I’ve always wondered how many of my SQLs are NOT using baselines.
Of course, when you run DBMS_XPLAN.DISPLAY_CURSOR for the statement, it can put a nicely formatted note for you, something like:
- SQL plan baseline "SQL_PLAN_01yu884fpund494ecae5c" used FOR this statement
But can you find it globally for ALL SQLs in the shared pool ?
With SQL Plan Management being relatively new, it is inevitable that many people run into problems using it.
Most of those problems are caused by our (relative) ignorance: SPM does change the way how ORACLE runs SQL statements and it simply takes time to get used to how it works. Yet, some of the consequences of using SPM are truly bizarre and will surprise many people (including, probably, some ORACLE developers).
In this post I will describe the 3 scenarios where we have recently run into problems using SPM and you can judge for yourself …
In the previous post we learned what SPM building blocks or baselines are. In this post, we are going to see how we can get them to work to provide plan stability for our SQL statements.
In a nutshell, SQL Baselines will be used with any SQL statement when the following 3 conditions are met:
- SQL Plan Management must be active
- SPM Baseline for the SQL must exists (duh!)
- SPM Baseline must be ACCEPTED
In the previous post we discussed the WHAT and WHY of ORACLE 11g SQL Plan Management, in this post will will begin discussing the HOW. In my personal experience, in order to understand (and appreciate) any new thing that comes from ORACLE you really need to go down into actual nuts and bolts and see how that thing works.
So, let’s get our hands dirty and go into the actual mechanics of SPM.
One of our developers called me recently and said that they had an interesting problem on their hands.
The essence of a problem was that the schema upgrade script misfired and did not create a very important index on one of the largest tables. As a result, critical queries that were supposed to hit that index were now doing full table scans and dragging on for minutes.
The weird thing was that a developer, realizing her mistake, connected to the schema and created the index manually. Yet even after that was done, target queries still full table scanned the table, completely ignoring the index.