RE: JDBC performance issues after moving to Java 8 -- JAVA400-L

There is a Mutex trap in QMGTOOLS you might want to look into. iDoctor
comes in handy here too since the Mutex is found in a PEX trace.

http://www-01.ibm.com/support/docview.wss?uid=nas8N1019977

--
Jim Oberholtzer
Agile Technology Architects

-----Original Message-----
From: JAVA400-L [mailto:java400-l-bounces@xxxxxxxxxxxx] On Behalf Of Mark
Murphy
Sent: Friday, July 06, 2018 1:15 PM
To: Java Programming on and around the IBM i <java400-l@xxxxxxxxxxxx>
Subject: Re: JDBC performance issues after moving to Java 8

The utility Charles pointed me to is not showing a DNS issue. what we are
seeing is that the program sits in TIMA status most of the time, and
periodically shifts over to RUN just long enough so that if we sit WRKACTJOB
and repeatedly press F5 it will go to RUN and immediately back to TIMA every
once in a while. Another thing we are seeing is that there are generally
three Mutexes out there for the job, but once again if we repeatedly refresh
the screen, that periodically will drop down to 1, and immediately back to
three.

On Fri, Jul 6, 2018 at 1:20 PM Jim Oberholtzer <midrangel@xxxxxxxxxxxxxxxxx>
wrote:

I can't answer the changes to Java question, Tim would have to respond
to that. My situation is experiential. Assuming Networking (as Tim
pointed
out) is not the issue, we've found some applications do run better in
the 64bit version.

Paging is clearly not an issue on your system.

--
Jim Oberholtzer
Agile Technology Architects

-----Original Message-----
From: JAVA400-L [mailto:java400-l-bounces@xxxxxxxxxxxx] On Behalf Of
Mark Murphy
Sent: Friday, July 06, 2018 11:47 AM
To: Java Programming on and around the IBM i <java400-l@xxxxxxxxxxxx>
Subject: Re: JDBC performance issues after moving to Java 8

I'm not sure it's an issue with the pool it's in, we have 10Gb in
*BASE, and 42Gb in *INTERACT, and very low paging and faulting (0 in
*MACHINE pool, 10 in *BASE, 6 in *INTERACT) those are total numbers
between DB and Non-DB.

I have thought about using 64bit. We aren't running out of memory
though that I am aware of, and I was under the impression that using
the 64bit JVM is less performant unless you need it. Have changes been
made to Java to invalidate that?

On Fri, Jul 6, 2018 at 11:25 AM Jim Oberholtzer
<midrangel@xxxxxxxxxxxxxxxxx>
wrote:

Consider moving to the 64bit version. I've had some success with
improvements that way. Depending on how memory intensive the
application is you might have to add memory to the shared memory
pool where it's running.

Which brings up another point, you should be moving these processes
into a shared memory pool and getting them out if *BASE.

--
Jim Oberholtzer
Agile Technology Architects

-----Original Message-----
From: JAVA400-L [mailto:java400-l-bounces@xxxxxxxxxxxx] On Behalf Of
Mark Murphy
Sent: Friday, July 06, 2018 7:58 AM
To: Java Programming on and around the IBM i
<java400-l@xxxxxxxxxxxx>
Subject: JDBC performance issues after moving to Java 8

We have a Java application running on IBM i v7.1. It was working
fine with Java 6, but when we moved to Java 8 in preparation for an
upgrade to v7.3, the program started taking over an hour to run. As
far as I know, the only change was to move to Java 8. We are using
the 32 bit JVM for both Java 6 and Java 8.

Any ideas on where to look for issues? We are not getting any errors
that I know of, and everything is running through properly. It just
takes over an hour for the program to run now when it used to take
minutes. The application is using JDBC to communicate with a MS SQL
Server. It is basically synchronizing data for another process that
runs

on Windows.

--
This is the Java Programming on and around the IBM i (JAVA400-L)
mailing list To post a message email: JAVA400-L@xxxxxxxxxxxx To
subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/java400-l
or email: JAVA400-L-request@xxxxxxxxxxxx Before posting, please take
a moment to review the archives at

https://archive.midrange.com/java400-l.

--
This is the Java Programming on and around the IBM i (JAVA400-L)
mailing list To post a message email: JAVA400-L@xxxxxxxxxxxx To
subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/java400-l
or email: JAVA400-L-request@xxxxxxxxxxxx Before posting, please take
a moment to review the archives at
https://archive.midrange.com/java400-l.

--
This is the Java Programming on and around the IBM i (JAVA400-L)
mailing list To post a message email: JAVA400-L@xxxxxxxxxxxx To
subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/java400-l
or email: JAVA400-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at https://archive.midrange.com/java400-l.

--
This is the Java Programming on and around the IBM i (JAVA400-L)
mailing list To post a message email: JAVA400-L@xxxxxxxxxxxx To
subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/java400-l
or email: JAVA400-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at
https://archive.midrange.com/java400-l.

--
This is the Java Programming on and around the IBM i (JAVA400-L) mailing
list To post a message email: JAVA400-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/java400-l
or email: JAVA400-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at https://archive.midrange.com/java400-l.