This parameter sets a target time for
members to
bring a newly added secondary
cluster caching facility into
peer state with the primary
cluster caching facility.
Once it is in peer state, the secondary
cluster caching facility is
able to take over the role of the primary
cluster caching facility if
the primary
cluster caching facility fails
or is shut down for maintenance. A lower setting for
cf_catchup_trgt causes
members to
aggressively perform catch up activity, thus minimizing the window
during which a
DB2® pureScale® instance
is running without a secondary
cluster caching facility that
is ready to take over the role of the primary. A lower setting has
a larger impact on database transactions and the overall performance
of the system since more system resources (for example, I/O bandwidth)
are used for catch up activity.
Note: Although
the default setting for
cf_catchup_trgt balances
system performance with high availability, the following rule of thumb
can be used to tune this parameter:
- If availability is more important than performance, use a lower
setting for cf_catchup_trgt.
- If performance is more important than availability, use a higher
setting for cf_catchup_trgt.