On Thu, 24 May 2018 12:55:01 -0500, Tom Marchant wrote: > >>OK, y'all probably know that I'm on a very back level system -- z/OS 1.12 >>and we're even back level on maintenance. I'm trying to get more up to >>date, mainly for "fun & profit". Anyway. There are a number of RACF modules >>which were hit and SMP/E tries to relink them. The problem seems to be that >>some CSECTs are missing in the relink. > >Was this something that happened during the APPLY of a PTF? >Did the PTF include JCLIN? >What MCS statements are in the PTF? > There's also the effect of the ++MOD CSECT(...) option to consider. If CSECT() is absent, the default is the name of the MOD element. ISPF SMP/E panels might show this.
Can you create a disposable instance of your CSI (Shmuel insists that includes PDSes)? In that do a ++DELETE of your load module and re-install the JCLIN to create it and note what happens. >AFAIK, normal behavior of SMP/E when applying a PTF that contains a >replacement ++MOD is that all LMODs that contain that module are relinked. >The original load module (or program object) is included, and the module >from the PTF is included. In this way, all of the CSECTs in the load module >are retained, and the one supplied by the PTF replaces the original. > >>They are in the LMOD in the running >>LINKLIB. In the output, I see the binder output which tries to relink the >>module. I see the missing CSECTs. The CSECTs which are not missing are have >>Binder " INCLUDE AOSBN(csect)" in the binder's input stream. The CSECTs >>which are needed are in AOSBN properly, but there is no INCLUDE for them. >> Are there any REPLACE commands in the listed SYSLIN? These are a consequence of the ++LMOD CSECT() option. >>I've never had anything like this happen before. I've been looking around >>in the current SMP/E manual and I _might_ be able to "fix" all the LMOD >>entries using UCLIN. But this seems to be dangerous to me. > >I wouldn't do that. > >>Any suggestions other than a stiff shot of something "nice" and "soothing"? > >PMR. > Against 1.12? John's last idea is more promising. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
