32760 is optimal for load libraries; I don't know of any reason to use it for PDSE.
I would expect FORCE SYSTEM DETERMINED BLOCKSIZE to break things, and not just for RECEIVE. Consider old programs with hard wired buffer sizes and other programs creating output for them. If SDB increases the block sizes for the creating programs, the reading programs will fail on S001. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 ________________________________________ From: IBM Mainframe Discussion List [[email protected]] on behalf of Wendell Lovewell [[email protected]] Sent: Wednesday, March 3, 2021 10:35 AM To: [email protected] Subject: Re: TSO RECEIVE and System Determined Blksize - New RFE Shmuel, I believe that's the way it works for existing files that have been opened and closed, whether SDB=Y or N. SDB comes into play when a file is allocated but not written to by the allocating program. This is from the ISMF help for the SDB setting: Use the FORCE SYSTEM DETERMINED BLOCKSIZE (SDB) field to specify that the system ignore a user-specified block size if no program opens the new data set for writing while it still is allocated. This prevents the user from overriding a system-determined block size. Possible values: Y If no program opens the data set for writing while the new data set still is allocated, then the system will discard a BLKSIZE value coded by the user. The system will attempt to calculate an optimal block size. If a program opens the data set for output while it still is allocated, then the user-specified BLKSIZE will take effect and override a system-determined block size. N If the user specifies a BLKSIZE value, it will take effect and override a system-determined block size. This is the normal way for the system to run. Since IEBCOPY is adjusting the blksize, it must be that RECEIVE did the allocation first and deallocated the empty file. Then SDB says "I can do better" & changed the blksize. Then when IEBCOPY is called it says "that's not right, I'll fix it". Makes me wonder if anyone uses SDB. Plus, how is 32720 an "optimal" block size? Optimal for disk sales? :) ---------------------------------------------------------------------- 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
