dbms_stats.gather_system_stats
Hi All,
We have an OLTP environment with around 1000 tables and 500 indexes. I have a question around system stats.
I understand system stats need to be gathered while there is load on the system, and I understand it does not need to be done frequently, instead possibly only once.
We have never run this in our environment and during some recent statspack reports we noticed db_scattered_read, db_sequential_read as in the top 5 timed wait events, multiple times on multiple reports.
After reading various documents we can potentially improve this by running gather_system_stats or by amending the following parameters: