1. Tuna Helper Proven Process for SQL Tuning Janis Griffin Senior DBA, Confio Software1

2. Who Am I? Senior DBA for Confio Software
Twitter
Current – 20+ Years in SQL Server & Oracle
DBA and Developer
Specialize in Performance Tuning
Review Performance of 100’s of Databases for Customers and Prospects
Common Thread – Paralyzed by Tuning

3. Agenda Introduction Challenges Identify - Which SQL and Why
Gather – Details about SQL
Tune – Case Study
Monitor – Make sure it stays tuned

4. Tuna Helper – Proven Process for SQL Tuning
Give a man a fish and you feed him for a day.
Teach a man to fish and you feed him for a lifetime.
Chinese Proverb

5. Introduction SQL Tuning is Hard This Presentation is an Introduction
3-5 day detailed classes are typical
Providing a Framework
Helps develop your own processes
There is no magic tool
Tools cannot reliably tune SQL statements
Tuning requires the involvement of you and other technical and functional members of team

6. Challenges SQL Tuning is Hard – did I mention that?
Requires Expertise in Many Areas
Technical – Plan, Data Access, SQL Design
Business – What is the Purpose of SQL?
Tuning Takes Time
Large Number of SQL Statements
Each Statement is Different
Low Priority in Some Companies
Vendor Applications
Focus on Hardware or System Issues
Never Ending

7. Identify – Which SQL Highest Wait Times (Wait Types)
Tracing a Session / Process
User / Batch Job Complaints
Queries Performing Most I/O (LIO, PIO)
Queries Consuming CPU
Queries Doing Full Table Scans
Known Poorly Performing SQL

8. Identify – End-to-End Understand the business aspects of statement
Who registered yesterday for SQL Tuning
Why does the business need to know this
How often is the information needed
Who uses this information
Understand the technical aspects
Review ERD
Understand tables and the data (at a high level)
Understand application architecture
Understand the entire process
What portion of the total time is database
Where is it called from in the application

9. Identify – End-to-End Time9

10. Execution Model (cont)
CPU
SPID 60 – Running
(Needs to perform IO)
SPID 51 - Running
Waiter List
SPID 52 – OLEDB
SPID 53 – WRITELOG
SPID 54 – PAGEIOLATCH_SH
SPID 57 – LCK_M_X
SPID 59 – WAITFOR
SPID 60 – PAGEIOLATCH_SH
CPU Queue
SPID 51 – Runnable
SPID 61 – Runnable
SPID 59 – Runnable

11. Wait Time Tables (SQL 2000) http://support.microsoft.com/kb/822101
sysprocesses
loginame
hostname
programname
spid
dbid
waittype
waittime
lastwaittype
waitresource
sql_handle
stmt_start
stmt_end
cmd
sysdatabases
dbid
name
::fn_get_sql
text

12. Wait Time Tables (SQL 2005/8)
sysprocesses
loginame
hostname
programname
spid
dbid
waittype
waittime
Lastwaittype
waitresource
sql_handle
stmt_start
stmt_end
cmd
dm_exec_sessions
login_time
login_name
host_name
program_name
session_id
dm_exec_requests
start_time
status
sql_handle
plan_handle
start/stop offset
database_id
user_id
blocking_session
wait_type
wait_time
dm_os_wait_stats
wait_type
waiting_tasks_count
wait_time_ms
dm_exec_query_stats
execution_count
total_logical_writes
total_physical_reads
total_logical_reads
dm_exec_query_plan
query_plan
dm_exec_sql_text
text

13. Problems with DMVs Can give an overall performance view
Can find costly queries and other problems.
Contents are from instance startup
No way to tie some information to a timeframe.
Are these Query stats from today or last week?
Need to sample data periodically
Be able to go back to 3:00 pm yesterday when problem occurred.
How often and when does this query execute?

14. Measure Database Wait Time
Understand the total time a Query spends in Database
Measure time while Query executes
SQL Server helps by providing Wait Types 14

15. SQLs – Wait Types - Wait Time
Almost 10 hrs on CPU
3.5 hrs on CXPACKET
Over 6.5 hours of Total time

16. Wait Time Scenario Which scenario is worse? SQL Statement 1
Executed 1000 times
Caused 10 minutes of wait time for end user
Waited 90% of time on “PAGEIOLATCH_SH”
SQL Statement 2
Executed 1 time
Waited 90% on “LCK_M_X”

17. Identify – Simplification
Break Down SQL Into Simplest Forms
Complex SQL becomes multiple SQL
Sub-Queries Should be Tuned Separately
Tuned SQL in Stored Procedures Separately
Get the definition of views
Understand Distributed Queries

18. Identify – Summary Determine the SQL Understand End-to-End
Measure Wait Time
Simplify Statement

19. Gather - Metrics Get baseline metrics Collect Wait Type Information
How long does it take now
What is acceptable (10 sec, 2 min, 1 hour)
Collect Wait Type Information
Locking / Blocking (LCK)
I/O problem (PAGEIOLATCH)
Latch contention (LATCH)
Network slowdown (NETWORK)
May be multiple issues
All have different resolutions
Document everything in simple language

