老李分享： Oracle Performance Tuning Overview 翻译

老李分享： Oracle Performance Tuning Overview 翻译

 

poptest是国内唯一一家培养测试开发工程师的培训机构，以学员能胜任自动化测试，性能测试，测试工具开发等工作为目标。如果对课程感兴趣，请大家咨询qq：908821478，咨询电话010-84505200。

1 性能优化概述

This chapter provides an introduction toperformance tuning and contains the following sections:

本章节给出了性能优化的简介，它包含了一下几个部分。

• Introduction to Performance Tuning
• 性能优化的介绍
• Introduction to Performance Tuning Features and Tools
• 性能优化工具和特点的介绍

1.1 性能优化简介

This guide provides information on tuning anOracle Database system for performance. Topics discussed in this guide include:

这个小节提供了对一个oracle数据库系统进行性能优化的知识。包括一下几个主题

• Performance Planning
• 性能规划
• Instance Tuning
• 实例优化
• SQL Tuning
• Sql优化

1.1.1性能规划

Before starting on the instance or SQL tuningsections of this guide, make sure you have read Part II, "Performance Planning".

在学习本小节的实例和sql优化部分之前，请先阅读II部分：性能规划

Based on years of designing and performanceexperience, Oracle has designed a performance methodology. This brief sectionexplains clear and simple activities that can dramatically improve systemperformance. It discusses the following topics:

基于多年的设计和性能方面的经验，oracle已经设计了一套关于性能方面的方法体系。本节描述了能够显著提高系统性能的清晰和简单的方案。该方案涉及到一下主题:

• Understanding Investment Options
• 理解能够在调优方面起作用的所有因素
• Understanding Scalability
• 理解可扩展性
• System Architecture
• 系统架构
• Application Design Principles
• 应用设计准则
• Workload Testing, Modeling, and Implementation
• 负载测试 建立模型，实现
• Deploying New Applications
• 部署新的应用

1.1.2实例优化

Part III, "Optimizing Instance Performance" ofthis guide discusses the factors involved in the tuning and optimizing of anOracle database instance.

本节的III部分，“优化实例性能”讨论涉及到oracle数据库实例调优和优化的各种因素。

When considering instance tuning, care mustbe taken in the initial design of the database system to avoid bottlenecks thatcould lead to performance problems. In addition, you need to consider:

当考虑实例优化的时候，数据库系统初始化设计必须要注意防止一些瓶颈，这些瓶颈可能会带来性能方面的问题。此外你需要考虑一下几点：

• Allocating memory to database structures
• 给数据库分配内存
• Determining I/O requirements of different parts of the database
• 确定数据库的不同部分的I/O需求
• Tuning the operating system for optimal performance of the database
• 调优操作系统使数据库系能达到最佳

After the database instance has beeninstalled and configured, you need to monitor the database as it is running tocheck for performance-related problems.

在安装和配置数据库实例之后，你需要监控数据库，因为它时刻运行着来检查跟性能相关的一些问题。

1.1.2.1性能准则

Performance tuning requires a different,although related, method to the initial configuration of a system. Configuringa system involves allocating resources in an ordered manner so that the initialsystem configuration is functional.

对于系统的初始化配置来说 性能调优需要一种与之不同的尽管相关的方法。要以有序的方式分配资源来配置一个系统，这样系统的初始化配置才能起作用。

Tuning is driven by identifying the mostsignificant bottleneck and making the appropriate changes to reduce oreliminate the effect of that bottleneck. Usually, tuning is performedreactively, either while the system is preproduction or after it is live.

通过找到最重要的瓶颈且做一些合适的变更来减少或消除瓶颈的影响来驱动调优的进行。通常来说，调优是需要有反应来验证的，这个系统要么是在试运行中，要么是已经在生产运行了。

1.1.2.2基线

The most effective way to tune is to have anestablished performance baseline that can be used for comparison if aperformance issue arises. Most database administrators (DBAs) know their systemwell and can easily identify peak usage periods. For example, the peak periodscould be between 10.00am and 12.00pm and also between 1.30pm and 3.00pm. Thiscould include a batch window of 12.00am midnight to 6am.

最有效的调优方式是建立一个性能的基准线，它能够被用来在性能问题出现的时候做比较。大部分的数据库管理员对他们的系统熟知而且能轻易的判断出系统使用的高峰时段。例如：高峰时段可能是上午10：00到下午12:00，也可能是

下午1.30到3.00.这可能包括凌晨12:00到6:00的批处理窗口。

It is important to identify these peakperiods at the site and install a monitoring tool that gathers performance datafor those high-load times. Optimally, data gathering should be configured fromwhen the application is in its initial trial phase during the QA cycle.Otherwise, this should be configured when the system is first in production.

