Plans for extended recovery facility toleration

Db2 can be used in an extended recovery facility (XRF) recovery environment with CICS® or IMS.

All Db2-owned data sets (executable code, the Db2 catalog, and user databases) must be on a disk that is shared between the primary and alternate XRF processors. In the event of an XRF recovery, Db2 must be stopped on the primary processor and started on the alternate. For CICS, that can be done automatically, by using the facilities provided by CICS, or manually, by the system operator. For IMS, that is a manual operation and must be done after the coordinating IMS system has completed the processor switch. In that way, any work that includes SQL can be moved to the alternate processor with the remaining non-SQL work. Other Db2 work (interactive or batch SQL and Db2 utilities) must be completed or terminated before Db2 can be switched to the alternate processor. Consider the effect of this potential interruption carefully when planning your XRF recovery scenarios.

Plan carefully to prevent Db2 from being started on the alternate processor until the Db2 system on the active, failing processor terminates. A premature start can cause severe integrity problems in data, the catalog, and the log. The use of global resource serialization (GRS) helps avoid the integrity problems by preventing simultaneous use of Db2 on the two systems. The bootstrap data set (BSDS) must be included as a protected resource, and the primary and alternate XRF processors must be included in the GRS ring.