Oracle Data Guard

Maximum Data Protection at

Minimum Cost
 Agenda

 HA / DR Challenges
 Data Guard Architecture & Features
 Data Guard 10g New Features
 Maximum Availability Architecture
 Summary
 Cost of Downtime
 Hotel Reservation System: $100K an hour
 Online Auction: $225K an hour
 Worldwide Consolidated Enterprise System: $1.2M an hour
 Credit Card Sales Authorization: $2.6M an hour

• Revenue
• Financial performance
• Productivity
• Damaged reputation
• Other expenses…
 Causes of Downtime
System
Failures
Storage Failures
Unplanned
Downtime Data
Human Errors

Failures Corruptions

Site Failures
System
Planned Changes
Downtime
Data
Changes
 Oracle’s Integrated HA Solutions

System Real Application Clusters

Failures
Unplanned Automatic Storage Management
Downtime Data
Flashback
RMAN & Flash Recovery Area
Failures H.A.R.D
Data Guard

System Online Reconfiguration

Planned Changes Rolling Upgrades
Downtime
Data Online Redefinition
Changes
 Oracle’s Integrated HA Solutions

System Real Application Clusters

Failures
Unplanned Automatic Storage Management
Downtime Data
Flashback
RMAN & Flash Recovery Area
Failures H.A.R.D
Data Guard

System Online Reconfiguration

Planned Changes Rolling Upgrades
Downtime
Data Online Redefinition
Changes
 Agenda

 HA / DR Challenges
 Data Guard Architecture & Features
 Data Guard 10g New Features
 Maximum Availability Architecture
 Summary
 What is Oracle Data Guard?
 Oracle’s disaster recovery solution for Oracle data

 Feature of Oracle Database Enterprise Edition (EE)

 Automates the creation and maintenance of one or more

transactionally consistent copies (standby) of the production (or
primary) database

 If the primary database becomes unavailable (disasters,

maintenance), a standby database can be activated and assume
the primary role

 Requires same releases of Oracle database and operating system

for the primary and standby servers
 Oracle Data Guard Focus

 Data Failures & Site Disasters:

– Data Protection
– Data Availability
– Data Recovery

Data is the core asset of

All 3 are important! the enterprise!

• Also addresses human errors & planned maintenances

 Oracle Data Guard Architecture
Physical Standby
Database
Sync or Async
Redo Shipping Backup
Production
Database Redo Apply DIGITAL DATA STORAGE

DIGITAL DATA STORAGE

Network
Broker

Logical Standby
Transform Database Open for
Redo to SQL Reports

SQL
Apply
Additional
Indexes & MVs
 Data Guard Process Architecture (9i)
Oracle Net Physical/Logical
Transactions Standby
Database

LGWR RFS MRP/ LSP

Online Redo Logs

FAL
Primary Standby
Transform Redo
Redo
Database Logs
to SQL for
SQL Apply
Backup /
ARCH Reports

Archived Redo Logs ARCH

Archived Redo Logs
 Data Guard Redo Apply

Primary Data Guard Physical Standby

Database Broker Database

Redo Apply Backup

Network DIGITAL DATA STORAGE

Redo Shipment
Standby
Redo Logs

 Physical Standby Database is a block-for-block copy of the primary database

 Uses the database recovery functionality to apply changes
 Can be opened in read-only mode for reporting/queries
 Can also be used for backups, offloading production database
 Data Guard SQL Apply
Additional
Indexes &
Materialized Views

Primary Logical Standby

Database Data Guard Database
Broker
Transform Redo
to SQL and Apply
Continuously
Network Open for Reports

Redo Shipment Standby

Redo Logs

 Logical Standby Database is an open, independent, active database

 Contains the same logical information (rows) as the production database
 Physical organization and structure can be very different
 Can host multiple schemas
 Can be queried for reports while logs are being applied via SQL
 Can create additional indexes and materialized views for better query performance
 Switchover and Failover

 Primary and Standby role transitions

 Switchover
– Planned role reversal
– No database reinstantiation required
– Used for maintenance of OS or hardware
 Failover
– Unplanned failure (e.g. disasters) of primary
– Primary database must be reinstantiated / flashed back [10g]
 Initiated using simple SQL / GUI interface
 Data Guard automates the processes involved
 Flexible Data Protection Modes
Protection Mode Data Loss Protection Redo Shipment

