Otherwise, the first thing to check is to ensure that your cluster interconnect is configured correctly and that there is no other traffic on this interconnect. The interconnect should be a private network that only lets one node see any other node in the cluster. The interconnect should not be used for anything else as this can cause performance problems.
If the interconnect is configured correctly, and you still cannot determine where your bottlenecks lie, you can query GV$SYSTEM_EVENT to see the top wait events causing your performance problems. For the most part, working with wait events in RAC is just like working with wait events in a single-node environment.
Dig deeper on Oracle database design and architecture
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.