wwmk230531p00013 Jul5

DBA Masterclass
Enable Ultimate Performance with

Automatic Database Tuning
Introducing Today’s Speaker & Panelists
Björn Bolltoft
Product Manager - Oracle
Observability & Management
Vladimir Demkin Alaa Fahmy Osama Ahmad

EMEA Core/Converged EMEA Core/Converged EMEA Data Management
Database Expert Database Expert Business Development Leader
Copyright © 2023, Oracle and/or its affiliates

Agenda
➢ Introduction
➢ Automatic Database Tuning
➢ Hands-on Lab Overview
➢ Q&A

Events Update
Next DBA Sep

Recording Available Masterclass Session 2023
DBA Masterclass
Session #1: Recording
https://youtu.be/FBSgfNS7654
Session #2:
https://youtu.be/YUYGdpSNzDg
https://oracle.com/goto/uk/dba-masterclass

Go to LiveLabs to view today’s workshop
1. Open up a web browser

2. Enter the link:
https://bit.ly/dba-masterclass-auto-tune
3. Run the workshop on your own
tenancy
OR on our LiveLabs Sandbox

Staying Informed…
https://bit.ly/emea_oracle_dba

Database 23c Page – New Features Information Updated Regularly
https://www.oracle.com/database/23c/
DBA Masterclass - Enable Ultimate
Performance with Automatic Database
Tuning
Fool-proof Approach for Managing Database Performance
Björn Bolltoft
Consulting Product Manager
12 Copyright © 2023, Oracle and/or its affiliates

Agenda
• Find, Fix € Validate
• The Find
• Find : Theory to Practice
• The Fix
• The Validate

FIND, FIX & VALIDATE
Top Challenges
Database Management
Handle more than

39%
? 50 DBs each
Key Takeaway:
Improve & Ensure Higher Quality of Service
Downtime resulting
78%
from untested changes
15 Copyright © 2023, Oracle and/or its affiliates Source: From Database Clouds to Big Data: IOUG Survey on Database Manageability
Top Challenges
Database Management
?
39% Automate
Handle more than
50 DBs each
Key Takeaway:
Improve & Ensure Higher Quality of Service
Downtime resulting
78% Validate
from untested changes
16 Copyright © 2023, Oracle and/or its affiliates Source: From Database Clouds to Big Data: IOUG Survey on Database Manageability
Past DB Tuning Process
Inefficient, Inaccurate, Manual
Find a Implement
Default value is 100,
problematic a possible
metric Fix …so let’s set it to 50…
Will optimizer_index_cost_adj
improving my sequential reads…
Wait for
end-user
Did it work? feedback Are customers complaining?
The “trial and error” method can consume more than 50% of the DBA time
Future Generation DB Tuning Process
Effective, Accurate, Automated
Built-in, self-diagnostics Automates complex and

engine: Automatic Find
Diagnostics Fix
Tuning
time consuming task of
Database Diagnostics application tuning
Monitor (ADDM) – (Tuning Pack)
(Diagnostics Pack)
SQL
Performance Validates tuning activities
Validate
Analyzer (Real Application Testing)
(SPA)

Future Generation DB Tuning Process
Effective, Accurate, Automated
STEP 1 - Find Issues STEP 2 - Fix Issues

Tuning
Diagnostics
SQL Performance
Analyzer (SPA)
STEP 3 - Validate Fix

Oracle’s Performance Methodology
• “DB Time” based

• Identify components/operations consuming most DB Time
– Oracle kernel is instrumented to track DB Time in every component of database
– Ignores Idle Wait Events
• In order to tune system, reduce “DB Time” consumed for the same workload
• Scoped analysis possible
– Instance wide
– Targeted for session/process or SQL, etc.
– Single tool usage results in fewer clicks
• DB Time exposed through EM interface for easy drill-down and analysis
• In most cases, 2-3 Clicks required for problem resolution
– Single tool usage results in fewer clicks
• Documentation: 2-Day DBA Class

Database Time (DB Time)
• Total time in database calls by foreground sessions

• Includes CPU time, IO time and non-idle wait time
• DB Time <> response time
• New metric for Oracle performance analysis
Database time is total time spent by user processes either

actively working or actively waiting in a database call.

