On Thu, 26 Oct 2017 07:30:07 +0000, Baguley, Nicholas: Absa <[email protected]> wrote:
>We need to echo or trace the TSO commands processed in a batch TSO process... >We are issuing an ADDUSER command under TSO and it returns a RC=8. >In itself not a "biggie". We run TSO via an ATTACH of IKJEFTnn(1B in this >case) so it is a subtask of an IMS address space. >The ADDUSER command is passed to IKJEFT as a PARM on the attach svc/macro as >opposed to SYSTSIN. > >We don't see the command "echoed" to SYSTSPRT as you "normally" do when using >SYSTSIN. >Is anyone aware of a mechanism of switching on tracing or diagnosing PARM= >input to IKJ? > >NB - this works fine in 99% of cases. We suspect either we are not building up >the ADDUSER command correctly(syntax error) or we have a RACF issue. >Unfortunately my next opportunity to make a program change and <SPLAT> the >command to the syslog is a couple of weeks away. >Maybe the assumption within the the bowels of TSO was that if input is via >PARM then there would be a jcl deck or job output to inspect. Whether the command comes in via the PARM on the ATTACH, or via SYSTSIN, it should be echoed on SYSTSPRT as far as I remember. You might try passing your command to a clist or REXX exec that will echo it or provide other tracing info, then execute it, then provide further echoing. It might even use output trapping to guarantee it catches any output produced by the command and can echo that. For example, if you're currently doing an ATTACH with a parameter of: ADDUSER rest-of-command you might instead try: %RUNME "ADDUSER rest-of-command" where the RUNME clist or exec that you would provide would take the quoted string, remove the outer quotes, echo the command, then run it, etc. I have a few comments, though, on the idea of running the TMP and RACF commands within an IMS address space. First, I didn't think that the TMP liked being run as a subtask. I thought it expects to be the jobstep task, and does some things that depend on that. I could be wrong, though and it's not anything I can check these days, being retired and with no access to the z/OS internals or even a system to play with. Next, it seems odd to me that the user ID associated with an IMS address space would have the necessary authority to run RACF commands. That certainly can be done, but I think it's very uncommon. Of course, if your process is working 99% of the time as you describe, then I suppose that this must all work, in general :) I hope this helps you resolve your problem. -- Walt ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
