JRun - Fast on notebook, slow on server

[CraigEHughes]

Hi, Please help...

I have developed a struts J2EE application using JRun 4 and MS SQL
server on Windows 2003 Server. Part of the application is an early
morning routine which updates 73K records from a local Flat file
provided by an AS400 hub.

Now my problem... on my Dell Inspiron 6000 notebook, this routine runs
in about 62 seconds, however, the same application, same flat, file
takes 7 minutes on a production spec server?

The details of these two machines are are follows:

Notebook:
Memory: 512MB
Processor: Intel Pentium M 1.73 Ghz

Server:
Memory: 2Gig
Processor: Intel Xeon 3.00 Ghz

Any ideas?

 

[David Segall]

Hi, Please help...

I have developed a struts J2EE application using JRun 4 and MS SQL
server on Windows 2003 Server. Part of the application is an early
morning routine which updates 73K records from a local Flat file
provided by an AS400 hub.

Now my problem... on my Dell Inspiron 6000 notebook, this routine runs
in about 62 seconds, however, the same application, same flat, file
takes 7 minutes on a production spec server?

The details of these two machines are are follows:

Notebook:
Memory: 512MB
Processor: Intel Pentium M 1.73 Ghz

Server:
Memory: 2Gig
Processor: Intel Xeon 3.00 Ghz

Any ideas?

Are you sure your app is the only one running on the server in the
"early morning"? It's a favourite time for scheduling batch jobs.

 

[Chris Uppal]

I have developed a struts J2EE application using JRun 4 and MS SQL
server on Windows 2003 Server. Part of the application is an early
morning routine which updates 73K records from a local Flat file
provided by an AS400 hub.

Not sure what you mean by "a local Flat file provided by an AS400 hub". I'm
assuming it's a file containing unindexed records accessed via normal (for
Java) file-based IO.

Now my problem... on my Dell Inspiron 6000 notebook, this routine runs
in about 62 seconds, however, the same application, same flat, file
takes 7 minutes on a production spec server?

So many possibilities. Which have you already eliminated ?

Can the server read the file (just as a file, without doing any DB work) as
fast as the notebook ? If not then it might be:
the different OSs' implementations of how they read data off the AS400.
Different network paths between the three machines.
Different network loads between the two runs.
...

If it can read the raw data approximately as fast:
Are the databases identical (i.e. identical contents) ?
Do you have the same indexes on the same tables (easy to miss) ?
Do you have the same configuration (logging, charset, default isolation-
level, etc, etc, etc) on the two machines ?
Is there something wrong/misconfigured with the server's disks ?
...

And so on...

The first step I'd take is to copy the file onto each machine's local disk
(time it), and then run the batch from that. That should tell you where the
bottleneck is. Then when you run the batch, look to see what's maxed-out on
the server. You can also use the Windows system monitoring to see what's
different.

-- chris