Maximum Protection Zero Data Loss LGWR SYNC – Synchronous

Double Failure Protection redo shipping to >= 2 sites

Maximum Availability Zero Data Loss LGWR SYNC – Synchronous

Single Failure Protection redo shipping

Maximum Performance Minimal data loss LGWR ASYNC or ARCH –

Asynchronous redo shipping

E.g. ALTER DATABASE SET STANDBY TO MAXIMIZE PROTECTION;

Balance cost, availability, performance, and transaction protection

 Automatic Resynchronization
 Network connectivity problems may occur
 Data Guard automatically resynchronizes standbys after
network connectivity restored
– Implicit
 ARCH process idling away on the primary ‘pings’ all standbys
on a regular basis to see if they are missing any redo data
 If so it sends them the missing redo data

– Explicit
 Gap discovered during apply process in physical standby
 Based on FAL_SERVER and FAL_CLIENT settings, primary
notified, and it sends missing redo data
 Agenda

 HA / DR Challenges
 Data Guard Architecture & Features
 Data Guard 10g New Features
 Maximum Availability Architecture
 Summary
 Data Guard 10g New Features
 General new features
– Real Time Apply
– Flashback Database Integration

 SQL Apply new features

– Zero Downtime Instantiation
– Support for additional datatypes
– Rolling Upgrades

 Data Guard Broker & Enterprise Manager new features

– RAC integration
– Simplified browser-based interface focused on best practices
 Real Time Apply
An up-to-date
Oracle Net Physical/Logical
Transactions
Standby
Database

LGWR RFS MRP/ LSP

Standby
Redo
Logs
Online
Redo Real Time
Logs
Primary Apply
ARCH
Database
ARCH

Archived
Redo Logs
Archived
Redo Logs
 Oracle 11g New Features

 Active Data Guard.

 Up to 30 DR sites (11gR2).
 Existing Site Recovery Tradeoffs
Primary Database Standby Database Reporting on
Redo delayed data
Shipment

Delayed
Apply

 Log apply may be delayed to protect from user errors but:

– Switchover/Failover gets delayed
– Reports run on old data
 After failing over to standby, production DB must be rebuilt
Enhanced DR with Flashback Database
Primary Database Real Time Standby Database
Real Time
Redo Apply Reporting
Shipment

No Delay!

Flashback Flashback
Log Log

Primary: No reinstantiation
after failover!

 Flashback DB removes the need to delay application of logs

 Flashback DB removes the need to reinstantiate primary after failover
 Real-time apply enables real-time reporting on standby (at extra cost of
resources!)
 SQL Apply Enhancements
 Logical standby databases can now be created from an
online backup of the primary database
 Support for additional datatypes:
– NCLOB
– LONG
– LONG RAW
– BINARY_FLOAT
– BINARY_DOUBLE
– IOT-s (without overflows and without LOB columns - mind Rolling
Upgrades)
SQL Apply – Rolling Database Upgrades
Upgrade

Redo Patch Set

Clients Upgrades
A B Logs A B
Queue

Major
Version X Version X X X+1 Release
1 2 Upgrades
Initial SQL Apply Config Upgrade node B to X+1

Redo Redo Cluster

Upgrade A B A B
Software &
Hardware
Upgrades

X+1 X+1 X X+1

4 Switchover to B, upgrade A 3 Run in mixed mode to test
Enterprise Manager New Features

 Broker & EM can now create and manage configurations

that contain RAC primary and RAC standby databases
 Streamlined browser-based interface that enables
complete standby database lifecycle management
 Focus on:
– Ease of use
– Management based on best practices
– Pre-built integration with other HA feature
 Example – Ease of Use

 Creation of Data Guard Standby using Grid

Control: just some mouse-clicks
(I like vanilla ksh-scripting, though)
 Example – Ease of Use

 Switchover using Enterprise Manager is now

literally two mouse clicks
Switched!
 Agenda

 HA / DR Challenges
 Data Guard Architecture & Features
 Data Guard 10g New Features
 Maximum Availability Architecture
 Summary
 Maximum Availability Architecture
 Oracle recommended High Availability architecture
– HA blueprints based on Oracle technologies:
 Database, Application Server, Enterprise Manager, …
– Evolves with new Oracle versions and features
– Based on HA requirements of customers
– Intended for Oracle customers interested in HA

 Best practices to define and meet SLA requirements

– Architectural, configuration and operational best practices
– Minimize scheduled and unscheduled downtime
– Efficiently recover from outages and restore fault tolerance
– Hardware and OS independent

