We have a query with CPU less than 20 and for a particular day it was stuck in MLK step for 20 minutes. It tries to put a read lock on around 10 different tables.
Later runs have executed within a second. However for 1 day it was stuck in the MLK step for about 20 minutes. The DBQL does not show any writes on any of those tables during that time. Can use some help trying to figure out what was the issue for that particular day.
Maybe there was a not yet commited Begin Transaction followed by a RowHash-lock INSERT/UPDATE/DELETE?
Do a Rewind in Viewpoint and check which session caused the block.
As well as a concurrent backup, were there any load utilities running against one/more of the tables involved (Fastload, Multiload etc.) ?
Thanks for taking the time to reply guys. It seems that the System was heavily degraded for that period which might have caused the issue. Unfortuately the queries we were looking at got stuck in the MLK step. Yes the previous "pseudo-lock" steps ran successfully