containing psets can be an SRD for its pset compartments. Lastly, a server, an npar, or a vpar containing fss groups can be an SRD for its fss group compartments.

A complex with nPartitions can hold multiple SRDs. For example, if the complex is divided into nPartitions, named Par1 and Par2, Par1's compartments could be vpars, while Par2's compartments are psets.

Each compartment holds a workload. gWLM manages the workload by adjusting the compartment's resource allocation.

Size

The amount of a resource a compartment actually has.

For CPU, size may differ from the allocation when gWLM is deployed in advisory mode.

SRD states

An SRD can be in one of two states: Deployed or Undeployed.

When deployed, an SRD can be in one of two modes: advisory or managed.

See mode

system

An item (such as a server, an nPartition, or a virtual partition) that is running an instance of the operating system.

	T

target

The value that drives a policy, influencing its resource requests to gWLM.

For a target CPU utilization, gWLM attempts to keep a workload's CPU utilization below the target by adding CPU resources when the workload is using too much of its current CPU allocation. For example, assume a workload has a utilization policy with a target of 80% and a size of 5 CPUs. If the workload is consuming 4.5 CPUs, its utilization percentage is 4.5/5, or 90%. gWLM would attempt to allocate additional CPU resources to the workload to meet the target. A size of 6 CPUs would result in a utilization percentage of 4.5/6, or 75%, meeting the target.

A target can also be a value the workload should stay above, such as x transactions per second. Adding resources in this case helps the workload maintain the number of transactions.

	U

undeploy

Disable gWLM's management of resources in a specified SRD.

If an SRD is in managed mode, undeploying stops the migration of system resources between compartments in the SRD. If the SRD is in advisory mode, gWLM no longer provides information on what requests would have been made.

user

A person with an account on the given system.

utilization policy

A policy for managing a workload's compartment. This type of policy has a target based on utilization. With a CPU utilization policy, gWLM attempts to keep a workload's CPU utilization below the target by adding CPU resources when the workload is using too much of its current CPU allocation. For example, assume a workload has a utilization policy with a target of 80% and an allocation of 5 CPUs. If the workload is consuming 4.5 CPUs, its utilization percentage is 4.5/5, or 90%. gWLM would attempt to allocate additional CPU resources to the workload to meet the target. An allocation of 6 CPUs would result in a utilization percentage of 4.5/6, or 75%, meeting the target.

You can set a priority for utilization policies to ensure gWLM attempts to satisfy the policies in a particular order. The highest priority is 1. Lower priorities are 2, 3, and so on--up to and including 1000.

	V

virtual partition (vpar)

A software partition of a server or of a single nPartition, where each virtual partition can run its own instance of the operating system. A virtual partition cannot span an nPartition boundary.

	W

weight

A value you assign to a policy that determines how resources are distributed in the following two scenarios:

• gWLM addresses priority levels from highest to lowest, allocating resources to all requests at a given priority level before considering lower priority requests. If, at some priority level, all requests cannot be satisfied, the remaining resources are distributed so that the total resource allocation for each workload is as near the proportion of its weight relative to the sum of all the weights as possible.

• If gWLM has satisfied all resource requests at all priorities and there are resources still to be allocated, it will distribute the remaining resources by weight. Again, this is so that the total resource allocation for each workload is as near the proportion of its weight relative to the sum of all the weights as possible.

workload

A collection of applications whose performance is to be managed as a single unit. You can base a workload on an npar, a vpar, a pset, or an fss group. gWLM then manages resource allocation for the npar, vpar, pset, or fss group.

gWLM manages a workload by adjusting the system resource allocation for its compartment.