Db2 resource availability considerations
Several options for the catalog and directory, for data in group buffer pools, and for Db2 restart can be used for data that requires high-availability.
Db2 resource availability considerations for a data sharing group are basically the same as for a single Db2 subsystem.
Critical Db2 data
Critical Db2 data should be placed behind high-availability storage controllers, such as the IBM® Enterprise Storage Server®. Consider placing the catalog and directory behind a 3990 control unit with dual-write capability for hardware duplexing. Another possibility is to use storage controllers with high-availability characteristics (such as controllers that use RAID technology) for critical Db2 data.
Group buffer pool data
Assign data that require high availability to group buffer pools that reside on non-volatile coupling facilities. Use group buffer pool duplexing, which has minimal impact on performance and can avoid hours of recovery time.
Data availability at restart
For faster restarts, take more frequent checkpoints. Checkpoint frequency is the most important factor in Db2 restart time.
If your installation sometimes has problems with units of recovery (URs) that take a long time to back out after a failure, make plans to reroute work to other members of the group. Another solution is to postpone backout processing for those long-running URs until Db2 is up and receiving new work.