Re: Trigger

[theis_richard@xxxxxxxxxx]

Norbert wrote:
> Every trigger program I use takes such a lot of performance, so I have to
> inactivate the trigger.
> Is this a general trigger problem, or are there some things I have to
> pay attention to ? Thanks.
Generally speaking, you have to pay attention as carefully as any other
frequently called program: avoid returning with LR on, avoid ACTGRP(*NEW),
and so on.  Could you elaborate: is your trigger OPM or ILE, RPG or COBOL,
opening files or not (I suspect it does), ...etc ?
A trigger add an other performance constraint; as it has to be fired for
each record, it forces the unblocking.

Alex wrote:
> All trigger programs are written in CL and do nothing but pass the
> trigger parameters over to the "real" trigger program.  This allows
> modifying the "real" trigger without having to have an exclusive lock on
> the file.
Good idea, but as far as performance is concerned, CL could not be the best
choice (RPG with LR off avoids reinitialisation).
As Alex outline it, a trigger should not do a lot of work synchronously,
but only "trigger" operations that ideally are done asynchronously in
background.  But it depends of the needs...

Best regards
Richard THEIS
AS/400 Education, France
  theis_richard@fr.ibm.com


+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---