Extreme Availability With Oracle Database 12c: Platform Technology Solutions, Oracle Server Technologies

Extreme Availability With
Oracle Database 12c
Platform Technology Solutions,

Oracle Server Technologies
1 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.
Oracle Database 12c
Key New Features for High Availability
 Global Data Services

 Data Guard Enhancements
 RMAN Enhancements
 Other HA Enhancements
 GoldenGate Update

Oracle Database 12c
Extreme Availability
 Oracle Database 12c introduces significant new (HA) capabilities that

– Drastically cut down planned and unplanned downtime
– Eliminate compromises between HA and Performance
– Tremendously boost operational productivity
 These take Availability to unprecedented new levels

– Next-generation Maximum Availability Architecture (MAA)
– Optimized for Oracle

Maximum Availability Architecture
Production Site Application Continuity Active Replica
– Application HA
RAC Active Data Guard
Global Data Services – Data Protection, DR
– Scalability
– Service Failover / Load Balancing – Query Offload
– Server HA
Flashback
– Human error GoldenGate
correction – Active-active
– Heterogeneous
Edition-based Redefinition,
RMAN, Oracle Secure Backup
Online Redefinition, Data Guard, GoldenGate – Backup to tape / cloud
– Minimal downtime maintenance, upgrades, migrations

Oracle Database 12c
High Availability Key New Features


Database Workload Management
Challenges in Replicated Environments
 Challenges
– No seamless way to efficiently use all databases
– No automated load balancing and Service HA
Active Standby Primary Active Standby – No complete centralized workload management
 Resulting in
– Sub-optimal resource utilization
GoldenGate
– Application outages when replicas fail
– Workload on replicas managed in silos
Solution: A holistic automated workload management framework for replicated databases

Global Data Services
Load Balancing and Service Failover for Replicated Databases
• Extends RAC-style service failover, load

balancing (within and across data
centers), and management capabilities to
a set of replicated databases
• Takes into account network latency,
replication lag, and service placement
policies
• Achieve higher availability, improved
manageability and maximize
performance

GDS for
Global Data Services High Availability
Active Data Guard Example
 Reporting client routed to ‘best’ database
Reporting Service – Based on location, response time, data, acceptable
data lag
– Reports will automatically run on least loaded server
Call Center Service
 Reporting client failover
– If preferred database not available, will route to
Active Data Guard another database in same region or a remote
database
 Global service migration

– Automatically migrates services based on
failover/switchover - if primary database is down, start
Call Center service on the new primary

Global Data Services GDS for
Use Case: Active Data Guard without GDS High Availability
Critical E-Commerce App accessing What happens when
Active Data Guard Standby Active Standby is down?
Order Capture Order History View Order Capture Order History View
Orders History Orders History

Service Service Service Service
Data Guard Data Guard
Primary Active Standby Primary Active Standby

GDS for
Global Data Services High Availability
Use Case: Active Data Guard with GDS
Order Order History

Capture View
When Active Standby is down …

 GDS fails over History Service Global Data Services
to primary, redirects
connection through FAN/FCF
Orders History History
Service Service Service
Data Guard
Primary Active Standby

GDS for
Global Data Services Workload Management
GoldenGate Example
 Call Center Client connections and

requests transparently routed to the
Call Center Service closest / best database
 Runtime load balancing metrics give client
GoldenGate real-time information on which database to
issue next request
 If a database fails, its global services
restarted on another replica

Global Data Services (GDS) GDS for
Automated Workload Management Workload Management
for Replicated Databases
 Takes into account – Database instance load, network latency,

workload management policies (replication lag, and DB cardinality,
region affinity, Database role, load balancing goals)
 Based on two foundations:
– Services: proven concept for dynamic workload management for RAC
– Replication: Active Data Guard, Oracle GoldenGate
 Benefits:
– Higher Availability via Service failover across local / global databases
– Better Scalability via multi-database integrated load balancing
– Better Manageability via centralized administration of global resources

GDS for
Global Data Services Workload Management
Concept
Unified Framework
Sales Service Sales Service

Global Data Services Configuration
Sales Global Service
GoldenGate
GoldenGate
Without GDS With GDS

(Database pools)

Oracle Global Data Services Capabilities
Automated Workload Management
Region based Workload Routing
Lag based Workload Routing
Connect-time Load Balancing

on a Set of Replicated Run-time Load Balancing
Databases
Inter-database Service Failover
Role based Global Services
Centralized Workload Management