Fundamental Concepts
Active Session =
Session currently spending time in a database call
Database Time (DB Time) =
Total time session spent in all database calls
Average Activity of the Session (% Activity) =
The ratio of time active to total wall-clock time
Browse Read Reviews

Add to Cart Checkout
Books For One Book
= time spent in database TIME

Diagnostics
The Find
Automatic Workload Repository
Challenge Solution
• How does a DBA monitor and • AWR

be aware of problems as they • Benefit:
arise By automating the gathering
• Capture associated database of database statistics for
statistics for problem problem detection and tuning,
detection and tuning AWR serves as the foundation
for database self-
management.

AWR HUB & AWR Explorer
Challenge Solution
• Default retention period of 8 • AWR HUB

days prevents diagnosis of • Benefit:
long term performance The AWR HUB is set up with
problems the same objects as in a
• Compare performance during standard AWR schema of any
this quarter’s books close with Oracle database. The
last quarter’s enhancement lies in the
• Increasing AWR retention partitioning, that allows for
period increases storage quick loads, efficient querying
overhead and cost in critical and when requested, effective
production environments purging of unwanted data.

AWR performance comparisons from the past
Single report
• One datapoint
• Is it good or is it bad?
control file sequential read

160
140
120
100
80
60
40
20
0
Total(s)

Single report - several snapshots

Still one datapoint
Is it good or is it bad?
300 control file sequential read
250
200
150
100
50
0
Total(s)

AWR Compare period Report

• Is it good or is it bad?
– You need to know the history
• Is this expected?

170
165
160
155
150
145
140
135
130
125
1 2

Several Report - Analytics

• Trend analysis
• Excel or homegrown analytics
• We have the history
• Is this good or is it bad?
• Is it an expected behavior?
• How do we corelate?

400
300
200
100
0
0:00
3:00
6:00
9:00
12:00
15:00
18:00
21:00
0:00
3:00
6:00
9:00
12:00
15:00
18:00
21:00
0:00
3:00
6:00
9:00
12:00
15:00
18:00
21:00
0:00
3:00
6:00
9:00
12:00
15:00
18:00
21:00

AWR Explorer
AWR Explorer
Visualize AWR data for advanced diagnostics
• Enables DBAs to perform advanced performance

analysis
• Integration of Oracle performance tools plus data
visualization
• Provides different aspects of Oracle Database
performance data which can be very helpful in
issue detection
• Visualize performance trends easily without
needing to toggle between hourly AWR reports
• Generate AWR report, ASH report, Perfhub report,
SQL report, etc.

AWR Explorer Features
Top Activity
Performance trend of the database with
capabilities to filter by various ASH
dimensions:
• Top SQLs
• Top wait events
Zoom on time dimension

• 1 hour – 7 days

Metrics
System metric values from
DBA_HIST_SYSMETRIC_HISTORY
• Average active sessions
• CPU
• IO
• Network
• Logons.
• ….
Inherit time dimension

Load Profile
System statistics information in

• DBA_HIST_SYSSTAT
• DBA_HIST_SYS_TIME_MODEL
Provide an overview of the global health of the
database
• Time Model Overview,
• Connections,
• Logons

Wait events
Top 10 wait events sorted by wait time

• Waits per Second
• Average Wait Time
Add new charts to display more wait events
View a histogram of the wait event with a

break down of the percentage of waits by
their duration, within a selected snapshot
range

Load Profile
All the database parameters that were

changed within the snapshot range
Corresponding change history.