找到这些这些高峰时段和安装一个能够收集高负载时段性能数据的监控工具都是非常重要的。理想情况下，数据采集应该是从应用的质量保证生命周期中的初始化试验阶段开始配置。否则，应该在系统首次应用于生产时配置。

Ideally, baseline data gathered shouldinclude the following:

理想情况下，基准线数据应该包括以下内容：

• Application statistics (transaction volumes, response time)
• 应用程序统计信息（交易量，响应时间）
• Database statistics
• 数据库统计信息
• Operating system statistics
• 操作系统统计信息
• Disk I/O statistics
• 磁盘 输入输出 统计信息
• Network statistics
• 网络统计信息

In the Automatic Workload Repository,baselines are identified by a range of snapshots that are preserved for futurecomparisons. See "Overview of the Automatic Workload Repository".

在自动工作量知识库里面，基线由一系列的快照形成，这些快照会被保存以便日后比较。祥见“自动工作负载库概述”。

1.1.2.3症状和问题

A common pitfall in performance tuning is tomistake the symptoms of a problem for the actual problem itself. It isimportant to recognize that many performance statistics indicate the symptoms,and that identifying the symptom is not sufficient data to implement a remedy.For example:

性能调优的一个常见陷阱是错误的把问题的症状（表现）当成是问题的实质了。认识到大量性能统计信息预示着问题的症状是很重要的。当然，同时认识到症状并不能给问题补救提供充分的数据也非常重要。

• Slow physical I/O
• 缓慢的物理输入输出

Generally, this is caused bypoorly-configured disks. However, it could also be caused by a significantamount of unnecessary physical I/O on those disks issued by poorly-tuned SQL.

通常情况下，这是由较次配置的磁盘引起的。然而他也可能由大量的不必要的磁盘物理输入输出引起，而这些输入输出是由垃圾sql导致。

• Latch contention
• 闩锁争用

Rarely is latch contention tunable byreconfiguring the instance. Rather, latch contention usually is resolvedthrough application changes.

通过重新配置实例来实现闩锁争用问题可调是很少的，相反闩锁争用通常通过改变应用程序来解决。

• Excessive CPU usage
• CPU使用率过高

Excessive CPU usage usually means that thereis little idle CPU on the system. This could be caused by an inadequately-sizedsystem, by untuned SQL statements, or by inefficient application programs.

过高的cpu使用率通常意味着系统会有很少空闲的cpu。这可能由系统规模不适当 或未经优化的sql 或者 低效率的应用程序导致。

1.1.2.4何时调优

There are two distinct types of tuning:

两种不同方式的调优

• Proactive Monitoring
• 主动监测
• Bottleneck Elimination
• 消除瓶颈

1.1.2.4.1主动监测

Proactive monitoring usually occurs on aregularly scheduled interval, where a number of performance statistics areexamined to identify whether the system behavior and resource usage haschanged. Proactive monitoring can also be considered as proactive tuning.

主动监测通常发生在一个定期的计划间隔，那里，一定数量的性能统计数据会被检测来确定系统的运行和资源利用率是否已经被改变。主动监测也可以被看作是主动调优。

Usually, monitoring does not result inconfiguration changes to the system, unless the monitoring exposes a seriousproblem that is developing. In some situations, experienced performanceengineers can identify potential problems through statistics alone, althoughaccompanying performance degradation is usual.

Experimenting with or tweaking a system whenthere is no apparent performance degradation as a proactive action can be adangerous activity, resulting in unnecessary performance drops. Tweaking asystem should be considered reactive tuning, and the steps for reactive tuningshould be followed.

Monitoring is usually part of a largercapacity planning exercise, where resource consumption is examined to seechanges in the way the application is being used, and the way the applicationis using the database and host resources.

通常，监控并不改变系统的配置，除非他暴露出了正在发生的一些严重的问题。在一些情况下，有经验的性能工程师能够只通过统计信息解决潜在的问题，尽管伴随着性能的下降。当没有显著的性能下降时，主动调整和试验系统是危险的，会造成系统不必要的性能下降。调整系统需要考虑反应调谐，而且反应调优的步骤要被跟踪。

监测通常是更大规模的规划工作中的一部分,那部分里面，资源消耗被检测来观测应用被使用的方式和 数据库及 主机资源的改变。

1.1.2.4.2瓶颈消除

Tuning usually implies fixing a performanceproblem. However, tuning should be part of the life cycle of an application—through the analysis, design, coding,production, and maintenance stages. Oftentimes, the tuning phase is left untilthe system is in production. At this time, tuning becomes a reactivefire-fighting exercise, where the most important bottleneck is identified andfixed.

