In a 2008 thread on people using IEFBR14 to do deletes there as this: "Indeed, how should Allocation know whether the program about to execute wants to "do something" with the dataset(s) before deleting it/them? Perhaps Allocation could be "educated" to issue HDELETE iff the dataset is migrated *AND* DISP=(,DELETE) *AND* PGM=IEFBR14."
I have a set of application folks who use this method for tape datasets (reusing the dataset names, rather than making them unique for whatever reason), and of course no one wants to change the batch application because of testing requirements, busy with other things, etc. I believe HSM installs a catalog locate exit, as does ABR, to handle the 'dynamic recall' . Why can't the exit check for this specific set of conditions: 1. jobstep program is IEFBR14 2. At offset 0 of the location of IEFBR14 in storage, we have '1BFF07FE' : SR 15,15 followed by BCR 15,14 3. Status is OLD or MOD 4. normal disposition AND conditional disposition are DELETE If all of the above are met, then do the equivalent of HDELETE for the dataset Seems to me that verifies in most cases that we're dealing with the real IEFBR14 and not a replacement (although I grant it's possible to leave those bytes in place by coding an IEFBR14 replacement with a different entry point, I believe it in practice to be pretty unlikely), and that the dataset is intended to be deleted no matter what (both dispositions are DELETE). The workaround for now is to try to make sure the dataset to be deleted is either retained until the delete reference (often the next day), OR to do HRECALLs ahead of the job. As I said, this is probably beating a dead horse but I didn't see this specific proposal mentioned. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html