Active Session History
Challenge Solution
• DBA on a production system • Active Session History

and get an emergency call like • Benefit:
“The Database is dead slow!” Enables targeted performance
• Identify the root cause of analysis of transient problems
performance issues at a precise • Built into the Oracle kernel and
point in the past—even when highly optimized
the sessions have disconnected.
• Increasing AWR retention
period increases storage
overhead and cost in critical
production environments

Active Session History (ASH)
Each sample represents a second of session activity

Sum the seconds to compute DB Time
Average Activity of the Session (% Activity) =
DB Time/Elapsed Time
Browse Read Reviews

Add to Cart Checkout
Books For One Book
= time spent in database TIME

Active Session History (ASH)
Query for Khalid Browse and Add Checkout

Husaini Novels Read item to cart using
Reviews ‘one-click’
Time SID Module SQL ID State Event
7:38:26 213 Book by author qa324jffritcf WAITING db file sequential read
7:42:35 213 Get review id aferv5desfzs5 CPU
7:50:59 213 Add to cart hk32pekfcbdfr WAITING buffer busy wait
7:52:33 213 One click abngldf95f4de WAITING log file sync

User Session Performance Dimensions
SQL PL/SQL Resources Identifiers Attributes

SQL
Consumer
SQL ID PL/SQL Wait Class Instance
Group
Session Top Level

Plan Hash Wait Event Services Module
Attribut PL/SQL PL/SQL
es
ASH Operation Object User Session Action
Blocking Parallel
OpCode Client
Session Process
Session
Resourc
Identifie
es
rs Program Trans. ID
Session Type

Automatic Database Diagnostics Monitor (ADDM)
Challenge Solution
• Perform accurate and timely • ADDM

diagnosis of the problem • Benefit:
before making any changes to ADDM makes it possible for
a system the Oracle Database to
• Diagnose the root causes of diagnose its own performance
performance problems and determine how any
identified problems can be
resolved.
• Root cause analysis,
Correction recommendations
Impact and benefits analysis

Automatic Database Diagnostics Monitor (ADDM)
Continuous Evolution in Database Performance Management
Compare Real-Time Enhanced ADDM

ADDM Period ADDM ADDM Real-Time Spotlight
ADDM
• Diagnose persistent • In-depth • Hung or extremely • Proactively detect • Three tabs:
performance issues performance slow databases & diagnose Findings,
comparison across transient high- recommendations
• Uses AWR snapshots • Uses a normal and
two periods impact problems and DB
diagnostic mode
• Regular interval parameters
• Relies on AWR data connection • Built inside the DB
• Automatic / Manual • Findings ordered
• Manual • Manual • Automatically
by Overall impact,
runs every 3
recommendations
seconds
by Overall benefit
• Database
Parameter
provides flexible
opt-in filtering
ADDM Spotlight
Challenge:
Production system changes have the potential to disrupt availability and performance SLAs. DBAs need to weigh the
costs vs. benefits of making high-impact system changes
Solution:
Provide insights on what changes need to be implemented to improve performance by providing synopsis of ADDM
findings/ recommendations and their impact/benefit over a longer period
Benefit:
• Gives you confidence to implement recommendations
• Up-level junior staff with expert advice
• Up-level system optimization from tactical to strategic
Features covered:
• Event Annotations
• Findings, Recommendations & Database Parameters tabs

Findings, Recommendations & Database Parameters tabs
• Findings ordered by Overall impact,

recommendations by Overall benefit
• Recommendations grouped by
implementation domain:
• Database Parameters, SQL, Schema
Objects
• Top SQL finding has
recommendations by SQL_ID (can be
many)
• Database Parameter provides flexible opt-
in filtering:
• High impact & non-default

Compare Period ADDM
SQL Commonality
AWR Snapshot Regressed SQL

Period 1
I/O Bound
AWR Snapshot Undersized SGA

Compare Analysis Report
Period 2 Period ADDM
• Full ADDM analysis across two AWR snapshot periods

