SQL Server Optimization for Developers

Download Report

Transcript SQL Server Optimization for Developers

SQL Server
Optimization for
Developers
ANIL DESAI ([email protected] | HTTP://ANILDESAI.NET)
AUSTIN .NET USER’S GROUP, 04/14/2014
Presentation Overview

Database Performance Goals and Challenges

Monitoring and Optimizing Performance



Understanding indexes

SQL Profiler and Database Engine Tuning Advisor
Tuning Database Queries

Understanding the Query Optimizer and Execution Plans

Seeing the effects of indexes
Application Design Best Practices
Performance Monitoring Process

Establish a
baseline
Repeat
(if desired)
Measure
performance
Best Practices:

Optimize for
real-world
workloads

Monitor/review
performance
regularly

Focus on
specific issues
Identify
bottlenecks
Make one
change at a
time
Monitoring and Troubleshooting
Scenarios
Server-Level
Issues
• Users are reporting slow performance
• Intermittent transaction timeouts
• “The server/application seem
sluggish”
ApplicationSpecific
Issues
• “Application A is running more slowly
than usual”
• “The End-of-Month report is taking too
long to run.”
Other Issues
• Ad-hoc reports are running slowly
• CPU, memory, disk, or network alerts
are being generated
Common Datbaase Questions
Which queries are
taking the longest to
run?
How can I optimize a
specific query?
Which are the most
cost-effective
upgrades I can make
to improve
performance?
How will my
development updates
affect production
performance?
Which queries are
using the most system
resources?
Why does
database/web site
access slow down
during particular times
or during certain
patterns of activity?
Is indexing configured
optimally for my live
(production) database
workloads?
Database Performance Tools
System/OS
Performance
Monitor
SQL Server
SQL Server
Management Studio
QueryLevel
Database Engine
Tuning Advisor
SQL Profiler / SQL
Trace
Resource Monitor
Database Engine
Tuning Advisor
Task Manager
Dynamic
Management Views
(DMVs)
Query Execution
Plans
Database Design Issues


Transaction processing (OLTP)

Favors normalized schema

Many tables, each with fewer columns

Optimized for write (transactional) activity
Reporting and Analysis

Centralized, consistent storage of required data

Favored by denormalized schema

Fewer tables with many columns in each

Data is aggregated from multiple sources into a data mart or data
warehouse

May store aggregates in warehouse
Understanding Indexes




Index types

Clustered Index

Non-Clustered Indexes

Columnstore indexes
Indexing strategies

Goal is ideal index coverage

Index maintenance can slow-down write operations (Insert, Update, Delete)
Referential Integrity

Primary Key (default = clustered index)

Foreign Key references

Constraints
Statistics (manual vs. automatic)
General Index Tuning Best Practices

Make tuning a part of your development process

Dev: Use synthetic workloads and test cases

Test: Use real-world databases, whenever possible

Production: Capture real usage statistics for analysis

Collect a representative workload, whenever possible

Consider all applications and workloads when tuning a database

Use naming conventions for indexes and related objects

Use query hints sparingly (NOLOCK)
Using SQL Profiler


Purpose / Features:

GUI for managing SQL Trace

Monitor important events

Capture performance data / resource usage

Replaying of workloads / transactions

Identifying performance bottlenecks

Correlation of data with System Monitor

Workloads for Database Tuning Advisor
Examples:

Generate a list of the 100 slowest queries

Monitor all failed logins (Security)
Database Engine Tuning Advisor

Automatic workload
analysis for Physical Design
Structures (PDS)


Data Source:

File (Profiler Trace or .SQL
files)

Table

Plan Cache
Tuning Options

Keep existing PDS

Advanced Features:
Partitioning, indexed views,
etc.
Demo: Optimizing Indexes



Generate sample queries / tables

View query execution plans

View the effects of indexes on common queries
Capture Performance Data with SQL Profiler

SQL Profiler traces, events, and filters

Using SQL Load Generator to generate database load

Capturing and storing
Analyzing and optimizing with Database Tuning Advisor

Analyzing index usage reports

Saving and applying index recommendations
Tuning Individual Queries


Query Analyzer Features

Execution Plan (estimated and actual)

Include Client Statistics (multiple trials)

Analyze in Database Engine Tuning Advisor (single query)

Trace query in SQL Profiler (single query)
Keep query logic close to the database

Filter returned data at the database layer

Minimize the size of result sets

Minimize round-trips to the server

Use standard (inner) joins, where possible

Consider strategic denormalization for core sets of data
Query Optimizer Details

Goal: Find the most efficient method to return the data

Come up with a plan quickly

Minimize CPU, memory, and I/O requirements

Use statistics and index details to improve plans

Query plan caching

Relational engine vs. storage engine

Execution Plan output


Save as .sqlplan file for later analysis

Output in graphical, text, and XML formats

Can store and export plans using SQL Profiler (ShowPlan XML event)
Can use query hints
Understanding Execution Plans

Optimizing individual queries

Rewrite query logic

Use other objects (views, stored, procedures, etc.)

