Re: Question on how the AS/400 handles "spooling"

Good to hear from you too Dennis !

Thanks for the feedback and validation :-)

Hopefully we will both learn something from this !

Chuck

Dennis Lovelady wrote:

> Hi, Chuck:
>
> Nice to see your smiling face.
>
> >to save so I restarted it at page 1901 and it took FOREVER to do
> >...
> >and while this was going on it CREAMED our other printers. A
> >printer would ask for form type or get alignment or end of forms
> >messages and take 5 or 10 minutes to do anything. There were INCREDIBLE
> >pauses between each spool file printed on these other printers. The
> >system was NEVER over 20 % CPU and that was the peak. The writer for the
> >5si did show towards the top of the list but with less than 4% CPU.
> >Interactive and batch response were GREAT.
>
> >WHAT was going on with spooling ?
>
> It was not taking a break - it would have been blowing through the spool
> data, counting pages, until it came to the start point.  Page 1901 doesn't
> sound like a big deal - I'd expect the /400 to take that in stride, and in a
> hurry.  However....  :^)
>
> The other system performance being good (and CPU utilization being low) does
> not surprise me at all.  Work Management defaults for spool jobs set them at
> priority 52, if I recall (maybe that was MY standard, rather than IBM's?), so
> spool jobs will not impact other work under most circumstances.
>
> It sounds like the spooler is operating in the manner of an MRT program, and
> that's a little frightening, and I don't truly believe that is the case.  But
> MRTs would do this.  Bruce?  Simon?
>
> >HOW is this handled on the system ?
>
> Your spooling subsystem ACTLVL is probably set pretty low (WRKSBSD) and
> TIMESLICE may be set too high (WRKCLS) - you might want to check these.  I
> would expect the system to go into a long wait for disk accesses now and
> then, which should throw out the job and allow other printers to have some
> CPU.
>
> If you mean, "how does the system find the right page, and continue from
> there?" my philosophy (never having seen da code) is that it goes through all
> the same steps as the actual print process, but sends the output to the bit
> bucket until it reaches the designated point (p. 1901 in your case), where it
> starts sending data to the printer.  As I said, I have not seen the code for
> this, but based upon your analogy and upon watching the effects of certain
> actions over the years, I have come to this conclusion.
>
> >In all my years with the AS/400 I have NEVER seen anything like this. It
> >was like the one BIG job (and all in all this wasn't THAT big of a job)
> >was TOTALLY dominating QSPL or whatever, to the detriment of other
> >writers but yet otherwise not impacting the system AT ALL (???????
> >!!!!!!!!).
>
> I suppose, overall, my response is not as much of a helper as it is in
> support of your question.  This is interesting, to say the least.  It'll be
> interesting to hear what those who know QSP* have to say.
>
> Dennis
> --
> Dennis Lovelady                    Simpsonville, SC
> mail: dennis@lovelady.com
> URL:  http://lovelady.piedmont.net
> ICQ:  5734860
> --
> "He is so unlucky that he runs into accidents which started out to happen to
> somebody else."
>         - Don Marquis
>
> +---
> | This is the Midrange System Mailing List!
> | To submit a new message, send your mail to MIDRANGE-L@midrange.com.
> | To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
> | To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
> | Questions should be directed to the list owner/operator: david@midrange.com
> +---

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---