multiple standalone jobs in one job - any performance gain?

[raiva1]

Hi All,

I have four mappings to load four different tables which are independent of each other, and each has different source

TableSrc1 --------------> TableTgt1
TableSrc2 --------------> TableTgt2
TableSrc3 --------------> TableTgt3
TableSrc4 --------------> TableTgt4

Now I want to know what and how different is it when I create 4 different jobs for each mapping and when I create 1 job which has all the four mappings in it?

Does creating 1 job for all the four mappings give me any performance benifit?


Thanks!!!!!

Rai

[chulett]

Welcome! :D

Does creating 1 job for all the four mappings give me any performance benifit?

No. In my opinion, there are no 'pros' and all 'cons' to this approach. If there are any dependancies on any of these jobs, you now need to wait for all jobs to complete. The bigger issue to me is restartability - one load has a problem and you need to rerun all four.

[hemachandra.m]

Having 4 load in the same job will not improve performance.

But for performance have 4 loads in 4 different jobs and in the sequencer make them to run in parallel. This will improve the performance a lot.

[raiva1]

Thanks chullet and hemachandra !!!!.

I had not thought about the restartability issue and sequencer.
I will explore the sequencer option now though.

Rai

[chulett]

Well... running them in a Sequence job or in a single Server job with no dependencies between the job streams will both run all four 'in parallel'. It's just that you have restartability and other options available in the Sequence job that you don't have in your original approach.