DSXchange

Hi,

Can anyone please help me out in solving the performance issue in the my job structure.

I have a big problem in extracting 30milion records from sysbase server and load it into the target database which is db2.

The datastage is on AIX server and we are using server jobs.

Can anyone suggest me how to increase the performance of extraction and loading. Present time taken is about 12hrs.

FYI there are inserts and updates in the target database.

Thanks a lot in advance.

Hi,

I think first of all you should try to seperate the insert and update to different stream. For insert process, try to use DB2 Loader utility to load data instead of normal insert function.

With this design, it should improve your job performance.

Cheers,

Hi ann_nalinee,

Check the Transaction size of DB2 input Tab.
try to split insert and update.
for update and insert the default should be 1.
hence this checking for each and every record.

or better use before sql to delete entire table and load.

Regards,
Laknar

As noted, you first need to determine the bottleneck. Split the job, land the data. Tune both halves individually. Use a hashed file to determine inserts v. updates. Consider a bulk loader for the inserts and possibly the updates as well.