HP3000-L Archives

April 1995, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"David W. Martin" <[log in to unmask]>
Reply To:
Date:
Mon, 10 Apr 1995 14:47:00 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (40 lines)
All:
 
Problem: We have some shared files on our HP3000 system that contain important
dates for processing data on our system. As part of the file equation defining
these files from jobs,  user logins, etc. we are using GMULT, NOLOCK and SHR.
 
FILE X.Y=X.Y;SHR;NOLOCK;GMULTI;OLD
 
Now every so often we might have a JOB running that accesses the file X in
group Y. The file contains one six byte record MMDDYY. Now this JOB runs a
Fortran program that opens the file, reads the only record in it and continues
processing. Now the file is still open in the Fortran program. Next while the
JOB is running a user logs on and runs a Fortran program with the same file
equation setup and that Fortran program accesses the same file as the JOB is
using, but the program gets an EOF and fails because of the state the job left
the file in. This state is caused by the use of GMULTI or least that is what I
think it is after reading some things on this list.
 
Anyhow someone has suggested that we use MR as well - to allow multirecord
access. This is supposed to prevent the above scenerio...will it? Can anyone
confirm that it would or has this person suggested something that will not work
 
or has been misinterrpted to work?
 
Anyhow it is REALLY frustrating. One solution might be to have the Fortran
program in the JOB to close the file immediately after opening it and reading
the one record. Thus the next Fortran program would get a file that is 'fresh'
rather than one with a file record pointer pointing to EOF.
 
There is some file system and program interaction that I am missing! Could
someone explain it?
 
 ============================================================================
David W. Martin                           The Standard Disclaimer Applies:
Lockheed Engineering & Sciences Company   Opinions expressed are my own...
Intercompany Programs Office              blah blah blah.
Houston, Texas
Email: [log in to unmask]
 ============================================================================

ATOM RSS1 RSS2