Unbreakable Architecture + Best Practices = Maximum Availability

 MAA Best Practice Publications

 Best Practices on:

 RAC/ Data Guard configuration
 Redo data transport mechanisms
 Instance Recovery
 Switchover/Failover
 Media recovery
 White papers1:
 SQL Apply configuration  MAA – detailed
 Network configuration  Media Recovery
 Integration of HA technologies  Site/Network configuration
 Fast-Start Checkpointing
 SQL Apply Best Practices
 Role Management

1. Ref. http://otn.oracle.com/deploy/availability/htdocs/maa.htm for latest updates

 Some Data Guard MAA Results
 Data Guard synchronous transport is viable for LAN and even for
MAN (~450 km / 10 ms RTT)
– Zero data loss capability
– Minimal impact on primary system throughput (less than 3% in tests)

 Asynchronous transport is viable for all types of networks

– US coast-coast deployment (~4000 km / 80 ms RTT) has almost no
performance impact (less than 1% in tests) with an average transaction loss
potential of 1 second in the event of a disaster

 Using SSH compression over a WAN in the asynchronous mode

reduces the network traffic by 35-60%
 Switchover can be done in 1 minute, failover can be done as fast as
15 seconds
 Data Guard and Streams
 Streams and Data Guard are independent features of Oracle Database
Enterprise Edition, based on some common underlying technology

 Data Guard: Disaster Recovery & Data Protection

– Transactionally consistent standby databases
– Zero data loss
– Automated switchover/failover
– Various data protection modes

 Streams: Information Sharing/Distribution

– Fine granularity and control over what is replicated
– Bi-directional replication
– Data transformations
– Heterogeneous platforms
Financial Services Company Using Data Guard & Streams

Streams
Master for information
Database distribution

Data Feed

Data Guard Data Transformation

for DR

Product Delivery Databases

for Client Access

Physical Standby Database

Disaster Recovery with Remote Mirroring

 Remote Mirroring (host-based and storage-based) is

another popular way to protect enterprise data

 Pro’s
– 1 solution for all datatypes
 Con’s
– Bandwidth (up to 30x)
– No protection from human failure (DML/DD)
– Cost of hard- and software
 Data Guard Advantages

1. Network Efficiency 4. Functionality

– Transmits only redo data – Comprehensive DR solution

2. Better suited for WANs 5. Flexibility

– No protocol converters, – No vendor lock-in for underlying
based on standard TCP/IP storage

3. Better Data Protection 6. ROI

– Preserves write-order – Extract value out of DR
consistency, avoids logical investment, integrated natively
and physical corruptions with the database
 Agenda

 HA / DR Challenges
 Data Guard Architecture & Features
 Data Guard 10g New Features
 Maximum Availability Architecture
 Summary
 Oracle Data Guard Benefits
1. Disaster Recovery & High Availability
– Easy failover/switchover between primary and standby databases

2. Complete data protection

– Enables zero data loss, safeguard against data corruptions, human errors

3. Efficient utilization of system resources

– Standby databases can be used for reporting, backups, queries

4. Balance data availability against performance

– Flexible data protection/synchronization modes

5. Automatic resynchronization after restoration of network connectivity

– Automatic archive gap detection and resolution with no manual intervention

6. Centralized and simple management

– Push-button graphical interface for management and monitoring
 Resources
 Maximum Availability Architecture white papers:
http://otn.oracle.com/deploy/availability/htdocs/maa.htm

 HA Portal on OTN: http://otn.oracle.com/deploy/availability/

 Data Guard home page on OTN:

http://otn.oracle.com/deploy/availability/htdocs/DataGuardOverview.html

 Oracle Documentation – Data Guard:

http://download-west.oracle.com/docs/cd/B12037_01/server.101/b10823/toc.htm

 Oracle Documentation – High Availability Best Practices:

http://download-west.oracle.com/docs/cd/B12037_01/server.101/b10726/toc.htm
 Upcoming events
 Miracle Database Forum - October 28-30 - www.miracleas.dk

– Data Guard 10g new features

– Performance Issues of Data Guard Configurations

 UKOUG annual conference - November 1-3 - www.ukoug.co.uk

– Fast Worldwide Consolidation of large databases

– Performance Issues of Data Guard Configurations

QUESTIONS
ANSWERS