Active Data Guard 12c

A Technical Overview
Name & Title

[email protected]
 Agenda

 High Availability and Data Protection Challenges

 Data Guard
 Active Data Guard
 Customers
 Summary and resources

2 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Oracle Maximum Availability Architecture (MAA)

Enterprise Manager Cloud Control

Primary Site – Integrated Management Active Replicas
– Site Guard
Oracle RAC Global Data Services Active Data Guard
– Scalability – Data Protection,
– Service Failover / Load Balancing
– Server HA – Disaster Recovery
Application Continuity – Query Offload
Flashback – Application HA
GoldenGate
– Human error – Active-active
correction – Heterogeneous
– Zero downtime maintenance

Online Redefinition, Data Guard, GoldenGate

– Minimal downtime maintenance, upgrades, migrations www.oracle.com/goto/maa

3 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Use Cases: Data Guard and Active Data Guard
Real-Time Data Protection and Availability – Eliminate SPOF
Exact copy Disaster Recovery
of primary Manual or Automatic Failover

Exact copy Query & Report Offload

of primary Open Read-Only
Oracle
Database Snapshot Convert to Test Database
(open read-write)
Standby Single Command Refresh

Data Guard New DB Standby First Patching,

Database Rolling Maintenance
Version
Redo Transport
Exact copy
SYNC or ASYNC of primary
Offload RMAN Backups

Exact copy
Source for thin snaps/clones
of primary
Note: A single DR copy may be multi- Exact copy Extract offload, source for
purposed for any combination of the of primary GoldenGate ALO mode
use cases described
Far Sync,
Redo GoldenGate Downstream

4 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Agenda

 High availability and data protection

 Data Guard
 Active Data Guard
 Customers
 Summary and resources

5 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Data Guard
Capabilities Included with Oracle Database Enterprise Edition (EE)

Data Protection High Availability Performance and ROI

Zero or sub-second data loss Extreme throughput -
Automatic database failover
protection supports all workloads
Strong isolation using Dual-purpose standby for
Automatic client failover
continuous Oracle validation development and test

Lost-write detection Standby-first patch apply Integrated management

Universal support – all data

Database rolling maintenance
types and applications
Comprehensive monitoring
Select platform migrations
with Enterprise Manager

6 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Active Data Guard Architecture
Oracle Aware Process Maintains an Exact Physical Copy of Production

Primary Database Active Standby Database

Open Read-Only
SYNC or ASYNC
Oracle Instance (in memory) Oracle Instance (in memory)
database redo

Redo
Recovery Recovery Apply
files files

Oracle data files Oracle data files

7 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Data Guard Transport for Zero Data Loss
Data Guard 12c FASTSYNC Process Architecture

User Transactions
Data Guard Processes
Queries, Updates, DDL
NSS – transmits redo from primary log buffer
Commit

RFS – receives redo from primary and writes to log file

MRP – recovery process on standby database
Primary
Online
Redo Logs

SGA LGWR Standby

Redo
Redo Logs
Buffer
NSS RFS MRP
Oracle Net

Primary Database Standby

Database

8 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Data Guard Transport for Near-Zero Data Loss
Data Guard 12c ASYNC Process Architecture

User Transactions
Data Guard Processes
Queries, Updates, DDL
TT – transmits redo from primary log buffer
Commit

RFS – receives redo from primary and writes to log file

MRP – recovery process on standby database
Primary
Online
Redo Logs

SGA LGWR Standby

Redo
Redo Logs
Buffer
TT RFS MRP
Oracle Net

Primary Database Standby

Database

9 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Data Guard 12c Automatic Resynchronization
Transactions
Oracle Net

Redo SYNC
TT
Buffer RFS MRP
SGA
NSS ASYNC
Standby
Redo
LGWR Logs
Standby
Primary
Online Database
Database Redo
ARCH
Logs

ping Alternatively, a singe RMAN

ARCH RFS command uses an incremental
backup of the primary database
to resync the standby
Archived
Redo Logs RECOVER DATABASE FROM SERVICE

10 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Data Guard Protection Modes
Balance Data Protection with Performance and Availability

Mode Risk of data loss Transport If no acknowledgement from standby:

Zero Data Loss
Maximum Stall primary until acknowledgement is
Protection Double Failure SYNC received from replica
Protection

Zero Data Loss SYNC Stall primary until acknowledgement is

Maximum
Availability Single Failure FASTSYNC received or timeout threshold period
Protection Far Sync expires – then resume processing

Maximum Potential for Primary never waits for standby

Performance
ASYNC acknowledgement
Minimal Data Loss

11 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Even Faster, Remove Delay for Human Response

 Automatic
Data – Failure detection
Guard
Observer – Database failover
– Client failover using
SYNC
SYNC/ /ASYNC
ASYNC FAN/TAF
– Reinstatement of failed
primary as a new
Site
SiteAB Site B
standby database
Primary
Standby Standby
Primary

12 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Original version
Reduce Planned Downtime New version
General Process for Database Rolling Maintenance
Database A Database B
Install new version in separate Oracle
Synchronize
homes and defer transport PROD STBY

