Charles, >What I don't understand is the exact mapping of SUBSYS names to record-level >functionality.
If *I* understand you right, then you're confusing SUBSYS as indicated in SMFPRM and a subsystem defined to the SSI. The SMF SUBSYS types *I* am aware of are JESx, STC, TSO, OMVS, ASCH. That's it. The type assigned to the address space the SMF record is either written for or that writes it determines under which subsys it will appear. I was going to say that you can see the subsystem it is classified to in the TYPE column of SDSF (alternate panel for DA), but that's not true here. SUBSYS OMVS shows up there as STC. (Any idle BPXOINIT is TYPE(OMVS) and shows up in SDSF as TYPE(STC).) TYPE30 can be written for any SUBSYS type. As can TYPE80. 100, 101, 102, 119: Same thing. If an application that is TYPE OMVS (let's say one of my much-hated USS applications, say, WBI) has calls into DB2, then they would get written for TYPE OMVS if that address space was started using the USS interfaces for creating an address space. When in doubt, have your customer have one SYS statement that only specifies the types that are to be written. Have your customer remove all TYPE statements from any SUBSYS statements they have. In our installation, all TYPEs are specified like that and the SUBSYS statements look like this: SUBSYS(STC,INTERVAL(002000),NODETAIL) SUBSYS(TSO,DETAIL) I think I specified STC nodetail because we do collect TYPE30 and did not need the DD stuff that gets written whenever an HSM or DB2 terminates (that quickly fills up all SMF buffers). I have no clue why TSO has DETAIL (hysterically grown). The SUBSYS type an address space is classified to are seen in IEFUSI (and in all SMF exits). I use them to set a different memlimit in iefusi for the different types of asids. From the memlimit value (as shown in SDSF) I can see what TYPE an asid is classified to. The exception is the ever-growing list of address spaces that consistently override what the installation has specified with ridiculously high values that are NOT supported by any paging subsystem. And I doubt they all have code in place that prevents a wait03C. HTH, Barbara ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
