Thanks Dan!
You mentioned "Adding to a DB2400 table is usually not a bottleneck as
table-write does
not generate any object locks by design. Lots of i jobs can be adding to
the same table simultaneously without any waits."
From the job performance page on Systime i, I saw many connections on the
logging table, it probably because "lots of i jobs are adding to the same
table simultaneously". This is normal and won't post any problem on the
system right?
From:
"Dan Kimmel" <dkimmel@xxxxxxxxxxxxxxx>
To:
"Java Programming on and around the iSeries / AS400"
<java400-l@xxxxxxxxxxxx>
Date:
05/17/2010 12:52 PM
Subject:
RE: java stored procedure document debugging info question
Sent by:
java400-l-bounces@xxxxxxxxxxxx
It depends:
Is java calling the java stored procedure via JDBC? Does each user share
a JDBC connection with a fixed user profile? Or does each user construct
a JDBC connection with their own user profile?
Or is some non-java application on the i calling the java stored
procedure?
Adding to a DB2400 table is usually not a bottleneck as table-write does
not generate any object locks by design. Lots of i jobs can be adding to
the same table simultaneously without any waits.
On the other hand, you may be generating some queuing or long setup
depending on your architecture relative to the questions above.
-----Original Message-----
From: java400-l-bounces@xxxxxxxxxxxx
[mailto:java400-l-bounces@xxxxxxxxxxxx] On Behalf Of
Yan.Zhang@xxxxxxxxxx
Sent: Monday, May 17, 2010 10:36 AM
To: java400-l@xxxxxxxxxxxx
Subject: java stored procedure document debugging info question
Background info: System i V6R1 server, using Java stored procedures
Problem description: In a general java stored procedure I developed, I
eventually generate an SQL query statment, run it on the server,
generate a result set and return it. In order to get the information for
debugging purpose, I created a logging table under a particular schema
in the V6R1 server. When the SQL query is generated, I insert this
information together with other relevant information into that logging
table. So if this stored procedure does not run, I can debug it using
the logging table.
My question is, if multiple users login our system, and implicitly use
this stored procedure at the same time (or as a sequential way), how
does the operation of inserting the the query info into the logging
table affect the overall performance of the application? Will it become
a bottleneck as multiple jobs may need to wait to write to the logging
table?
If this design is not very good, how to efficiently document the
potential error information?
Thanks for your time reading and answering my question!
Yan
midrange.com