• Detects causes, measure effects, then correlates them
– Causes: workload changes, configuration changes
– Effects: regressed SQL, reach resource limits (CPU, I/O, memory, interconnect)
• Makes actionable recommendations along with quantified impact
FIND: Theory to Practice
DB Time Theory to Practice: EM Performance page
Active Sessions by wait class over

time
Colored area = Amount of DB time
, More the DB time , More the
Problem
“Click on the big stuff…”

DB Time Theory to Practice: EM Performance page
Active Sessions by wait class over

time
Colored area = Amount of DB time
, More the DB time , More the
Problem
“Click on the big stuff…”

ASH Theory to Practice: EM ASH Analytics Page
• Graphical ASH report for

advanced analysis
• Select any time period for
analysis
• Analyze performance across
many dimensions
• Provides visual filtering for
recursive drill-downs
• Different visualizations:
Stacked chart or Tree Map
• Collaborate with others using
Active Reports

ADDM to Practice: Automatic Diagnosis and Recommendations
• 54.1% of the impact is from the

SQL statements in the report
• Performance diagnostics data
provided for SQL causing high
DB resource usage

It’s now time for Q&A
Please use the Q&A Console

to submit questions
Tuning
The Fix
Automatic SQL Tuning
Gather Missing or Stale

Statistics
SQL Profiling Create a SQL Profile

Statistics Analysis Add Missing Access Structures
Access Path Analysis Modify SQL Constructs
SQL Restructure Analysis
Adopt Alternative
Alternative Plan Analysis Execution Plan
Parallel Query Analysis SQL Tuning Create Parallel SQL Profile Administrator
Advisor
Automatic Tuning Optimizer Comprehensive SQL

Tuning Recommendations
• SQL Tuning Advisor

• Gives suggestions on the various problems identified during the diagnosis phase
• Uses the same CBO but has more time budget to perform comprehensive analysis
• Identifies alternate execution plans using real-time and historical performance data
• Recommends parallel profile if it will improve SQL performance significantly (2x or more)

The Fix: ADDM tuning -> SQL Tuning Advisor
• Run the problematic

performing statements through
Tuning Advisor
• We will automatically capture
the problematic statement
• Top SQL statements are
automatically captured in AWR
• Seamless integration between
Diagnostics and Tuning

The Fix: SQL Tuning Advisor
• Gather Missing or Stale

Statistics
• Create a SQL Profile
• Add Missing Access Structures
• Modify SQL Constructs
• Adopt Alternative
Execution Plan
• Create Parallel SQL Profile

The Fix: Tuning Advisor – SQL Profile
• Comprehensive analysis and

multiple alternatives to
improve performance
• SQL Profiles can be
implemented transparently &
non-intrusively to tune SQL
statements

What is SQL Profile?
• Contains auxiliary information collected by the ATO for a SQL statement

• Customized optimizer settings
• Based on past execution history (e.g., first_rows vs. all_rows)
• Compensation for missing or stale statistics
• Compensation for errors in optimizer estimates
• Estimation errors occur due to data skews and correlations, complex filters and joins
• Doesn’t require any change to the SQL text
• Ideal for Packaged Apps
• Persistent: Works across shutdowns & upgrades
• Transportable across databases (10.2)
• Force Matching for Literals: This setting may be useful for queries that use only literal values
• SQLTUNE_CATEGORY: Enables you to test a profile in a restricted environment before
making it available to other sessions. Can be enforced at session level or system level, this
defaults to system level
What SQL Plan Baselines?
• A Plan Baseline is a set of accepted plans that the optimizer is allowed to

use for a SQL statement
• A plan includes all plan-related information that the optimizer needs to
reproduce an execution plan, such as)
• SQL plan identifier
• Set of hints
• Bind Values
• Optimizer environment
• The primary goal of SQL plan management is to prevent performance
regressions caused by plan changes

• SQL plan baselines are proactive but restricts your potential gain
• SQL profiles are reactive but can improve your performance

The Fix: Tuning Advisor – Optimizer Statistics
• Comprehensive analysis and

multiple alternatives to
improve performance
• SQL Profiles can be
implemented transparently &
non-intrusively to tune SQL
statements
• As an example, let’s gather
statistics in pending mode…

• The Fix – Gathering Optimizer Statistics
• However, this could negatively impact other SQL in the

workload
• If only there was a fool-proof method to validate statistics

gathering recommendation….

