Im studying for the backup and recovery exam. One
quesiton from self test software states that SMON will use parallel query when
it detects a dread transaction with a 'large number' of rollback
blocks.
my questions:
1. is the number of parallel query slaves it uses
based on my parallel max servers? I have been setting this very high because
there are
times when I need it. I didnt see any harm in it.
Im concerned that a large rollback might suck up too many
resources.
2. What is considered a 'large number' of rollback
blocks?
|