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

One final note applies only to Serviceguard configurations where the primary node and the failover
node are both on the same system complex. In this situation failover works, but you should always
fail back to the primary, instead of making the failover node the new primary (do not use the
rotating standby model.) This is because you cannot seize rights that have already been seized from
the same complex.
Additional GiCAP high availability solutions
In the previous examples, we saw that GiCAP enables you to move capacity from one server partition
to another server. However, there are additional ways and configurations that enable you to use
GiCAP to provide high-availability solutions:
• You can use GiCAP to move capacity from one or more nonproduction servers, such as test servers,
during a failover situation. You have a set of standby servers that are part of a group and can pool
their resources to provide failover capability.
• You can combine GiCAP with TiCAP in a similar scenario. If you have a set of nonproduction
servers, and some of those servers are provisioned with TiCAP (perhaps for use during peak
production times), then all TiCAP from the entire group is available during a failover situation.
However, because the TiCAP is available to all group members, it may be inadvertently consumed
such that an insufficient balance is actually available in a failover situation. Take care to ensure
enough TiCAP is available if you are relying on it to provide failover.
• Systems with full usage rights can also be part of a GiCAP group and can be used as donor
systems, contributing usage rights to the group and allowing additional activations on member
systems with iCAP components.
• The examples show failover from nPartitions to nPartitions and from virtual partitions to virtual
partitions, but this is not a requirement. Rights that are seized from either an nPartition environment
or a virtual partition environment can be deployed on any member of the GiCAP group, regardless
of the target environment type.
Summary
You can use HP iCAP in combination with HP Serviceguard to reduce the cost of implementing highly
available solutions for mission-critical environments. The rich functionality of both products provides a
wide spectrum of possible solutions:
• For many Serviceguard configurations, the purchase and installation of iCAP processors provides—
at a much reduced price—standby processing capacity for certain specific hardware failures. It also
provides the foundation for more sophisticated high-availability solutions.
• When dynamically resizable partition configurations are used, the purchase and installation of
iCAP components (cores and cells with memory) enables standby partitions to be provisioned cost
effectively and to share resources under some situations. Redundancy of components can be
provided without doubling the price.
• Adding TiCAP to a system with iCAP components provides additional flexibility and standby
capacity. This part of the solution is not limited to a particular configuration but provides extra
standby capacity only for processor cores. It is used in situations in which you can easily predict the
duration of the failure because it requires incremental purchases of TiCAP. The requirement to do
failback is compelling for the same reason.
• GiCAP extends these capabilities by enabling multiple servers of a group to share resources
(usage rights and TiCAP). Sharing these resources can be beneficial in either load balancing or
failover solutions.
• GiCAP in combination with TiCAP provides the most cost-effective and flexible solution and can
potentially be used in all HA configurations.
37