20. Gather – Execution Plan
SQL Server Management Studio
Estimated Execution Plan - can be wrong for many reasons
Actual Execution Plan – must execute query, can be dangerous in production and also wrong in test
SQL Server Profiler Tracing
Event to collect:
Performance: Execution Plan (2000) , Showplan XML (2005/8)
Works when you know a problem will occur
DM_EXEC_QUERY_PLAN
Real execution plan of executed query

21. DM_EXEC_QUERY_PLAN

22. Example SQL Statement Who registered yesterday for SQL Tuning
SELECT s.fname, s.lname, r.signup_date
FROM student s
INNER JOIN registration r ON
s.student_id = r.student_id
INNER JOIN class c ON
r.class_id = c.class_id
WHERE c.name = 'SQL TUNING'
AND r.signup_date BETWEEN ' :00'
AND ' :59'
AND r.cancelled = 'N'
Execution Time – 1:30 to execute
Wait Types – Waits 90% on PAGEIOLATCH_SH

23. Execution Plan
CREATE NONCLUSTERED INDEX [reg_alt] on [dbo].[REGISTRATION] ([CANCELLED],[SIGNUP_DATE]) 23

24. Database Tuning Advisor
2 Indexes & 4 Additional Statistics

25. Gather - Relationships
CLASS
class_id
name
class_level
REGISTRATION
class_id
student_id
signup_date
cancelled
STUDENT
student_id
fname
lname

26. Gather – Table Information
Table Definition
Where does it physically reside
Large columns?
Data Profile Task / Viewer – Integration Services
Existing Indexes
Names of all existing indexes
Columns those indexes contain
Statistic Gathering & Rebuild Schedules

27. Gather – Summary Metrics Plan Table Relationships Table Information
How long does it take currently
What does the query wait for (wait types)
Plan
DM_EXEC_QUERY_PLAN
Actual Execution Plan
Do not use Estimated Plans unless necessary
Table Relationships
Table Information
Columns, Existing Indexes & Triggers

28. Tune – Create SQL Diagram
SQL Tuning – Dan Tow
Great book that teaches SQL Diagramming
registration
.04 5 30 1 1
student
class
.002
select count(1) from registration where cancelled = 'N'
and signup_date between ' :00' and ' :59'
3562 / = .0445
select count(1) from class where name = 'SQL TUNING'
2 / 1000 = .002

29. Tune – New Execution Plan
create index IX_CLASS on CLASS(name)
Metric – Takes 0:20 to execute
Why would an Index Scan still occur on REGISTRATION?

30. Gather – Existing Indexes

31. Tune – New Execution Plan
create index IX_REGISTRATION on REGISTRATION(class_id)
Metric – Takes 0:02 to execute

32. Tune – Alternative from SSMS
create index IX_ALT_REG on registration(cancelled, signup_date)
Metric – Takes 0:08 to execute

33. Query 2 Who cancelled classes within the week
SELECT s.lname, c.name, r.signup_date cancel_date
FROM student s
INNER JOIN registration r ON
s.student_id = r.student_id
INNER JOIN class c ON
r.class_id = c.class_id
WHERE r.signup_date BETWEEN ' :00'
AND ' :59'
AND r.cancelled = 'Y'

34. Tune – Execution Plan

35. Tune – Create SQL Diagram
registration
.004 5 30 1 1
student
class
select count(1) from registration where cancelled = 'Y' and
signup_date between ' :00' and ' :59'
378 / = .0047
select count(1) from registration
where signup_date between ' :00' and ' :59'
28797 / = .3599
select count(1) from registration where cancelled = 'Y'
523 / = .0065

36. Tune – New Execution Plan
create index IX_ALT_REG on registration(cancelled, signup_date)

37. Better Execution Plan Execution Stats – 20 Logical Reads
CREATE INDEX reg_alt ON registration(class_id)
INCLUDE (signup_date, cancelled)
Execution Stats – 20 Logical Reads
Metric – Takes 0:00 (sub-second) to execute 37

38. Monitor Monitor the improvement Monitor for next tuning opportunity
Be able to prove that tuning made a difference
Take new metric measurements
Compare them to initial readings
Brag about the improvements – no one else will
Monitor for next tuning opportunity
Tuning is iterative
There is always room for improvement
Make sure you tune things that make a difference
Shameless Product Pitch - Ignite

39. Summary Identify Gather Tune Monitor What is the Bottleneck
End-to-End view of performance
Simplify
Gather
Metrics – Current Performance
Wait Time
Execution Plan
Object Definitions and Statistics
Tune
SQL Diagrams – Dan Tow
Monitor
New Metrics, Wait Time Profile, Execution Plan

40. Confio Software Wait-Based Performance Tools Igniter Suite
Ignite for SQL Server, Oracle, DB2, Sybase
Provides Help With
Identify
Gather
Monitor
Based in Colorado, worldwide customers
Free trial at