Global Data Services: Concepts
Clients
(Users)
 Global Service: Database Service provided by
Clients
(Users) multiple databases with replicated data
Data Center #2
Data Center #1
EMEA Region
– Local service + {region affinity, replication lag,
App/Middleware APAC Region App/Middleware
database cardinality, load balancing goals}
 Global Service Manager (GSM): Provides main GDS
GDS Framework
functionality: service management and load balancing
– Regional listener to incoming database connections
Global Service
– At least one GSM per region or multiple GSMs for HA
Global Service GDS
Managers GDS
Catalog
Managers – All databases/services register to all GSM Listeners
Catalog
Primary Standby
 GDS Catalog: stores all metadata, enables centralized
global monitoring & management
SALES POOL (sales_report_srvc, sales_entry_srvc)
GDSCTL – Global service configuration stored in GDS Catalog
and EMCC Replication
 GDS Region: Group of databases and clients in close
network proximity, e.g., East, West
SALES Databases
 GDS Pool: Databases that offer a common set of
global services, e.g., HR, Sales
 GDSCTL: Command-line Interface to administer GDS
*GUI support of GDS in EMCC12.1.0.5

Global Data Services: Concepts
 GDS Region: Group of databases and clients in close network proximity, e.g., East, West
 GDS Pool: Databases that offer a common set of global services, e.g., HR, Sales
 Global Service: Database Service provided by multiple databases with replicated data
– Local service + {region affinity, replication lag, database cardinality}
 Global Service Manager (GSM): Provides main GDS functionality: service management
and load balancing
– Clients connect to GSM instead of database listener
– At least one GSM per region or multiple GSMs for High Availability
– All databases/services register to all GSM Listeners
 GDS Catalog: stores all metadata, enables centralized global monitoring & management
– Global service configuration stored in GDS Catalog
 GDSCTL: Command-line Interface to administer GDS

Application Workload Suitability for GDS
GDS is best for applications that:
 Are replication-aware
– Designed to work in replicated environments
 Can distribute their work into read-only and read-write services
– Example: read-only loads on an Active Data Guard standby
 Are aware of and avoid or resolve update conflicts
– Example: Oracle GoldenGate multi-master configurations
 Can tolerate some replication lag
– Many reporting applications can work with slightly out-of-date data

Oracle Database 12c


Zero Data Loss Challenge
Synchronous Communication Leads To Performance Trade-Offs
Commit Ack Network Ack
Commit Network Send
Primary Standby
The longer the distance, the larger the performance impact

Data Guard Async – Before 12c
Some Data Loss Exposure Upon Disaster
ASYNC Standby
Primary

Active Data Guard Far Sync – New in 12.1
Zero Data Loss For Async Deployments
 Far Sync: light-weight Oracle instance: standby control file, standby

redo logs, archived redo logs, no data files
 Receives redo synchronously from primary, forwards redo
asynchronously in real-time to standby
 Upon Failover: Async standby transparently obtains last committed
redo from Far Sync and applies: zero data loss failover
 Second Far Sync Instance can be pre-configured to transmit in
reverse direction after failover/switchover
 Terminal standbys required to be Active Data Guard Standbys

Active Data Guard Far Sync
Operational Flow
Standby
Primary
SYNC
ASYNC
Commit
Ack
Far Sync
Instance
No Compromise Between Availability and Performance!

Data Guard Far Sync
No Compromise Between Availability and Performance
Standby
Primary Standby
Primary
async
sync
Far Sync Far Sync

Instance Instance
(Active)
(Idle) (Active)
(Idle)

Data Guard Far Sync
Standby between continents with zero data loss
Far Sync
Instance
(Idle)
Far Sync
Instance
(Active)
async
London
Primary Reading Standby
sync
Bridgeport
New York

Active Data Guard Far Sync
Benefits
 Best data protection, least performance impact
 Low cost and complexity
 Best way to implement a near DR + Far DR model
 Relevant to existing Data Guard ASYNC configurations
 Data Guard Failover? No Problem! Just do it – No Data Loss!

Active Data Guard Real-Time Cascading
Eliminates Propagation Delay
 Prior to 12.1, Standby 1 waits till log switch before forwarding redo
from archived logs to Standby 2
Primary SYNC or ASYNC Standby 1 ASYNC Standby 2
 In 12.1, Standby 1 forwards redo to Standby 2 in real-time, as it is

received: no propagation delay for a log switch
 Standby 2 (Active Data Guard Standby) is up-to-date for offloading
read-only queries and reports

Data Guard Fast Sync
Reduced Primary Database Impact for Maximum Availability
Primary Standby
LGWR NSS RFS  For SYNC transport: remote site
Commit acknowledges received redo before
Commit writing it to standby redo logs
Acknowledge Redo Standby
Logs Redo
Logs
 Reduces latency of commit on primary
 Better DR – increased SYNC distance
Acknowledge
 If network round-trip latency less than

returned on receipt
Primary Standby
LGWR NSS RFS time for local online redo log write,
Commit
synchronous transport will not impact
Commit
Acknowledge Redo
primary database performance
Standby
Logs Redo
Logs

