Peformance issue in parellel job while writing to DB2 stage.
Posted: Wed Apr 15, 2009 1:13 am
Hi,
I am facing with problem analyzing the time taken by one job which has 2 lookups one sort, one aggregator and few transformer stage.
This job is taking 1hrs 30 min to complete (for processing 85000 rows). I have the following observations :
1> All the lookups and source query is taking less than 10 mins to fetch the data.
2> The job is taking 1hrs 30 min for processing 0 rows.(No rows were fetched from source and written to target)
3> If I run the same job by replacing the target table (DB2table ) with dataSet or sequential file than the job is running in 13 mins ( For processing same number of rows - 85000) and 7 mins for processing 0 rows.
4> In Insert / Update of target table all the fields used in where clause are indexed.
I am not able to figure out why this job is taking this much time to complete.
What may be the probable cause of such behavior??
What can be done to bring down run time of this job??
Thanks in advance....
I am facing with problem analyzing the time taken by one job which has 2 lookups one sort, one aggregator and few transformer stage.
This job is taking 1hrs 30 min to complete (for processing 85000 rows). I have the following observations :
1> All the lookups and source query is taking less than 10 mins to fetch the data.
2> The job is taking 1hrs 30 min for processing 0 rows.(No rows were fetched from source and written to target)
3> If I run the same job by replacing the target table (DB2table ) with dataSet or sequential file than the job is running in 13 mins ( For processing same number of rows - 85000) and 7 mins for processing 0 rows.
4> In Insert / Update of target table all the fields used in where clause are indexed.
I am not able to figure out why this job is taking this much time to complete.
What may be the probable cause of such behavior??
What can be done to bring down run time of this job??
Thanks in advance....