Hello Everyone,
(first, strange that I can't include images.... the button looks greyd-out).
running 11.2.0.2 on aix.
I have a query that ran for 3600 seconds, 3500 of which on cpu (sql).
| Total | Per Execution | Per Row |
|---|
| Executions | 1 | 1 | <0.01 |
| Elapsed Time (sec) | 3,597.00 | 3,597.00 | 29.48 |
| CPU Time (sec) | 3,241.75 | 3,241.75 | 26.57 |
| Buffer Gets | 660,102,894 | 660,102,894.00 | 5,410,679.46 |
| Disk Reads | 16,586 | 16,586.00 | 135.95 |
| Direct Writes | 0 | 0.00 | 0.00 |
| Rows | 122 | 122.00 | 1 |
| Fetches | 10 | 10.00 | 0.08 |
The EP has several nested loops with one full table scan (table has 1901 rows).
Is there any situation on which this rows/buffer gets ratio would be normal?
For me this looks like too much, given that there was actually no IO.
Thanks,
jpinela.