Archives of the TeradataForum
Message Posted: Thu, 08 Mar 2007 @ 16:37:16 GMT
Subj: | | Re: Info on DataStage EE For Teradata |
|
From: | | Geoffrey Rommel |
I haven't worked with DSEE for a few months, so I may not remember some details, but here goes.
| 1) When using Teradata Enterprise Stage Write does this use the Teradata Fastload utility? | |
Yes, it uses the FastLoad protocol, but it does not actually execute the 'fastload' client program.
| 4) When DataStage Partitioning is in effect and ETL SMP Server is configured for using multiple "nodes" how do sessions get created
on Teradata Target system? | |
I think you're asking how many sessions are created. This is governed by the RequestedSessions and SessionsPerPlayer options.
| 5) Based on above question can we end up with blocking on the Teradata system? If not please explain why. | |
Since FastLoad protocol is being used, I surmise that each stream will load its own subset of rows into an empty work table. Blocking would be
impossible until the final stage, when all have to be copied into the target table. At that point the usual rules for write locks would be
followed.
| 7) At what point is a connection created against the Teradata system when a DataStage job begins execution? | |
Near the beginning. Basically, as soon as the syntax is checked, all processes are started, including the Enterprise stage. If your logon
information is correct, the connection will be made then, and the stage will wait for the data to come in on the pipe.
You didn't ask about the Teradata API stage, but it makes a direct CLI connection to the server. Handy for small volumes, but do not use it for
more than 1000 rows or so.
|