When one of the RAC node is down, CPU usage goes up a lot due to capture querying gv$instance to check on the node status.
select count(*) from gv$instance where status = OPEN and thread# = :1
Capture querying gv$instance too frequently when one of the RAC node is down
STATUS: This issue will be addressed in future release of 6.2 to query less frequently on gv$instance when one of the RAC node is down.
