Archives of the TeradataForum
Message Posted: Thu, 30 Jan 2003 @ 14:18:22 GMT
We have an Account table which is Multiloaded to several times a night, by a number of different source accounting systems. Once the table has been fully loaded, a number of other jobs then reference it.
At present, none of these 'referencing' jobs can run while a Multiload is running on the Account table. We have a Read Lock on the Account table, and other jobs abend if we try to run them. I have 2 questions in this regard:
1) If we change our Read Locks to Access Locks in our 'referencing' jobs, should we be able to read the Account table even though it is being Multiloaded? (We will never be trying to read rows for the system being loaded at that point - we will always be reading rows for a system that has already loaded fully and successfully).
2) If Yes is the answer to Q.1, is there any risk that, even though we will always be accessing rows OTHER than those for the source system being loaded at the time, we might get an incorrect answer due to the fact that Multiload works at block level and some of the rows we want may be on the block multiload is using?
Julie Flanagan, AIB Bank.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|