Archives of the TeradataForum
Message Posted: Thu, 20 Feb 2003 @ 13:51:32 GMT
We have a standard that before multiloading to a permanent table, we always run a Release Multiload. A few times lately, however, Users left queries running overnight, and Release Multiload steps in our batch had to wait until queries had finished. (Unfortunately, on 2 occasions, this was 7-8 hours later!)
We need to allow our Users to leave queries running without interrupting the batch/nightly loads to our tables. (We considered killing User queries, but this is not an option). We are now considering removing the RM step.
What are other people's views on this? Do others run RM steps before all loads as a standard? (This was strongly recommended to us by NCR when we started out).
A possible alternative, put forward by one of our team members, is to check before a load for the presence of one of the Log or Work tables, using Show table command (to indicate something failed in a previous run). If the table does not exist, an error (3807) will be returned. We then know something has failed, and could use our existing logic to release mload. But it would mean that the release mload ONLY runs in an erroneous situation. Anyone any comments on this?
We assume many companies have user queries running at the same time as table loads and would be interested in hearing any suggestions or ways we can handle this.
Our DBA has suggested that Tpump may be of use (?) We do not have Tpump at present - so any comments in this regard most welcome!
We are using Teradata V2R3 at present, very shortly moving to V2R4.1.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|