Archives of the TeradataForum
Message Posted: Wed, 09 Dec 2015 @ 09:17:55 GMT
I am observing deadlock between a TPUMP job and delete running for same table at the same time.
TPUMP transaction/session are requesting row hash write locks on different AMPs at the same time. Since delete need a table level lock, they both entered into a deadlock situation where delete session was having locks on few AMPs and TPUMP having locks on others.
Target table is non-fallback and no index exists on it.
For TPUMP serialize is on so I?m wondering why TPUMP sessions need locks on more than 1 AMP at the same time.
Is pack factor playing any role here ?
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|