About a great number of Rosettanet data process question

, ,
1

Dear All,

Currently I’m testing RN module in webMethods 8.0.2, I found that I can’t send a

great number of RN data, such as 1000 3A4, anyone have successful experience for

it?

Below is my test environment description:

1.webMethods version:8.0.2

2.backend db:SQL Server 2008
In Oracle 10.0.5, I have also do the test,the result is less than 200 count and end with exception
In SQL Server 2008, less than 600 count, but jdbc pool’s count has been increasing, finally block process due to reached the maximum number of connections(600)

3.fix:installed all latest fix by UpdateManager and ensure installation is no problem

4.OS:Windows 20003 enterprise 64bit

I have successful experience in webmethods 7.1.2, the number of a day is about

50000, I can’t believe the process capacity of webMethods 8.0.2 as enterprise

integration solution, please kindly give me your advise, thanks.

2

The issue have been resolved by SoftwareAG consultant, below is solved method:

The schema changes suggested are the following:

  1. PRTSTEPQUEUE Table
    DROP PRTSTEPQUEUE .PRTSTEPQUEUE_IDX;
    CREATE UNIQUE CLUSTERED INDEX PRTSTEPQUEUE_IDX on PRTSTEPQUEUE (INSTANCEID, STEPID, POS, EVENTTYPE, SUBPROCESSINDEX, SUBPROCESSSTEPITER, SUBPROCESSMODELID);

The change here is to a “clustered” index.

  1. PRTUUIDPIDXREF Table
    CREATE UNIQUE INDEX PRTUUIDXREF_IE1X ON PRTUUIDPIDXREF(UUID, MODELID, INSTANCEID, MODELVERSION);

This is a new index that will prevent table scans which increases the chances for deadlocks.

  1. PRTPROCESS Table
    DROP PRTPROCESS.PRTPRCSS_IE1X ;
    CREATE clustered INDEX PRTPRCSS_IE1X ON PRTPROCESS(CURRENTTIME, STATUS);

This changes the index to a “clustered” index which prevents table scans during Storage Cleaner execution.