SQL
Performance
Analyzer (SPA)
The Validate
SPA Overview
• Helps users predict the impact of system

SQL Workload
changes on SQL workload response time STS
• Low overhead capture of SQL workload to SQL

Tuning Set (STS) on production system SQL plans + stats
SQL plans + stats
• Build different SQL trials (experiments) of SQL Pre-change Trial Post-change Trial
statements performance by test execution
• Analyzes performance differences Compare
SQL Performance
• Offers fine-grained performance analysis on Analysis Report

individual SQL
• Integrated with STS, SQL Plan Baselines, & SQL
Tuning Advisor to form an end-to-end solution

SPA Report Example
2 5
1
4

SPA Challenges
Running SPA on:
Test System: Safe but… Production System: Easier but…

Test Production
Database Database
• Requires separate HW • Could be resource intensive and
impact production performance
• Data in test system should be
same as production • Changes needs to be manually
scoped to private session
• Lengthy, error-prone task
• Could take a long time to finish
• No resource control by default
DBA

SPA Quick Check
• Optimized for use on prod systems SPA Quick Check

Optimized • Optimal Trial or Explain Plan mode
• Disable multi-executions, full DML execute disabled

Optimal Trial Per SQL Time
Mode, no DML Limits, Limits
execute Controlled/ testing scope to
Optimized Scoped private session
• Per SQL time limits

Context-aware
Controlled • Testing scoped to private session

Change-Aware change testing
• Associate with Resource Consumer Group

Production
Database
Pre-selected
STS and
default SPA
• Context-aware change testing workflows, such as, settings
Change-Aware • Optimizer gather statistics DBA
• Init.ora parameter changes

SPA Quick Check
Optimized
Minimizes use
Identifies subset Test-executes of production Multiple No full DML
SQL workload with only SQLs with resources executions (execute Select
plan changes first plan changes dramatically – up disabled part of workload)
to 10x reduction

SPA Quick Check
Change-aware
Change-aware: Knows what change is being tested
In-line with routine DBA tasks such as statistics

gathering, init.ora parameter changes
Intelligently limits impact to private test session

SPA Quick Check
Controlled
Per-SQL time limit – protects from runaway SQL
Resource throttling - Associate with

Resource Consumer Group
Testing scope limited to private session

Optimizer Statistics Refresh - Gather Statistics
Go to Optimizer Statistics Page


• Choose Gather Statistics


• Choose the extent
• Choose to validate with SPA


• Choose object according to
dialog


• Choose object according to
dialog
• Submit

Optimizer Statistics Refresh- Remedy and publish
• Remedy regression
• SQL Plan Baseline
• Tuning advisor

• Tuning advisor
• Implement (for this example)

• Tuning advisor
• Validate again

• Tuning advisor
• Validate again
• Publish Statistics

Use Case 2: SPA Quick Check - Init.ora Parameter Validation
Problem:
• My manager wants me to improve database performance by tuning the optimizer_index_caching
parameter. The assumption is that will help
improve the performance significantly…
• However, I’m not sure if this parameter will help, how do I quickly and accurately validate the impact of the
above parameter for my workload?
Solution: You can use SPA Quick Check to validate the performance impact of
the parameter…

SPA Quick Check - Init.ora Parameter Validation
• Select Administration >

Initialization Parameters
• Change Parameter of interest
• Category Optimizer
• Select Validate with SPA
• Select “SQL Performance
Analyzer Task”

