Processes Mysteriously Being Deleted

[Jan van den Ende]

Robert,

>>>
I can see from the batch logfile that the process was at a WAIT statement at the same time the BUGCHECK ocurred.
<<<

Please DO realise that batch job logs are only flushed ever so often.
A killed job often does NOT get the chance to write the last part of the logfile.

So, concluding that the process deletion happened during WAIT is by no means warranted.

Just my EUR 0,02

Proost.

Have one on me.

jpe

[Robert Atkinson]

Am aware of log flushing - set to 5 seconds on all of our processes anyway - but the termination time and the date/time in the log file coincide to show the process was running a WAIT.

Rob.

[Jon Pinkley]

Rob,

RE:"Am aware of log flushing - set to 5 seconds on all of our processes anyway - but the termination time and the date/time in the log file coincide to show the process was running a WAIT."

Sorry, even though I am not from Missouri, please show me. Where did you get the termination time? My guess is the job completion time from the retained queue info. What do you mean by "date/time in the log file". To me these are both ambiguous.

And quite a lot can happen in 5 seconds.

Jon

[Robert Atkinson]

"Completed 27-DEC-2006 18:50:12.88 on queue BETA$OPERATIONS"

ROB$ type gbsinvrun.log;997
$! RUNNING SYLOGIN.COM
$!
.......
27-DEC:18:48:45> $ DELETE /LOG SYX:TMPLOC.INT;*
%DELETE-I-FILDEL, SY1:[LIVE.DAT]TMPLOC.INT;1 deleted (290 blocks)
27-DEC:18:48:46> $ COPY /LOG SYX:INVTEL.WRK DEV$INTAREA:*.TW3_22J7261
%COPY-S-COPIED, SY8:[LIVE.DAT]INVTEL.WRK;1 copied to DEV$INTAREA_ROOT:[20080122]INVTEL.TW3_22J7261;1 (11281 blocks)
27-DEC:18:48:46> $ DELETE /LOG SYX:INVTEL.WRK;*
%DELETE-I-FILDEL, SY8:[LIVE.DAT]INVTEL.WRK;1 deleted (70035 blocks)
27-DEC:18:48:46> $!
27-DEC:18:48:46> $ WAIT 00:10:00

[Jon Pinkley]

Thanks,

I assume there was also a bugcheck and it had its finger pointed at the DBSINVRUN job?

I have to agree that the log as shown seems to indicate that the process was in a DCL wait at the time.

Is anything throwing Special Kernel Mode AST's at the process? Does DCL have some file open that has RMS global buffers associated (I'm reaching here).

Hein, if you are still reading this thread, do you have any idea how a process could be generating an RMS bugcheck while it is in a WAIT if it isn't related to exec or kernel mode AST's being involved?

Jon

[Jon Pinkley]

I had a typo, I meant GBSINVRUN not DBSINVRUN.

[Willem Grooters]

What follows WAIT? It could be the process is deleted AFTER wait and before the buffers are flushed (a crash can ahppen within 5 seconds :)

[Wim Van den Wyngaert]

Willem,

In that case the wait is not working correctly. He asked to wait 10 minutes ...

Wim

[Wim Van den Wyngaert]

Rob,

Is the job keeping a task-to-task open ?
If it doesn't receives any information for 75 seconds, the remote stuff stops (75 coming from our already gone 6.2 systems). May be this caused the bugcheck because the abort is about 60 seconds after the wait started.

Fwiw

Wim

[Robert Atkinson]

Presumably you're talking about DECnet or something?

Nothing like that happens in the process.