Re: Setting library list using Initial command

On 19-Nov-2014 04:11 -0600, Frank Broome wrote:

I know that the initial command is only running when connection is
established.
As I said, running a command like e.g. CHGLIBL works fine, running a
CALL to a program that changes the library list does NOT work.

After running the CALL to the program the job that has executed it
ends, whilst if I run command CHGLIBL the job stays and the lib list
is changed accordingly.

Running with iASP? What release of the OS?

I have now been able to catch the joblog and see what happens and
when I run the call to the program the job ends with:
"Tried to refer to an object that no longer exists"
From program . . . . . . . . . : QZRCSRVS
From library . . . . . . . . : QSYS
From module . . . . . . . . : QZRCSRVS
From procedure . . . . . . . : ProcessExitPt
From statement . . . . . . . : 14
To program . . . . . . . . . . : QZRCSRVS
To library . . . . . . . . . : QSYS
To module . . . . . . . . . : QZRCSRVS
To procedure . . . . . . . . : ProcessExitPt
To statement . . . . . . . . : 14

restated in symptom kwd form:

msgMCH3402 T/QZRCSRVS TM/QZRCSRVS TP/ProcessExitPt stmt/14

I have made sure that I am NOT doing any RCLACTGRP or RCLRSC in my
program, so I am not sure which object it's trying to refer to??

Any clues would be appreciated!

Only having the compiler listing [and source-code] from that IBM code, would allow someone to /know/ [though possibly only guess] what was the object that was destroyed; an object deleted since the pointer was obtained originally for that object, possibly by that same IBM routine. However from the naming of the procedure, *one might infer* that there is likely an Exit Point defined for the service; and that the attempt by the server to invoke the program defined for that exit point presumably is failing due to a stored pointer, possibly to the exit-point program, that refers to an object that was deleted since.

The Work With Registration Info (WRKREGINF) should record the Exit Point (EXITPNT) named QIBM_QZRC_RMT and there may be an Exit Program might that had been established for that Remote-Command/Program-Call service. Yet if that named program was since deleted, perhaps as a copy from the QRPLOBJ library, then the expectation would be that both invocations [a Requested Function of either X'1002' (remote-command) and a X'1003' (distributed-program-call)] should fail identically.

I suppose as an error manifest in an IBM program, possibly an object was deleted from a QPTFOBJ# library; an object that was required and activated only for the program-call method, but not via the command method.?

I would investigate if there is a program defined for the exit point, and if there is then also determine the creation date of that program; review if any bound objects have been deleted since. I would probably also obtain a trace of the failing request, to see more information than was [so limited apparently, as] provided by the joblog.

In any case, the server [or the server job serving that connection] may need to be stopped\restarted to activate either changed IBM code or a new exit program. In the case of PTF application, the Special Instructions for PTFs provided by IBM, if the PTF is applied *IMMED instead of delayed, almost surely [still] include the stated requirement to end and then restart the remote command server features.