" DB2 performance tuning and optimization"
1. Performance optimization should span the entire life cycle of the application system, not when users feel poor performance
2. For any program, the slowest and most time-consuming operation is from the disk (the number of revolutions of the disk has not been greatly improved). In any database system, there must be a large amount of I/O , so I/O is the cause of The key to database performance problems.
3. Data access in DB must go through the buffer pool: the read data must be read to the buffer pool before it can be submitted to the application; the written data must also be written to the buffer pool before I/O can be performed . The buffer pool is the parameter that affects DB performance the most.
4.Greate BufferPool statement can add more buffer pools
5. Lock-related problems: lock waiting, lock escalation, and deadlock-related problems
Lock waiting : When the lock requested by the application for an object cannot be obtained immediately, the application enters a queue waiting for service, waiting for the application occupying the lock to submit or roll back to release the lock;
Lock escalation: from row level to table level
6. Multiple inserts into the same table can use value followed by multiple (), but Informix does not support it.