Upgrade or perform other maintenance

Synchronize
on B then resynchronize with production PROD STBY

Switch production to B, outage is limited

SWITCHOVER
to the time needed to switch roles STBY PROD

Upgrade A via redo stream Synchronize

and synchronize STBY PROD

13 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Reduce Planned Downtime
Transient Logical
 Oracle Patchsets and new Database
Releases
‒ e.g. 11.2.0.1 > 11.2.0.3
Standby-First Transient Logical ‒ e.g. 11.2.0.3 > 12.1
Standby-First
Patch Assurance Rolling Upgrade  Upgrade Process
 Patch Set Update ‒ Begins with physical standby
 Critical Patch Update ‒ Uses logical replication between
 Patch Set Exception old and new versions
Data
 Exadata bundled patch ‒ Reverts to physical standby when
 Exadata storage server Guard both primary and standby are at
software patch Switchover new version
 Uses physical replication
between old and new
environments
 See My Oracle Support Note Tech Refresh & Tech Refresh and Migrations
1265700.1 Select Migrations  See My Oracle Support Note 413484.1 for
primary/standby cross-platform combinations
‒ Uses physical replication between old and new
environments

14 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Integrated Management
Enterprise Manager
 Simple to operate
– Single click failover
switchover, and
convert to snapshot
 Simple to monitor
– Apply and transport lag
 Automatic notification if
thresholds are exceeded
 Wizard-driven creation

15 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Oracle Multitenant – Manage as One

11g 11g

11g 11g
VS
11g 11g

11g 11g

 One Data Guard standby database

11g 11g
 One Broker/Enterprise Manager configuration
nPrimaries nStandbys  One failover/switchover command

16 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Agenda

 High Availability and Data Protection

 Data Guard
 Active Data Guard
 Customers
 Summary and resources

17 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Active Data Guard
Advanced Capabilities in an Option License for Oracle Database 12c

Data Protection High Availability Performance and ROI

Zero data loss at any distance Automatic block repair Offload read-only

Automated rolling database

Real-time cascade Offload read-mostly
maintenance
Transparent failover for in-
Offload incremental backups
flight transactions
Service management for Offload transport
replicated databases compression
Intelligent load balancing for
replicated databases

18 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 The Zero Data Loss Challenge
Trade-off Performance for Protection

Commit Ack Network Ack

Commit Network Send

Primary Standby

19 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Geographic Separation and Disaster Recovery
DR Deployed at Distances Too Great for Synchronous Replication

Distance Between Primary and DR sites

Greater than
300 miles Less than 100
miles

100-300 miles
233 Responses in an IOUG Survey

20 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Typical WAN Deployment with Data Guard 11g
Accept Data Loss in Order to Avoid Performance Impact

Standby
Data Guard London
ASYNC
Primary
Data loss failover
New York

21 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Active Data Guard Far Sync
Zero Data Loss Protection at ANY Distance

Far Sync

Far Sync Standby

SYNC Data Guard London
ASYNC
(compressed)
Primary Zero data loss failover
New York

22 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Active Data Guard Rolling Upgrades
Automates the Transient Logical Rolling Upgrade Process

Start Switchover Finish

DBMS_Rolling PL/SQL Package replaces forty two manual steps

23 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Planned Maintenance Supported
Using Active Data Guard DBMS_ROLLING

Oracle Database 12.1.0.1 onward Oracle Database 12.1.0.2 onward

Change XML-CLOB to Binary XML Database rolling upgrades
Add Partitioning to Non-Partitioned Tables New patch sets, e.g. 12.1.0.2 to 12.1.0.3
Change Basicfile LOBs to Securefile LOBs New releases, e.g. 12.1.0.2 to 12.2
Alter table to be OLTP-compressed

24 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 In-Flight Work: Dealing With Outages
Current Situation

 Database outages can cause in-

End User
flight work to be lost, leaving
Application users and applications in-doubt
Servers
 Impact of outage is visible and
leads to:
Database
 User frustration
Servers
 Duplicate submissions
 Rebooting mid-tiers

25 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Transaction Guard
Preserve and Retrieve COMMIT Outcome
 API that supports known commit
outcome for every transaction
End User
 Without Transaction Guard, upon
Application failures – transaction retry can
Servers
cause logical corruption
 With Transaction Guard,
Database
applications can deal gracefully
Servers with error situations, vastly
improving end-user experience
 Used transparently by
Application Continuity

26 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Application Continuity
Masks Unplanned/Planned Outages

 Replays in-flight work on

End User
recoverable errors
Application
Servers  Masks many hardware, software,
network, storage errors and outages
when successful

Transaction Replayed
Database
Servers
 Improves end-user experience and
productivity without requiring
custom app development

27 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Databases in Replicated Environments
Challenges
 No seamless way within Oracle to
balance workload when identical
services are available on multiple
replicas
GoldenGate

 Automatic service failover in replicated

environments is not available for all use
Active Data Guard cases
– Active Data Guard read-only services
– GoldenGate

28 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Global Data Services
Load Balancing and Service Failover for Replicated Databases

 Extends RAC-style service failover,

Global Data Services load balancing and management
capabilities to a set of replicated
databases
GoldenGate
 Takes into account network latency,
replication lag, and service placement
policies
Active Data Guard
 Achieve higher availability, improved
manageability and maximize
performance

29 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.