If used, protection log and command log datasets for use in Adabas cluster environments are specified in the session start-up JCL. Session initialization writes the log dataset names to the parallel participant table (PPT) and ADARES uses dynamic allocation to access them.
The same is true of the Work dataset: it is specified in the session start-up JCL. Session initialization writes the Work dataset name to the PPT and, in the event of an autorestart, a nucleus is able to access the Work dataset of a peer nucleus using dynamic allocation.
Dynamic allocation is currently available only in cluster environments (when the NUCID is greater than zero). However, dynamic allocation is used by a noncluster nucleus to determine that the previous nucleus session was a cluster session and PLOGs or CLOGs remain to be copied.
Files that are eligible for dynamic allocation by another database must be cataloged, rather than being accessed using explicit UNIT and VOLSER keywords in the JCL. If JOBCAT or STEPCAT DD statements are used to define a catalog search structure, these statements must be used in a consistent way on all database start-up jobs in the sysplex.
