Recoup 'erroneously available' pool addresses are addresses that are shown as available in the Recoup captured pool directories, but were located as in-use during Recoup chain-chase. Erroneously available addresses are protected during Recoup roll-in, but to improve database integrity, customers should consider running Recoup thru Phase 7 to identify diagnostic information about these addresses.
Recoup will also detect the erroneously available condition if it encounters a file address that was released by an application to the system after the last PDU run but before the start of Recoup phase 1 chain chase (in the pre-recoup FC33 set), and detects the same file address in a record structure chased by recoup during phase 1, or released again by an application during recoup phase 1 chain chase (in during recoup FC33 set).
Recoup Phase 5 uses the addresses for lost and EA to retrieve the data records online via the zseld (selective dump) function. That is put to the dump tape (RTA or RTL) which is then used by Phases 6 and 7 to generate reports. Phase 6 presents the zseld data for each record (file address followed by the first 20 bytes of the records data). Phase 7 uses that zseld data to create counts of lost or EA by record ID and then within each ID the "last program to file" name.
Common causes for erroneously available are:
1) Application release of a file address but not removing the references to that file address from the application data base. Then recoup chain chase detects the file address still reference in the application data base.
2) Application releases the same file address multiple times. In a PDU run, these error would be detected as an online multiple release of a file address, be in recoup this same set of actions will be identified as an erroneously available addresses.
3) A new application data base was loaded to the system, however an updated recoup descriptor or DBDEF (for TPFDF structures) was not properly loaded to the system. This causes data records in the new application to be identified as "lost" addresses and possibly released back to the pool of available addresses by a PDU run for lost addresses, On the next recoup run, the properly updated DBDEF and/or Recoup Descriptor was loaded and these addresses are detected during phase 1 chain chase run,