Data Guard
Other New Features in Oracle Database 12c
Rolling Upgrade With Active Data Guard DML on Global Temporary Tables
 Automate complexity through simple PL/SQL Package:  Temporary undo is not logged in redo logs
DBMS_ROLLING (12.1.0.1 onwards), with simple Init,
 Enables DML on global temporary tables on Active
Build, Start, Switchover, Finish procedures
Data Guard: more reporting support
 Additional Data Type Support: XML OR, Binary XML,
Spatial, Image, Oracle Text, DICOM, ADTs (simple  Set by default on Active Data Guard standby
types, varrays), …
Validate Role Change Readiness Unique Sequences

 Ensure Data Guard configuration ready for switchover  Primary allocates a unique range of sequence numbers
with automated health checks – verify no log gaps, to each Standby
perform log switch, detects any inconsistencies,
 Enables more flexible reporting choices for Active Data
ensures online log files cleared on standby, …
Guard

Oracle Database 12c


Multitentant: Backup as One
Recover at Pluggable Database Level
One Backup
Point-in-time recovery
at pluggable database level

Fine-grained Table Recovery From Backup
 Simple RECOVER TABLE command to
recover one or more tables (most
recent or older version) from an RMAN
backup
101  Eliminates time and complexity
01 010 associated with manual restore, recover
1 1 101001
1 011011110 & export
0 0 1 0 1 00 0 0
– Enables fine-grained point-in-time
1001101
1010 recovery of individual tables instead of
01 RMAN the contents of the entire tablespace
Backups

Table-Level Recovery From Backup (RMAN)
STEPS
 Determines which backup sets contains the table(s) or partition(s)

and which redo logs are required for recovery
 Creates an auxiliary instance and recovers specified table(s) or partition(s)
until the specified point in time or scn (location optional)
 Creates a Data Pump export dump file that contains the recovered
table(s) or partition(s)
 Imports the Data Pump export dump file into the target instance (optional)

Cross-Platform Backup & Restore
Simplified Platform Migration
Source Database (AIX) Destination Database (Solaris)
1001
00 1010000 1
1 0110111101
1 1 101
1 0000101 00
1011110110 1
Backup
101 1 1
1001
to Disk/Tape Restore Backup

(data files, optional (optional endian
endian conversion, conversion,
metadata export) metadata import)
 Simplifies procedure for platform migration

 Minimize read-only impact with multiple incremental backups

Pluggable Database Backup & Restore
Fine-Grained Backup & Recovery to Support Consolidation
 Backup and recover specific pluggable databases with new PLUGGABLE DATABASE keywords:
RMAN> BACKUP PLUGGABLE DATABASE <PDB1>, <PDB2>;
 Familiar BACKUP DATABASE command backs up CDB, including all PDBs
 PDB Complete Recovery
– RESTORE PLUGGABLE DATABASE <PDB>;
– RECOVER PLUGGABLE DATABASE <PDB>;
 PDB Point-in-Time Recovery
– RMAN> RUN {
– SET UNTIL TIME 'SYSDATE-3';
– RESTORE PLUGGABLE DATABASE <PDB>;
– RECOVER PLUGGABLE DATABASE <PDB>;
– ALTER PLUGGABLE DATABASE <PDB> OPEN RESETLOGS; }
 Familiar RECOVER DATABASE command recovers CDB, including all PDBs

Better Performance
Other New Features in Oracle Database 12c
 Enhanced Multi-section Backup capability: now supports image copies

and incremental backups
 More efficient synchronization of standby database using simple
RMAN command: RECOVER DATABASE … FROM SERVICE
 Enhanced Active Duplicate
– Cloning workload moved to destination server via auxiliary channels,
relieving resource bottlenecks on source
– Cloning can now use RMAN compression and multi-section capability to
further increase performance

Oracle Database 12c


Other HA Enhancements
Online Datafile Move Online Redefinition Enhancements

 Relocate a datafile while users are actively  Improved sync_interim_table performance
accessing data: ALTER DATABASE MOVE
DATAFILE …  Ability to redefine table with VPD policies
 Maintains data availability during storage  Improved resilience of finish_redef_table

migration  Better handling of multi-partition redefinition
Separation of Duties Additional Online Operations

 SYSDG / SYSBACKUP: Data Guard & RMAN  Drop index online / Alter index unusable online / Alter
specific administrative privileges index visible / invisible online
 No access to user data: enforce security  Drop constraint online / Set unused column online
standards throughout the enterprise
 Online move partition: ALTER TABLE … MOVE
PARTITION … ONLINE