Strategic demoralization

Data Retrieval: Table scan, index seek/scan

Index Usage

Covering indexes

Join conditions
Execution Plan Example
Execution Plans in Windows Azure
Client Statistics Example
Application Design Best Practices



Create an abstraction layer between business and database
objects

ADO.NET

Microsoft Enterprise Library

NHibernate

Entity Framework
Use caching wherever possible

Server-side (web services)

Application-level (middle tier)

Client-side (desktop or mobile apps)
Minimize transaction times
Dev Best Practices: Application
and Data Architecture
Architecture
•Standards-based communications
•Loosely-coupled tiers and components
Development /
Testing
•Test harness
•Unit tests that use data
•Query performance monitoring
Performance
and Monitoring
•Windows PerfMon Counters
•Instrumentation and Logging
•Alerts / pro-active corrections
•Auto-scaling
Reliability
•Error and event handling
•Transaction retry (random interval); Deadlock management
•User notifications (responsive UI / cancel and retry options)
Windows Azure and Cloud
Databases



Practical cloud benefits

Data redundancy and geographic distribution

Lower management overhead
Potential issues

Keeping data close to applications and services

Data synchronization

Network performance issues

Data security, legal issues, and regulatory compliance
Determine where/how to use cloud-based services

SaaS vs. PaaS vs. IaaS
Azure Database Services


SQL Azure Database

Cost-effective, managed database instances

Can be managed with standard tools (Visual Studio and SSMS)

Some limitations (CLR, Mirroring, Partitioning, Replication, Extended SP’s)
Other Services

Azure Virtual Machines (SQL Server templates)

Azure Web Sites (with gallery templates)

Azure HDInsight, Cache Service

Azure Backup and Recovery Manager

SQL Azure Reporting

Network, Active Directory, Service Bus, etc.
Managing SQL Azure Instances
ORM Considerations


General issues

Development efficiency vs. hardware/software efficiency

Latency, query inefficiency (outer joins), platform-specific optimizations

Frequency and number of server round-trips
ORM-generated queries can be inefficient


Difficult to tune or modify individual queries
Potential Solutions

Make sure entity relationships are correct

Can use views or stored procedures to improve performance in some
cases

Bypass the ORM for some types of operations
New Features in SQL Server 2014

Memory-optimized tables (In-Memory OLTP)

Buffer Pool Extension (for SSD usage)

Delayed durability

Async log writes can result in data loss

Enable at database-level; use with BEGIN ATOMIC … COMMIT

Resource Governor storage I/O limits

Updateable Clustered ColumnStore indexes

Primarily for data warehousing; supports data index compression

Azure storage for SQL Server data/log files

Backup to Azure; Backup encryption
Dev Best Practices: Managing Data

Large UPDATE or DELETE operations:



Large INSERT operations

Disable indexes and triggers (if present)

Use BULK INSERT, bcp, SSIS, or DTS

Change transaction isolation level (if appropriate)

Change recovery model
Use SQL to generate SQL


Use loops to minimize locking and transaction log growth
Example: INSERT statements
Schedule or delay non-critical operations
Dev Best Practices: Schema
Changes

Generate Scripts

Script specific objects using SQL Server Management Studio

Script the entire database using Generate Scripts


Can include schema and/or data
Schema changes

Use ALTER commands when possible

Drop and recreate objects, as needed

Make all scripts re-runnable

Check before and after state of all objects
Dev Best Practices: Performance
Testing


Build performance testing/optimization into the dev process

Develop load tests or test “harnesses”

Using synthetic load generation tools

Use representative test data
Consider caching effects:

Index maintenance (fragmentation)

DBCC DropCleanBuffers

DBCC FreeProcCache
Advanced Performance
Approaches

Database Federations


Vertical and horizontal data partitioning
Cross-Server queries

Use Linked Servers to query across databases

Potential performance issues

Data compression (row- or page-level)

Resource governor

SQL Server Analysis Services (SSAS)

Pre-aggregation for performance

Dependent on a denormalized schema (optimized for reporting)
Links and References

Presenter:

http://AnilDesai.net | [email protected]

Presentation slides and sample code
Microsoft TechNet Virtual Labs
 Sample Databases



AdventureWorks Sample Databases (CodePlex)

Microsoft Contoso BI Demo Dataset
Database-related tools

SQL Load Generator by David Darden (CodePlex)

Glimpse

Red Gate Software

Spotlight
Summary and
Conclusion