Archives of the TeradataForum
Message Posted: Wed, 14 Jan 2004 @ 10:16:23 GMT
in my opinions it is better to have several ids with tailored rights than only one id that does it all.
In our shop we have a different logon for each production job that we run. I think this is a good way to go. It enables you to do better system analysis and helps you to pinpoint problems (like runnaway queries, excessive use of spool, etc) much more easily. If you can tie back your logons to applications it helps you analyze who uses what on your system. This is very valuable knowledge if you have to implement changes & need to be sure to notify (or coordinate with) all affected parties. (Both of the above are dependent on the different logging features you have enabled as well.)
It also serves to contain problems. If you implement a new job & or say the data changes on an existing job it can lead to all kinds of problems, like for example out of spool errors. If you only use one user-id all jobs running concurrently will have to abort because of out of spool. If you have separate users only the problem job will fail.
Also, having different logons for the different jobs enables us to protect our data better. Every job is set up to only have exactly the rights it needs. I know that no other job will come along and manipulate my data, as it does not have the correct rights to do so. This makes it much harder to accidently or malicously manipulate data.
In our set-up it's not possible for a normal user (aka a person) to use one of the batch logons. I think this is important too, as it means only production jobs that leave a trail do data manipulation.
Hope you find these thoughts helpful!
|Copyright 2016 - All Rights Reserved|
|Last Modified: 23 Jun 2019|