Cost-Effective High-Availability Solutions with HP Instant Capacity on HP-UX

Using GiCAP to recover from failures (seizing core usage rights)

With GiCAP, you have another tool for providing cost-effective high-availability solutions: The ability

to “seize” usage rights from a partition that is down. Depending on the version of Instant Capacity,

this feature can also be used to provide disaster recovery when all partitions of a GiCAP member fail.

In the simplest high-availability solution, you create a group of servers that includes two types of

members: 1) servers to run your primary processing tasks and 2) servers that provide failover

processing (Active/Standby model). Both types of servers can contain iCAP components for the most

economical solution. When a failure occurs on a partition in one of the active servers, you can use a

GiCAP command to extract (seize) processor core usage rights from that partition and transfer those

usage rights to the Group Manager. Then, using normal activation commands on the failover server,

you can use those transferred usage rights to activate additional processor cores on the failover

servers to increase capacity. With iCAP 8.01, at least one additional partition on the original primary

server must be running in order to seize usage rights. With iCAP 8.02.01 and later, no partitions

need to be running.

The freedom that iCAP 8.02.01 and later versions provide you (no partitions need to be running to

seize rights from a member), enables you to recover from more serious outages. For example,

consider a GiCAP group consisting of an active server in one location and a failover server in

another location. With the Group Manager in a third location, if either of the group members

becomes unavailable due to a disaster, you can seize core usage rights from the unavailable member

and make them available to the remaining member.

The seizure of core usage rights from a fully unavailable member is temporary. After 10 days, the

usage rights are automatically restored to the member from which they were seized. (In the event a

system cannot reconnect to the Group Manager within the 10-day period, you can extend the

expiration of the seized usage rights. This extension is applied to the Group Manager by a codeword

obtained from HP support.) Seizure of core usage rights from a member with at least one accessible

partition (at the time of rights seizure) do not expire. Later examples show how this works.

Because rights seizure always operates at the nPartition level, there are additional considerations for

virtual machine and virtual partition environments. Keep the following points in mind when working

with virtual machines:

• If there are multiple VM guests for a hard partition, they will all be affected by a reduction of usage

rights for that partition.

• Rights seizure can only be specified for the VM Host and not for the VM guests.

• For best performance, you should maintain certain levels of CPU resources for an active VM Host.

Keep the number of cores on an active VM host above:

– The total entitlement for the virtual machines running on the host

– The maximum number of virtual CPUs configured for any virtual machine on the host

For an active VM Host, avoid deactivating cores below these levels. If the VM Host is down, you can

seize rights from it; however, this may reduce the CPU resources below the recommended levels. If so,

you will have to restore usage rights to the host before starting its virtual machines.

For information about virtual partition environments, see the “

Recovery from a failure involving virtual

partitions” sub-section.