HP Integrity Virtual Machines 4.2: Installation, Configuration, and Administration
• In the case of LAN failure, Serviceguard switches to a standby LAN or moves affected
packages to a standby node.
• In the case of SPU failure, your application is transferred from a failed SPU to a functioning
SPU automatically and in a minimal amount of time.
• For failure of other monitored resources, such as disk interfaces, a package can be moved
to another node.
• For software failures, an application can be restarted on the same node or another node with
minimum disruption.
Serviceguard also gives you the advantage of easily transferring control of your application to
another SPU to bring the original SPU down for system administration, maintenance, or version
upgrades.
11.3.2 VM Hosts as Serviceguard Nodes
When planning the VM Host and Serviceguard nodes, you need to consider the following:
• Proper memory and physical CPU allocation
The number of physical CPUs per Serviceguard node limits the number of virtual CPUs
(vCPUs) with which a guest can run. If your Serviceguard nodes differ in the number of
CPUs on each Serviceguard node, then it might not be possible to failover a VM package
running on a Serviceguard node with a large number of physical CPUs (pCPUs) to a
Serviceguard node with a smaller number of physical CPUs.
A set of Integrity VM servers can be grouped into an Integrity VM multiserver environment,
MSE. This grouping is required when a set of Integrity VM Hosts is also configured as a
Serviceguard cluster. The MSE environment must be configured consistently on each node
of the VM Host cluster. To set up an Integrity VM multiserver environment, use the
hpvmdevmgmt command on each of the servers in the group to establish each server entry
and the MSE group entry. For more details about MSE, see the Integrity VM manpages
section in this document or on your Integrity VM system.
• Online migrations
To be compatible with online migration, VM configured as Serviceguard packages must
conform with the restriction and limitations outlined in Section 10.3.4 (page 181).
• Virtual machine entitlement
Virtual machine entitlement is the minimum amount of processing power guaranteed to
the virtual machine from each vCPU. When allocating minimum entitlement to a virtual
machine, it is important to plan your failover strategy, so that a failover of a set of virtual
machine packages to a single Serviceguard node does not exceed the total available CPU
processing on that node.
• CPU processing overhead
Take into consideration the CPU processing overhead that both Serviceguard and the VM
Host require.
• The amount of physical memory on each Serviceguard node
The amount of physical memory on each Serviceguard node must also be considered when
planning your VMs as Serviceguard Packages configurations and failover strategy.
• Differing amounts of physical memory of each Serviceguard node
Consideration should be given to the impact and effects of configuring dynamic memory
and the minor memory overhead associated with Serviceguard running on the VM Host. If
your Serviceguard nodes differ in the amount of physical memory on each Serviceguard
node, it might not be possible to failover a virtual machine package running on a Serviceguard
node with a large amount of physical memory to a Serviceguard node with less available
memory.
11.3 VMs as Serviceguard Packages 195