Problem Description:
Project Feedback The database server's CPU lasts 100%, Tracing the wait events that are found in many active sessions is "Asynch descriptor resize", and these sessions are always in the active state, and the queries executed by these sessions are not complex and feel strange.
Review the data for the Oracle Bug 9829397,excessive CPU and many "Asynch descriptor resize" Waits for SQL using Async IO, this bug occurred on 11.2.0.1 and 11.2.0.2, 11.2.0.3 fixed the issue.
setting hidden parameters is also a workaround: set Disk_asynch_io to False to disable asynchronous I/O: Alter system set Disk_asynch_ Io=false
After modifying the parameter configuration and rebooting, the problem remains that the SQL is still in the execution state, but the wait event becomes: DB file scattered read
attempts to db_file_ Multiblock_read_count, problem solving.
Sql> show parameter db_file_multiblock_read_count
sql> alter system set DB_FILE_MULTIBLOCK_ read_count=10000; --attempted to change to a super-large value
Tested, and finally defined the value of db_file_multiblock_read_count as a maximum (or 256, 128?
Oracle Bug A script is always in the execution state, waiting for the event: Asynch descriptor Resize