slowness of DML, and ADDM complaining: Investigate the possibility of improving the performance of I
We are testing one new server and planning to move production to that.
This is oracle 10g and O/S is linux.
We are observing ADDM is reporting:
Action Investigate the possibility of improving the performance of I/O to the online redo log files.
The average size of writes to the online redo log files was 1057 K and the average time per write was 39 milliseconds.
Our SysAdmin is saying, the files are already on good I/O SAN, and since I/O is very high then latency also high.
He feels exessive log_buffer may be cause of this.