Oracle Database Connection

[kishorenvkb]

Hello All,

I have an extract job, which reads the data from an Oracle job. This job has some extensive transformation logic in it. Does this job retains the connection to Oracle even after the extract stage is completely done? The problem is... we are getting the "Snapshot is too old" error message from Oracle, even though we have completed reading the data from Oracle and it is performing some other stages in the job.

Along the same lines, does it make the connection and reserve the number of sessions requested to Teradata at the start of load job before even it gets to the Teradata Mload stage?

Please asssist.

[Sainath.Srinivasan]

Not sure what you are asking.

You are mentioning the "pipeline parallelism" behaviour of DataStage.

Are you updating the source table whilst you are reading ?

Ask your DBA to increase the segments and commit control timings.

[themixedup1]

In my experience, Datastage does not drop the connections to the servers instantly. It pulls through the data and leaves the connection open anywhere from 30 secs to 5 mins. Depending on what else it is processing. I've seen them drop faster from SQL Server but they seem to drop slower from Oracle depending on what you are doing.

[kishorenvkb]

Okay. From datastage, I am only reading the data from Oracle table. But, on the other side, the source system is updating the same Oracle table, to remove the historical rows. I am not concerned about that as I am only reading the current and future rows.
My question is that I want the connection to be droped from Oracle as soon as my Oracle stage is done extracting all the data that it should extract.

[chulett]

The connections are dropped when the job completes as that's the only time it is "done extracting all the data".