Re: "Stalling"

Dean,
If you're looking at the save job from WRKACTJOB, then seeing 0% against the
job, especially near the beginning of the job, is not unusual.  The 
system  load/dump tasks will be eating lots of CPU.  The total CPU being
used WILL be reflected in the total CPU at the top of the display, but not
against the save job.  If you have IBM's Performance Tools, you can see the
load/dump tasks (starting with 'LD') by using the WRKSYSACT command.
As for changing the job's priority - if there are no other jobs competing
for the system resources, it wouldn't really matter whether it was running
at priority 1 or 99.  If a job isn't CPU bound (ie. is limited by disk I/O)
then it isn't going to eat up a higher percentage of CPU just because you
change it's priority.



On Sun, 4 May 1997 DAsmussen@aol.com wrote:

> Chris,
> 
> In a message dated 97-05-04 04:04:35 EDT, you write:
> 
> > Occasionally, the machine will "stall". Their will be plenty of jobs
> >  sitting in batch, but all of them will be getting zero CPU. Even with the
> >  CPU utilization at <1%. Just as there is nothing apparently starting the
> >  stall, it appears to end for no reason. 
> 
> I would appreciate any insight into this as well.  I never noticed it until
> today when we ran our first Production run of our BPCS shop order purge
> program.  We're on a maxed-out 320 with the latest CUM of OS/400 3.1.01.
>  This program incorporates a command utility that I wrote that performs
> DSPPGMREF on all programs in the series to an outfile, then performs a SAVOBJ
> to a SAVF for any file open for update or output.
> 
> By chance, a file required for the series didn't get promoted to the
> production machine and I had to start and keep an eye on it manually.  During
> the SAVOBJ portion, the job had a status of EVTW and zero CPU.  ZERO CPU
> during a SAVOBJ??!!  The program also has INTENSIVE SQL against large files,
> and opens THIRTY of them.  After the SAVOBJ completed, the job still took
> less than 1% CPU with nobody on the system -- after I bumped the priority up
> to 20!  This job should DEVOUR resources, yet received little.  What's up
> with that?
> 
> TIA,
> 
> Dean Asmussen
> Enterprise Systems Consulting, Inc.
> Fuquay-Varina, NC  USA
> E-Mail:  DAsmussen@AOL.COM
> 
> "If a friend won't lend you fifty dollars, he's probably a close friend." --
> Anonymous
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
> * This is the Midrange System Mailing List!  To submit a new message,   *
> * send your mail to "MIDRANGE-L@midrange.com".  To unsubscribe from     *
> * this list send email to MAJORDOMO@midrange.com and specify            *
> * 'unsubscribe MIDRANGE-L' in the body of your message.  Questions      *
> * should be directed to the list owner / operator: david@midrange.com   *
> * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
> umidr
> 
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* This is the Midrange System Mailing List!  To submit a new message,   *
* send your mail to "MIDRANGE-L@midrange.com".  To unsubscribe from     *
* this list send email to MAJORDOMO@midrange.com and specify            *
* 'unsubscribe MIDRANGE-L' in the body of your message.  Questions      *
* should be directed to the list owner / operator: david@midrange.com   *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *