× The internal search function is temporarily non-functional. The current search engine is no longer viable and we are researching alternatives.
As a stop gap measure, we are using Google's custom search engine service.
If you know of an easy to use, open source, search engine ... please contact support@midrange.com.



Charlie,

Thanks for sharing the info on STRDBMON...

I wanted to share our experiences as we endeavored to improve the processing time of some of our long-running jobs...

Prior to starting this investigation we had two "batches" of AFI extract jobs (an AM batch and a PM batch) because we could not run 74 AFI extracts at one time. (Keep in mind that you must run ALL AFI Extracts in the same single-threaded jobQ to avoid co-mingling of plant data within GL Sessions). Running that many AFI jobs simply took far too long. Between 10 and 12 hours for both batches combined, every day...! The peak-time demands on the system simply required the AFI Extracts to be performed during off-peak hours.

We run 37 Inventory and 37 Production AFI Extracts each day from Machine Manager. Inventory was running quickly but Production was not. STRDBMON quickly identified and suggested an Index be created over PCP93 and we did so.

The performance enhancement if any was not sufficient so when I saw the FIPCEXTRCT jobs were building indexes over PCP59 and PCP62 I looked at the FI* CL programs to discover the key that the OPNQRYF command was requesting the index be built for and I manually created indexes to match those key values.

STRDBMON did NOT recognize nor suggest an index be created over PCP59 or PCP62. I wonder why it failed to suggest an index be created...?! We are still at V5R2...

Once I created the indexes over PCP59 and PCP62 several of the FIPCEXTRCT jobs did run significantly more quickly. There were still some jobs that took hours to run.

Over the holiday weekend we were able to purge tens of millions of records out of 3 key files: PCP59, PCP62 and PCP93, and reorg them to remove deleted records. When that failed to reduce the processing time for those few jobs that were still running a long time we decided to look closer still.

In the submission screen of the FIPCEXTRCT job there is a flag for "Post Material Movements". The few plants that were still having a problem had these flags selected. When we took those off, the jobs finished in under 2 minutes. Recent changes in our AFI rules meant that these transactions were no longer needed for our accounting practices.

We now only retain between 90 and 120 days of data in the family of PCP* transaction files i.e. PCP01, PCP02, etc... including FIP45 & FIP51. Reduced transaction file sizes, along with the indexes we built and the removal of the unneeded flag in the FIPCEXTRCT screen has made it possible for us to run 74 AFI extract jobs every day in under two hours. And they are all done before 6:00 AM every day. That is roughly an average of 1.6 minutes per job...!

A significant performance increase without changing a single line of code...!

There are probably many opportunities within S/21 where similarly significant performance enhancements could easily be realized. Has anyone else already found any they'd like to share...? We will be looking for more as we work next on automating our Inventory Monthend Close.

These are the 3 indexes we added for the AFI Extract:
PCP59 :
Company Code              K01 CONO59
Transaction #             K02 TRNN59

PCP62 :
Company Code              K01 CONO62
Transaction #             K02 TRNN62

PCP93 :
Company Code              K01 CONO93
General Ledger Update     K02 GLUP93
Mps/mrp Model Number.     K03 MODN93

Regards,
Jeff Klipa


----Original Message Follows----
From: CMassoglia@xxxxxxxxxxxxx
Subject: [SYSTEM21]   Improving System/21 Performance
Date: Mon, 29 Aug 2005 08:26:00 -0400

If you have not analyzed access path creation in System/21 using STRDBMON,
your system probably runs far slower that it should.  We substantially
reduced our month end processing time without changing a single program.
Some on demand and daily reports that used to take minutes to run now take
seconds.  Using STRDBMON, we were able to identify temporary access paths
created by SQL, OPNQRYF, and Query/400 that either took forever to create
or were created frequently.  All we had to do was create logical paths
using the suggested key structure.

If anyone is interested, I have a short 9 page PowerPoint presentation that
describes the technique in more detail.

Thanks.

Charlie Massoglia
Chief Information Officer
cmassoglia@xxxxxxxxxxxxx
517-780-4336
Fax:  517-789-4503



As an Amazon Associate we earn from qualifying purchases.

This thread ...


Follow On AppleNews
Return to Archive home page | Return to MIDRANGE.COM home page

This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].

Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.