Which three statements are true about Automatic Degree of Parallelism (Auto DOP) on Exadata?
B (not c): To enable the features in 11g Release 2, use the parallel_degree_policy parameter (by default this stuff is off - parameter is set to manual). For Auto DOP, setting this to limited is sufficient. If you want more functionality (in-memory parallel processing and parallel statement queuing parallel_degree_policy should be set to auto).
Incorrect:
Not A:
* What if you don't have Exadata? Chances are that if your system is large enough to warrant Auto DOP, it is similar enough to Exadata to behave nicely with this number.
* Auto DOP is only useful in systems with enough CPUs and disks to make concurrent parallel processing feasible. It was designed mostly for Exadata, or for systems of similar size. 48 CPUs and 84 disks sound like a nice start
Currently there are no comments in this discussion, be the first to comment!