The problem: A few minute spike on control file sequential read event.
The cause: A logon storm.
The proof: A logon trigger querying v$database (that’s a controlfile access). Hence, the wait event.
The solution: Either remove the query from the trigger or keep the query and find out the reason of excessive logons at that interval given that it will be eliminated on that side. The latter is the correct way to go.
The World’s Shortest Troubleshooting Scenario
Posted in Oracle.