Archives of the TeradataForum

Message Posted: Fri, 08 Jul 2011 @ 14:47:28 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Hi,

In most cases, the 'restart process' for a failed FL is what has been suggested. Drop/recreate the target table, drop error tables and then submit the job again. Note that you cannot 'delete the rows' from a table that is part-way through an FL process.

However, if the job fails in the 'application phase' - referred to in the FL manual as "phase 2" then I think your better approach is to just resubmit the FL job and allow the restart logic to handle this. This is because by the time that FL starts phase 2 then all of the data is already loaded to the Teradata system. All of the data is already in the target table, it just has to be sorted into rowhash order.

Looking at your original email below, I'm guessing that you're more familiar with Multiload (ML) than FL (myapologies if I'm mis-judging that). Remember that FL and ML work in different ways. With FL the data goes into the target table during phase 1. For most practical purposes, once phase 1 has completed, the client system (e.g. your etl server) is basically irrelevant, all of the work is done on the Teradata system. Yes, the FL processing on the Teradata system has to send messages back to the client to say "I've done this bit", but that's about all.

Cheers,

Dave

Ward Analytics Ltd - Information in motion (www.ward-analytics.com)

	<Prev	Next>		<<First	<Prev	Next>	Last>>