Thank you for your comments.
How do you use the ZIP ROLLOUT command to release memory? I tried the
following but it made no apparent difference in the amount of memory R:Base
was using:
SET VAR vcomspec TEXT = (ENVVAL('COMSPEC'))
--
ZIP ROLLOUT &VCOMSPEC /C E
--
RETURN
In this example the "E" program is simply a small batch file that contains
the EXIT command.
Am I using the wrong approach on this?
-----Original Message-----
From: Emmitt Dove [mailto:[EMAIL PROTECTED]]
Sent: Monday, February 10, 2003 2:12 PM
To: [EMAIL PROTECTED]
Subject: [RBASE-L] - Re: Performance and memory questions
>
>I run about 40 programs each morning. I recently redesigned my technique
>using the QUIT TO syntax. It was my understanding that this uses memory
>more efficiently or will release memory when the QUIT TO is executed. I
>watched the Performance tab in the Task Manager as it steadily increased
>from an initial value of 5,546 K when R:Base started up to 64,268 K as the
>programs progressed. Is this normal? Is R:Base releasing memory
correctly?
>Is there a better way to run programs?
Mike,
We've seen similar issues, and typically have found that either sub-par
program design or outright errors can be the source of the problems. One
thing you can do is DISConnect the database periodically, CLEAR ALL VAR
NOW, then reconnect. This gives R:Base the opportunity to reorganize
itself.
If you're using nested cursors, make sure you do all the DECLAREs outside
the loops, and within the loops do OPEN cursor RESET / CLOSE rather than
DECLARE / DROP inside the loop. If you're using WHILE loops make
absolutely certain that all variables are declared before starting the
loop. If you have WHILEOPT ON, try it OFF. Try replacing the
WHILE/ENDWHILE with IF/ENDIF and GOTOs.
>I have experienced frequent Windows 2000 memory crashes running R:Base
>lately. There was some indication that there might be a problem with the
>interaction between Norton Antivirus and the Acrobat Distiller software
that
>I used to create PDF files from R:Base reports. My programs often crash
>right at the point where printing begins. I have tried tweaking the system
>and making sure I have the latest versions and patches but I am still
>getting the crashes. Again, I did not have these problems back in October.
>There have been R:Base updates and Windows 2000 critical updates since
then.
>How can I tell what has changed?
Recently I was having the memory crashes on a Terminal Server running
Metaframe XP. I looked at the scratch path, and changed it, and the
problems stopped. It may have been a space issue, it certainly was not a
rights issue. And I cannot assure you that this was the fix, but it worked.
Desperately seeking answers and stability.
As are we all!
As the Good Dr. R> recently said to me, "If you're doing everything right,
R:Base will just work." Sure, anomalies do arise, and I can attest to that
fact. I can also tell you that if you can demonstrate a reliably
replicable issue, it will get fixed.
Emmitt Dove
Manager, DairyPak Business Systems
Blue Ridge Paper Products, Inc.
40 Lindeman Drive
Trumbull, CT 06611
(203) 673-2231
[EMAIL PROTECTED]
[EMAIL PROTECTED]
