Archives of the TeradataForum
Message Posted: Thu, 10 Dec 2015 @ 09:20:25 GMT
Thanks for the responses.
I believe that is what happening here. Within a "pack" not only common value rows but different rows with different values(of key) are being sent. This causes transaction to demand locks on more than 1 AMP and it conflicts with the all AMP delete query leading to deadlock.
In the viewpoint as well, I can see that AMPs participating for TPUMP sessions are 30 and more. That also proves the point and behavior.
Any suggestion to improve and reduce deadlock frequency in such scenarios where TPUMP is 24X7 job and all AMP bteq delete also comes in between?
|Copyright 2016 - All Rights Reserved|
|Last Modified: 23 Jun 2019|