• Select the latest

SPA_OPTPARAMETER… task


• Select comparison between
TIRD_TRIAL and
FOURTH_TRIAL

• Select comparison between
TIRD_TRIAL and
FOURTH_TRIAL
• Check the result
• If improvements, implement
• If regression and improvements
– Consider SQL Plan Base &
Implement
• If regression, discard
83 Copyright © 2023, Oracle and/or its affiliates | Confidential: Internal

Tuning
STEP 1 - Find Issues
Diagnostics STEP 2 - Fix Issues
SPA Quick
Check
STEP 3 - Validate Fix

It’s now time for Q&A
Please use the Q&A Console

to submit questions
Q&A
Learn More Oracle Cloud Free Tier
1. oracle.com/enterprisemanager
2. Video Series: youtube.com/OracleEnterpriseMgr Always Free

Services you can use for
3. Blog: https://blogs.oracle.com/post/top- unlimited time
activity-lite
https://blogs.oracle.com/observability/post/ad
dm-spotlight-strategic-advice-optimize-oracle-
dbms
4. Documentation:
https://docs.oracle.com/en/enterprise-
30-Day Free Trial
manager/cloud-control/enterprise-manager- Free credits you can use for
cloud-control/13.5/index.html more services
Hands-on-labs
5. Try it now
www.oracle.com/cloud/free

wwmk230531p00013 Jul5

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

wwmk230531p00013 Jul5

Uploaded by

Copyright:

Available Formats

DBA Masterclass

Enable Ultimate Performance with

Vladimir Demkin Alaa Fahmy Osama Ahmad

Copyright © 2023, Oracle and/or its affiliates

Copyright © 2022, Oracle and/or its affiliates

Next DBA Sep

Copyright © 2022, Oracle and/or its affiliates

1. Open up a web browser

Copyright © 2023, Oracle and/or its affiliates

Copyright © 2023, Oracle and/or its affiliates

12 Copyright © 2023, Oracle and/or its affiliates

13 Copyright © 2023, Oracle and/or its affiliates

Handle more than

Built-in, self-diagnostics Automates complex and

18 Copyright © 2023, Oracle and/or its affiliates

STEP 1 - Find Issues STEP 2 - Fix Issues

STEP 3 - Validate Fix

19 Copyright © 2023, Oracle and/or its affiliates

• “DB Time” based

20 Copyright © 2023, Oracle and/or its affiliates

• Total time in database calls by foreground sessions

Database time is total time spent by user processes either

21 Copyright © 2023, Oracle and/or its affiliates

Browse Read Reviews

= time spent in database TIME

• How does a DBA monitor and • AWR

24 Copyright © 2023, Oracle and/or its affiliates

• Default retention period of 8 • AWR HUB

25 Copyright © 2023, Oracle and/or its affiliates

control file sequential read

26 Copyright © 2023, Oracle and/or its affiliates

Single report - several snapshots

300 control file sequential read

27 Copyright © 2023, Oracle and/or its affiliates

AWR Compare period Report

control file sequential read

28 Copyright © 2023, Oracle and/or its affiliates

Several Report - Analytics

control file sequential read

29 Copyright © 2023, Oracle and/or its affiliates

• Enables DBAs to perform advanced performance

31 Copyright © 2023, Oracle and/or its affiliates

Zoom on time dimension

32 Copyright © 2023, Oracle and/or its affiliates

Inherit time dimension

33 Copyright © 2023, Oracle and/or its affiliates

System statistics information in

Inherit time dimension

34 Copyright © 2023, Oracle and/or its affiliates

Top 10 wait events sorted by wait time

Add new charts to display more wait events

Inherit time dimension

View a histogram of the wait event with a

35 Copyright © 2023, Oracle and/or its affiliates

All the database parameters that were

Corresponding change history.

36 Copyright © 2023, Oracle and/or its affiliates

• DBA on a production system • Active Session History

37 Copyright © 2023, Oracle and/or its affiliates

Each sample represents a second of session activity

Browse Read Reviews

= time spent in database TIME