Affinity for Capacity-Aware Placement Advice overview

Note

The Capacity-Aware Placement Advice (CAPA) Affinity settings require BMC Cloud Lifecycle Management 3.1 Service Pack 1. Affinity settings are ignored by CAPA in earlier versions of BMC Cloud Lifecycle Management.

This topic contains the following sections:

Background on Datastore placement

When a ComputeContainer (VM) is being placed, BMC Cloud Lifecycle Management polls the CAPA advisor to identify the VirtualDatastore (VMWare datastore) for placing each virtual disk. The general behavior is described in the Datastore placement section under Understanding CAPA.

There are two types of vDisks for a ComputeContainer:

  • Primary disks: These are in the provisioning template in BBSA. One VDS Placement call is made for all the primary disks.
  • Secondary disks: These are listed in the blueprint or in options. One VDS Placement call is made for each secondary disk.

There are two parameters in the Defaults and Thresholds tab of the Placement Advice Preparation Task, for the assumed size when the virtual disk size is not specified:

  • Candidate default disk size: This size is used to compute the placement of the virtual disk.
  • Disk PPA multiplier (applied to disk size for simulation): When the disk size is not known, the assumed "default disk size" is multiplied by this number to simulate the effect of the provisioned disk.

For the default disk size, CAPA uses a size (1 GB) that is small enough to reduce the chances of false failures, when a datastore has enough space. In effect, it means to simply select the datastore with the most free space.

For the Disk PPA multiplier, CAPA makes the assumed PPA just big enough to avoid a large number of repeated placements on a single datastore before the next utilization data feed. For instance, to make the PPA 10GB, the Disk PPA multiplier is 10.

Note

In BMC Cloud Lifecycle Management versions through 3.1, the primary disks are not visible to BMC Cloud Lifecycle Management. Hence, the VDS Placement call does not specify the size of the primary disks. TrueSight Capacity Optimization will assume a size for this placement according to the parameters above. In this case, modify the Candidate default disk size to the size you know as the total size of the VGP template, and use a Disk PPA multiplier of 1.

In BMC Cloud Lifecycle Management version 3.1 SP1, a new blueprint parameter was introduced called BMCBootDiskSize. This parameter can be used to set the size of the primary disks. For more information, see description of the enhancements in CLM 3.1 SP1.

Types of Affinities

There are two kinds of Affinities available in TrueSight Capacity Optimization Placement Advice:

  • Affinity for compute placement: Applies to compute resource placement (Placement advice calls for compute pools, virtual clusters, virtual hosts, virtual resource pools).
  • Affinity for datastore placement: Applies to VirtualDatastorePlacement calls.

Managing affinity settings for Datastore Placement

The default behavior of CAPA is described in the Datastore Placement section in Understanding CAPA. This behavior is called "Prefer SOI", and usually there is no need to modify it.

You can modify the settings and behavior of Affinity for datastore placement in the Placement Advice Preparation Task, as follows:

Affinity

Description

Prefer VM

Try to use same datastore for all vDisks of a VM if possible.

Strict VM

Use one datastore for all vDisks of a VM, or fail provisioning.

Prefer SOI (Default)

Try to use same datastore for all vDisks of all VMs of the SOI if possible.

Strict SOIUse one datastore for all vDisks of all VMs of an SOI, or fail provisioning
No affinityUse whichever datastore has most available space for each vdisk

If one datastore has a lot more space available than the others, then CAPA might still end up placing all the disks on that datastore. The above affinity settings produce a difference only when the top two datastores are not too far apart in terms of available space, and their difference in available space is less than the total sizes of the disks being placed.

CAPA takes the following actions for the scenarios given below:

  • Prefer SOI, Prefer VM, and No affinity affinity settings: If one datastore is exhausted, CAPA will try to use another datastore for the remaining disks.
  • Strict SOI and Strict VM affinity settings: If all the disks of a single SOI or single VM cannot fit on the datastore with the most space, CAPA will prompt BMC Cloud Lifecycle Management to fail the provisioning.

For more information, see Configuring placement options and Affinity settings for Compute and Storage resources.

Managing Affinity settings for Compute Placement

You can modify the Affinity settings for compute resource placement in the Placement Advice Preparation Task, as follows:

Affinity

Description

Prefer SOI (Default)

Try to use same compute resource for all VMs of the SOI if possible.

Strict SOI

Use one compute resource for all VMs of an SOI, or fail provisioning.

No affinity

Use the compute resource that has maximum available resources for each VM.

For more information, see Configuring placement options and Affinity settings for Compute and Storage resources.

Note

BMC Cloud Lifecycle Management can issue bulk provisioning requests where multiple SOIs are provisioned in a single transaction. In BMC Cloud Lifecycle Management versions prior to 3.1 Service Pack 1, the calls to CAPA failed to distinguish between different SOIs within one bulk provisioning request. Hence CAPA tried to place all virtual disks in the bulk transaction on the same datastore, and all virtual machines on the same compute resource. This issue has been fixed in BMC Cloud Lifecycle Management version 3.1 Service Pack 1.

Where to go from here

Configuring placement options and Affinity settings for Compute and Storage resources

Was this page helpful? Yes No Submitting... Thank you

Comments