Grasping at straws, Re: orphaned Telnet jobs -- MIDRANGE-L

"Lance Gillespie" <lgillespie@xxxxxxxx> wrote:

Maybe try running a WRKUSRJOB USER(*ALL)STATUS(*ACTIVE) JOBTYPE(*INTERACT)
and see if the unconnected TN5250e devices
that reject an incoming connection are
in DSC (disconnected) status?

No, nothing's in DSC status. And I'm sure that would haveshown up before, in a WRKACTJOB, anyway.


What I have been able to gather is this:

If the orphaned session is at SIGNON DISPLAY, it takes8-11 seconds to vary off. (Normal is about 2 seconds.)

If the orphaned session is ACTIVE, it takes 20-30 secondsto ENDJOB OPTION(*IMMED) (same as you'd expect w/ 3-pagejoblog)

An orphaned job call stack takes around 30 seconds to comeup. (Normal is less than a second.) It comes up like this:

Job:   CODEMAND07   User:   QVCODEMAND   Number:   889400

Thread: 00000002Program

Rqs   or
Lvl   Procedure   Library     Statement        Instruction
       QCMD        QSYS                            0476
       QVCODEMAND  WINTOUCHP   1000                0031
     < P_CODEMAND  WINTOUCHP
       CODEMAND    WINTOUCHP   0000000438
       QCMDEXC     QSYS                            0125
       QWCCLFEC    QSYS                            0057
     < RNI_CANCEL  WINTOUCHP
       CODEMAND    WINTOUCHP   0000001066
     < QRNX_CLOSE  QSYS        0000000023
       QDMCLOSE    QSYS                            0115
       QWSCLOSE    QSYS                            0183

Complete thread information not available.

CODEMAND is the application. Statement 438 is where it isattempting to call QCMDEXC on a SIGNOFF command. Statement1066 is the very last numbered statement in thecompilation listing, and is a generated O-spec for a fileit has no reason to be writing to at this point.

I've left some orphaned sessions up, in case anybody cansuggest anything else to look at.

To coin a phrase from a rather well-known 1958 movie thatstarred David Hedison,


HEEEEELLLLLLLPPPPP MEEEEEEEEEE!!!!

--
JHHL