A failover policy specifies
how a cluster manager should respond when a cluster element such as
a network interface card or a database server fails. In general,
a cluster manager will transfer workload away from a failed element
to an alternative element that had been previously identified to the
cluster manager as an appropriate replacement for the element that
failed. This transfer of workload from a failed element to a secondary
element is called failover.
Round robin failover policy
When
you are using a round robin failover policy, if a failure
occurs with one computer in the cluster domain (also called cluster
domain nodes or nodes); the database manager restarts
the work from the failed cluster domain node on any other node that
is in the cluster domain. The round robin failover policy is
available for both single and multiple database partition configurations.
Mutual failover policy
To configure a mutual failover policy,
you associate a pair of computers in the cluster domain (also called cluster
domain nodes or simply nodes) as a system pair.
If there is a failure on one of the nodes in this pair, then the
database partitions on the failed node will failover to the other
node in the pair. Mutual failover is only available when you have
multiple database partitions.
N Plus M failover policy
When you are using a N Plus M failover
policy, then if there is a failure associated with one computer
in the cluster domain (also called cluster domain nodes or
simply nodes) then the database partitions on the failed
node will failover to any other node that is in the cluster domain.If
roving HA failover is enabled, the last failed node become the standby
node once that failed node is brought online again. The roving HA
failover for N plus M failover policy is only supported for the case
where M=1. N Plus M failover is only available when you have
multiple database partitions.
Local restart failover policy
When
you use the local restart failover policy and a failure
on one of the computers in the cluster domain (also called cluster
domain nodes ornodes) occurs, the database manager
restarts the database in place (or locally) on the same node that
failed. The local restart failover policy is available
for both single and multiple database partition configurations.
HADR failover policy
When you configure a HADR failover policy, you are enabling
the Db2® High
Availability Disaster Recovery (HADR) feature to manage failover. If an HADR primary database fails,
the database manager will move the workload from the failed database to an HADR standby
database.
Custom failover policy
When
you configure a custom failover policy, you create a
list of computers in the cluster domain (also called cluster
domain nodes or nodes) onto which the database
manager can failover. If a node in the cluster domain fails, the database
manager moves the workload from the failed node to one of the nodes
in the list that you specified. The custom failover policy is
available for both single and multiple database partition configurations.