gil, Yes, jcl allocation is ideal. That is not possible here, as this, like ftp, or any other file transfer mechanism, coming into the mainframe, via a server, running on z/OS, must use dynamic allocation for allocating the dataset.
Your question on the S99WTDSN privilege is one I don't have an answer to. The product that is doing the serving is using bpxwdyn for it's allocations, however, I don't know, off the top of my head if there is a parm for BPXWDYN that would tell it to wait. Anybody know that one? Thanks gil for your comments and pointer to that setting for dynalloc. peter On Tue, 6 May 2025 13:42:05 -0500, Paul Gilmartin <[email protected]> wrote: >On Tue, 6 May 2025 11:29:33 -0500, Peter Vander Woude wrote: > >>We have a file transfer tool, that is used to send data to z/OS, however if >>the dataset is a gdg, and there is a disp=old on the gdg, the dynamic >>allocation fails and the transfer job also fails. >> >"to z/OS" From what platform where that transfer tool runs? > >>We can have them run a script prior to the actual transfer and I want to be >>able to check if there is an enq against the gdg base. Now SYSDSN will give >>me that, however, if there are multiple migrated gds's they all get recalled, >>which I do not want to do. >> >Will that suffer a TOCTTOU exposure? > >JCL allocation is ideal in that it delays until DEQ. >Have you S99WTDSN privilege? (But beware deadlocks.) >-- >gil > >---------------------------------------------------------------------- >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
