1. When to consider performance issues.
Performance considerations should run through all stages of System Construction: system design, database design, code development, and system O & M.
1. system design phase:
System Architecture: Layer 2, Layer 3, and Layer 3
System type: OLAP-report generation, mainly considering the processing speed of massive data
Related reading: Oracle queries associated tables based on foreign key names
OLTP-mainly considering memory (cache), concurrency, concurrency, etc.
Estimation of system capabilities in the system life cycle: concurrency, query response time, and so on. Negative example: the ticket booking system for the 2008 Beijing Olympic Games collapsed.
2. In the database design stage, the types and attributes of database objects are highlighted:
Partition.
Indexes and their types (OLTP, generally using B-Tree; OLAP, generally using Bitmap or Text indexes ).
Parallel.
Other Object Attributes (ASSM and concurrency attributes ).
Memory parameters.
Concurrency parameters.
I/O parameters.
Database architecture (single node, RAC, distributed ).
Hardware performance (CPU, memory, etc ).
3. Efficient SQL binding to variables (OLTP)
Misunderstanding: Improper Use of Hint: forced use of index, forced setting of driver table, forced optimizer Mode
4. Object Type: Modification of attributes and Parameters
2. DBA should be involved in all stages of system construction, and developers should consider performance issues as much as possible.
Iii. optimized content:
1. Parameter: OLTP: Memory: SGA, PGA
Parallel: PROCESSES, SESSIONS
2. Object Attributes: partition and parallel
3. Index types: B-Tree, Bitmap, and Text
4. Efficient SQL
5. SQL Execution Plan
6. Analyze the table: histogram, DBMS_STATS package
Iv. optimization tools
1. SQL _Trace
2.10046 events
3. Performance view, such as: V $ SQL, V $ SESSION, V $ LOCK, V $ SESSION_LONGOPS
4.10053 event: displays all detailed operations in the Oracle execution plan
5. Optimization Tool: Hint-optimizer Mode
Access path
Table connection Sequence
Concurrency
Statspack (AWR) report and ASH report