Oracle Database 12c


Oracle GoldenGate 12c
Low-Impact, Real-Time Data Integration & Transactional Replication
New DB/ Zero Downtime Upgrade
HW/OS/APP & Migration
Fully Active
Distributed Active-Active High Availability
DB
Oracle Database12c * Exact Copy Disaster Recovery for

of Primary Non-Oracle Databases
Log-based Reporting
Query & Report Offloading
Database
Changed Data
Data Real-time BI, Operational
Oracle & Non-Oracle Reporting, MDM
Warehouse
Database(s)
Data Integrator
ODS Global
Data Synchronization within the
Data Enterprise
Centers
Message Bus Message
Event Driven Architecture, SOA
Bus

GoldenGate Zero Downtime Migration/Upgrade
Seamless Migration and Upgrades to Oracle Database 12c
• Consolidate/migrate/ Oracle Database 12c

Switchover
maintain systems without
downtime
• Minimize risk with Compare &
Verify using
failback option Oracle GoldenGate
Veridata
• Validate data before
Real-Time Optional
switchover Replication Failback
• Use Active-Active For Migrations Data Flow
Nn-
replication for phased Non-
Oracle Oracle Oracle
Oracle 10.2 11.2
user migration ERP
ERP CRM DW

Oracle GoldenGate for Active-Active Databases
Increase ROI on Existing Servers & Synchronize Data
• Utilize secondary
Oracle Database 12c
systems for transactions
• Enable continuous
availability during
unplanned & planned
outages
• Synchronize data across Heterogeneous Bi-Directional
global data centers Real-Time Replication
• Use intelligent conflict Non- Oracle Oracle

detection & resolution Oracle 10.2 11.2
App App2 App3

Oracle Database 12c
Extreme Availability: Summary
 Oracle Database 12c offers a tremendously sophisticated

set of high availability (HA) capabilities
 These capabilities
– Further reduce downtime
– Significantly improve productivity
– Eliminate traditional compromises

Safe Harbor Statement
THE PRECEDING IS INTENDED TO OUTLINE OUR GENERAL

PRODUCT DIRECTION. IT IS INTENDED FOR INFORMATION
PURPOSES ONLY, AND MAY NOT BE INCORPORATED INTO ANY
CONTRACT. IT IS NOT A COMMITMENT TO DELIVER ANY
MATERIAL, CODE, OR FUNCTIONALITY, AND SHOULD NOT BE
RELIED UPON IN MAKING PURCHASING DECISIONS. THE
DEVELOPMENT, RELEASE, AND TIMING OF ANY FEATURES OR
FUNCTIONALITY DESCRIBED FOR ORACLE’S PRODUCTS
REMAINS AT THE SOLE DISCRETION OF ORACLE.


Extreme Availability With Oracle Database 12c: Platform Technology Solutions, Oracle Server Technologies

Uploaded by

Document Information

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

Extreme Availability With Oracle Database 12c: Platform Technology Solutions, Oracle Server Technologies

Uploaded by

Copyright:

Available Formats

Extreme Availability With

Oracle Database 12c

Platform Technology Solutions,

 Global Data Services

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

 Oracle Database 12c introduces significant new (HA) capabilities that

 These take Availability to unprecedented new levels

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

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

 Global Data Services

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

Solution: A holistic automated workload management framework for replicated databases

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

• Extends RAC-style service failover, load

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

 Global service migration

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

Orders History Orders History

Data Guard Data Guard

Primary Active Standby Primary Active Standby

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

Order Order History

When Active Standby is down …

Primary Active Standby

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

 Call Center Client connections and

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

 Takes into account – Database instance load, network latency,

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

Sales Service Sales Service

Without GDS With GDS

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

Lag based Workload Routing

Connect-time Load Balancing

Global Data Services

Role based Global Services

Centralized Workload Management

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

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

 GDSCTL: Command-line Interface to administer GDS

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

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

 Global Data Services

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

Commit Ack Network Ack

Commit Network Send

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

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

 Far Sync: light-weight Oracle instance: standby control file, standby

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

No Compromise Between Availability and Performance!

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

Far Sync Far Sync

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

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

 Best data protection, least performance impact

 Low cost and complexity

 Best way to implement a near DR + Far DR model

 Relevant to existing Data Guard ASYNC configurations

 Data Guard Failover? No Problem! Just do it – No Data Loss!

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

Primary SYNC or ASYNC Standby 1 ASYNC Standby 2

 In 12.1, Standby 1 forwards redo to Standby 2 in real-time, as it is