Hi Folks,
It still amazes me that even today I encounter *nix programmers that
have almost no idea what a "spoolfile" is, let alone the concept and how to
organize them. The closest they can get to it is a "nohup.out" file, which
is always overwritten with the next nohup process. What Paul just
demonstrated and is available on every "MPE/iX" system and what we
all(well...almost all) take for granted is beyond the grasp of most of the
programming world. Many days I feel I have taken a giant step backwards....
Just my two cents when I would have given almost anything to have had a
"spoolfile" for the aborted *nix process that I ran today.
Eric Sand
-----Original Message-----
From: HP-3000 Systems Discussion [mailto:[log in to unmask]]On
Behalf Of Paul Christidis
Sent: Wednesday, December 01, 2004 5:31 PM
To: [log in to unmask]
Subject: Re: [HP3000-L] Spool files
Hi Greg,
I found myself in a similar situation some years back. The 'solution'
that I've implemented, using NBSPOOL, is as follows:
1. A job is run every morning to delete any spoolfiles that based on a
variety of parameters. (priority, outclass, age, etc.)
2. A record of each spoolfile that qualifies for deletion is written on
the job's $stdlist as follows:
PURGE DEVICE=NOPRINT,<=TODAY-15;SHOW
25 Spool Files Qualify
SFID FILENAME JOB USER NAME SECTORS DEVICE PRI WHEN
CREATED
----------------------------------------------------------------------------
---
#O12482 $STDLIST #J3018 MANAGER.SYS 64 NOPRINT 8 11/15/04
00:10
#O12483 $STDLIST #J3019 MANAGER.SYS 32 NOPRINT 8 11/15/04
00:10
...
#O12841 $STDLIST #J3072 MANAGER.SYS 48 NOPRINT 8 11/15/04
05:00
#O12887 C6998ALL #J3082 PROG5.VXL 576 NOPRINT 8 11/15/04
06:00
#O12893 C7283ALL #J3085 PROG5.VXL 432 NOPRINT 8 11/15/04
06:10
...
3. The above $stdlist is emailed to myself and another programmer and
archived.
4. When a specific spoolfile is needed it can be located using the
recorded information.
5. The restore job can then use the 'SFID' to restore the specific spool
file.
I implemented the above around August of 2001 and have used it a few times
since then.
Regards
Paul Christidis
HP-3000 Systems Discussion <[log in to unmask]> wrote on 12/01/2004
04:34:45 PM:
> We just had an exciting experience... We had some apparently important
> reports that ran on Thanksgiving and did not get printed. We only keep
two
> day's worth of spool files on the system. So, we were restoring our STD
> file, and from that, all of our spool files from Friday. Then we began
to
> get reports of slow system response time, and our N class (MPE 7.0)
> consistently has sub-second response time for the application. After
> checking some of the more obvious possibilities, I called Beechglen. Jon
> Backus and Doug Werth found that the spooler was consuming all available
> CPU. Apparently, the spooler was having trouble both spooling to print,
and
> having restore write files to OUT.HPSPOOL. That sounds like a bug. So,
> beware!
>
> Now, among the many challenges we have with how we handle printing is
how we
> store and restore our spool files, such that getting back some of our
spool
> files means restoring all of them. There has to be a better way. Short
of
> buying a product, what are some better ways?
>
> Greg Stigers
>
> * To join/leave the list, search archives, change list settings, *
> * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|