How to accelerate a job running on Enterprise?

[Swordfish]

Hello,

I have JDE OneWorld XE release B7333 service pack 23. Can anybody tells me how to accelerate (make it finish faster) a job running on Enterprise server under Queue - QB7333? Because I have a scheduled job that is taking so much time when it runs... how can I accelerate it?
Note that I have SQL server 2005 as my DBMS on Enterprise.

 

[Larry_Jones]

On the View Submitted Jobs screen, pressing Find repeatedly tells the server that you're in a hurry and to speed up its processing

 

[DSauve]

Cruel, cruel!

 

[Sebastian Sajaroff]

Hi,

There are at least 3 ways to do it :

1. Tune that particular job (you'll need a DBA and
a JDE developer)

2. Tune your whole OS/DB infrastructure (only if you
notice that all jobs are slow, in that case you'll need
an OS/DB expert, typically a CNC).

3. This one is quite expensive, requires a spaceship
and a senior astrophysicist.

Move your servers close to a neutron star or a
black hole so that space-time framework distorts, you'll
probably notice that your jobs run 10x or 1000x faster.

Don't try to install them inside your black hole event
horizon unless you want them to be crushed by tidal
forces and lose contact forever.

Alas, I love astrophysics but I don't have the spaceship!

 

[Adrian_Chimirel]

Accelerating a job? Nope

Debugging & fixing it? Yep

PS Is it vanilla? I remember R12803 Identify New Entries was looking (in 911, of course) for a blank (empty) field, and replacing it with a(nother) blank value

But we caught it with SAP's Business Objects Data Integrator, that refreshes the Data Warehouses (it was replicating ALL these records (yes, millions), giving MSQL Server a really hard time).

 

[Mick K]

Some JDE batch processes are very inefficient, especially when deleting or updating many records. A JDE developer can usually identify where improvements may be made by looking at the code and at debug logs.
Is there a specific UBE with this problem ?

 

[moeshaker]

For short term solution, split the job into versions and split the data selection over different version and change the Queue to multithreaded.
We did this in cost rollup , split the job into 5 versions, each version data selection for a a specific range of items. the cost rollup used to take 24 hours not its done in 9 hours.

 

[fi0317]

some advice:
1. a good developer to design the UBE reports more effective.
2. modify the priority of the file JDE.INI or jas.ini (I am not sure which one)
3. create specfic queue for the report you taled about.