Usually, the purpose for tuning is to reduceresource consumption or to reduce the elapsed time for an operation tocomplete. Either way, the goal is to improve the effective use of a particularresource. In general, performance problems are caused by the over-use of aparticular resource. That resource is the bottleneck in the system. There are anumber of distinct phases in identifying the bottleneck and the potentialfixes. These are discussed in the sections that follow.

调优通常意味着修复性能问题。然而调优应该是应用程序整个生命周期的一部分，包括分析，设计，编码，生产，维护阶段。经常的情况是调优阶段会一直保留到到系统生产运行才离开。那个时候调优已经变成响应的工作，大多数重要的瓶颈已经被识别和修复。

通常情况下，调优的目的是为了减少资源消耗或者减少某项操作完成的耗时。无论哪种方式，目的都是为了提高某个特殊资源的有效利用。一般来说，性能问题都是由某个特殊资源过度使用造成的，那个资源正好是系统的瓶颈。辨识出瓶颈和潜在的问题有许多不同的阶段。这些东西将在以下部分被讨论：

Remember that the different forms ofcontention are symptoms that can be fixed by making changes in the followingplaces:

请记住竞争的不同形式是一些症状，这些症状能够被在以下地方做些改变来修复、

• Changes in the application, or the way the application is used
• 应用程序的改变，或者应用程序被使用方式的改变
• Changes in Oracle
• oracle的改变
• Changes in the host hardware configuration
• 主机硬件配置的改变。

Often, the most effective way of resolving abottleneck is to change the application.

经常情况下，解决瓶颈最有效地方式是改变一个应用程序。

1.1.3 SQL 优化

Part IV, "Optimizing SQL Statements" ofthis guide discusses the process of tuning and optimizing SQL statements.

本节的IV部分 讨论到了调优的过程 和sql的优化。

Many client/server application programmersconsider SQL a messaging language, because queries are issued and data isreturned. However, client tools often generate inefficient SQL statements.Therefore, a good understanding of the database SQL processing engine isnecessary for writing optimal SQL. This is especially true for high transactionprocessing systems.

许多客户端和服务端的应用程序开发者把sql当做是一种消息语言，因为查询发起后就会返回结果数据。然而客户端工具常常会产出一些低效率的sql。因此，对数据库sql处理工作机制有个很好的理解对写出最优的sql是必要的。这对高事物处理系统尤其管用。

Typically, SQL statements issued by OLTPapplications operate on relatively few rows at a time. If an index can point tothe exact rows that are required, then Oracle can construct an accurate plan toaccess those rows efficiently through the shortest possible path. In decisionsupport system (DSS) environments, selectivity is less important, because theyoften access most of a table's rows. In such situations, full table scans arecommon, and indexes are not even used. This book is primarily focussed onOLTP-type applications. For detailed information on DSS and mixed environments,see the Oracle Database Data Warehousing Guide.

通常情况下，由联机事务处理系统发起的sql语句在同一时间会操作相对少的行。如果一个索引能够定位它所需要的准确行，那么oracle能够通过可能的最短的路径来建立一个准确的计划来取到那些行。在决策支撑系统环境当中，查询是次要的，因为他们总是要获取表中的大部分行。。在这个情况下全表扫描是普遍的，索引甚至可能都没被用到。这里主要的关注点都在联机事务处理系统。若要获得决策支撑系统和混合环境更详细的信息，请看 oracle数据库数据仓库指南。

1.1.3.1查询优化器和执行计划

When a SQL statement is executed on an Oracledatabase, the Oracle query optimizer determines the most efficient executionplan after considering many factors related to the objects referenced and theconditions specified in the query. This determination is an important step inthe processing of any SQL statement and can greatly affect execution time.

当一个sql语句在oracle数据库上执行的时候，在考虑过涉及到的若干对象的许多相关因素和查询时的一些具体情况后oracle查询优化器会决定哪个是最高效的执行计划。这个决定是任何sql语句执行过程的重要一步，而且对执行时间影响很大。

During the evaluation process, the queryoptimizer reviews statistics gathered on the system to determine the best dataaccess path and other considerations. You can override the execution plan ofthe query optimizer with hints inserted in SQL statement.

在评估期间，查询优化器会检查系统收集的统计信息来决定最好的取数据途径和其他考虑。你可以在sql语句中插入hint提示来覆盖查询优化器的执行计划。