What I am seeing is all sorts of non-optimum block sizes, most of them are 3120. We are running a script from ISV which allocates the output dataset using the TSO ALLOCATE command (we have hard coded the optimum block size in the ALLOCATE command). Then the script issues the RECEIVE command. After the script runs, we see the output datasets with non-optimal block sizes. It appears that what we coded in the ALLOCATE is over-ridden by RECEIVE. By the way, the CBT ZAP is for the TRANSMIT command.
To answer your question, I would say RECEIVE is recreating the output datasets as they existed on the ISV's system, Which really astonishes me. This is a large software vendor who markets among other things performance monitoring software. These are very large output datasets. DASD space is wasted and additional I/O (remember it takes CPU to drive I/O) is required to stage and process these datasets. "Confidentially doc, I am the wabbit." Bugs Bunny Sent with Proton Mail secure email. On Sunday, December 7th, 2025 at 11:30 PM, James Mulder <[email protected]> wrote: > I would have guessed that when RECEIVE creates a target data set, it would > use the blocksize of the TRANSMITted data set. Are you seeing some other > behavior? > I would also guess that if you RECEIVE into a data set that you have already > allocated, which has a block size in the DSCB, RECEIVE shouldn't change it. > > I do know that TRANSMIT and RECEIVE hardcode BLKSIZE3120 in their DCBs for > the log data set. That really irked me several decades ago, and I showed the > owners at the time how to fix that, but they refused to do it. Of course, > they are all long gone from IBM now so I suppose I could just do it myself if > I ever get around to it. > I did change TRANSMIT and RECEIVE from 24-bit to 31-bit addressing mode in > z/OS 3.2 so that they could access ASCBs in 31-bit storage. > > Jim Mulder > > > No, that is not the issue I have. We are receiving large XMIT datasets from > > an ISV > > which require large amounts of DASD, greater than a 3390-9. If the XMIT > > datasets > > used SDB, the space requirements would be greatly reduced (I reblocked them > > after > > they were staged). > > > > Jim Mulder z/OS Diagnosis, Design, Development, Test IBM Corp. Poughkeepsie NY > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
