Jay,

Sounds like you want something that will do a proper "completion check" as
well as sequencing. It also sounds like your jobs may not run in an
immediate sequence, but have potentially long pauses between them.

You could take a couple different approaches here. One of the things you
could do is not only have your jobs check for the presence of a dummy file,
but check for some kind of completion status within that file. Have each job
echo BEGIN or START (or whatever word you choose) into the dummy file at the
beginning of the job, then echo DONE or COMPLETED into the dummy file at the
end of the job. The following job would read the dummy file as input and
check for the appropriate word. If it sees START instead of DONE, the job
dies.

Another approach would be to use *two* dummy files for each job. Echo the
job number into FILEA at the start of the job, and into FILEB at the end of
the job. The next job would read these two files and compare the job
numbers. If they don't match, kill the job with an appropriate error
message. I particularly like this approach because it could also let me know
which job I'd need to go check to find out where the problem occurred. :-)

Also, you may want to put ACDs onto your dummy files so only the $OWNER
(i.e. the batch job's user ID) has the capability to purge them. Just a
thought.

HTH,
Patrick
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Patrick Santucci
HP e3000 Systems Administrator
Cornerstone Consolidated Services Group

> -----Original Message-----
> From: Jay Willis [SMTP:[log in to unmask]]
> Sent: Tuesday, June 19, 2001 11:40 am
> To:   [log in to unmask]
> Subject:      [HP3000-L] Getting jobs to run in proper sequence
>
> Hi all,
>
> I am seeking suggestions on how we might get a series of jobs to run in
> the
> proper sequence. Currently the jobs, in turn, build/check for the
> existence
> of/purge dummy files. I don't care for this approach since a file may
> inadvertently be renamed, purged, etc. between jobs. This happened to us
> recently when a job aborted, thus not building the dummy file that the
> subsequent job was to check for, allowing the user to stream a job out of
> sequence. The users can make changes to the data, based on the reports the
> jobs produce, so that after job A runs, they go over the reports, make
> changes, then stream job B, and so forth.
>
> Any ideas/thoughts gleefully accepted.
> Jay
>
> * 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 *