We are managing VLDB ADABAS database serving the population of the country. Our national Database contains more than 150 ADABAS Files.
Many a times, we are tasked with producing transactional investigations using multiple large Protection Logs against multiple transactional files. And our Protection Logs are 30,000 cylinders generated on VTS.
Using ADASEL, we have been able to generate EXPA records for multiple files, but the EXPA outputs cannot exceed 20 cards per job, as we receive SEL034 error message: "
Syntax - DDEXPA data set number specified - must be 1 - 20 inclusive
"
We have been receiving cases having around 60-80 transactional files, and we have to run separate jobs to perform the needed investigations. We have separate selection criteria for multiple files, wherein each file selection generates its specific EXPA output.
At times, we have to search a range of historical PLOG tapes, which may not be in the VTS Cache. This adds to delays in providing transactional investigation reports to the management, as the same sets of tapes have to be processed for different suites of files in separate jobs.
Hence, we request you to provide a solution to ADASEL by increasing the numbers of EXPAs cards, just as the ASSO & DATA DD Cards (i.e. maximum of 99), which is as per the documentation: "
A maximum of 99 physical extents is now set for Associator and Data Storage data sets
"
Note:
We tested to find the maximum time to process ADASEL for each of the 20 files against a TAPE, and then compared it with the maximum time a single job processing ADASEL for all the 20 files generating separate EXPAs.
We noticed that the single job processes faster per tape, despite having 20 separate file selection criteria.
Which means, based on the existing ADASEL algorithm, even if more than 20 EXPAs are generated, the time to complete processing the tape(s) should remain more-or-less the same.
Use Case | PLOG, ADASEL, EXPA, JCL |
Hi Juergen,
I'm following up with this brainstorm idea.
Regards,
Zaffer
Hello Zaffer,
we received your idea.
However, as we have much more incoming ideas than we can resolve, we still need time to
analyze
decide on whether we will implement it
The roadmaps for upcoming versions will be finished by Sep/Oct.
Cheers,
Jürgen