Transcript Oracle EPM Disaster Recovery Best Practice

<Insert Picture Here>
Oracle EPM Disaster Recovery Best Practice
For Release 11.1.1.3
Agenda
•
•
•
•
•
DR Approach
Asymmetric Deployment
Database Recommendations
Application Configuration and Failover
Existing Documentations
Disaster Recovery Approach
Active/Standby Solution
Primary Site
DNS switch/Geo LBR
Application
Standby Site
Application
WAN
NAS/SAN
Mounted
NAS/SAN
Mounted
Disk Replication
BINARIES
CONFIGURATION
BINARIES
CONFIGURATION
Application
Data Storage
LDAP Data
Storage
Metadata
Storage
Security Information
Storage
Data Guard
Application
Data Storage
LDAP Data
Storage
Metadata
Storage
Security Information
Storage
Asymmetric Deployment
Production Site
Standby Site
Global Router
Firewall
Firewall
Web
Tier
Web
Tier
Firewall
Planning
Firewall
Disk Replication
Foundation
Essbase
Average
Latency and
Bandwidth
WAN
Planning
Foundation
Essbase
Financial
Management
Financial
Management
Data Guard
RAC
Firewall
RAC
Firewall
Hostname Requirement
• Production and standby sites are in completely
separate networks and the fully qualified hostnames
can be the same in both sites
• Or, production and standby sites have different DNS
that resolves the hostnames to the right IP in their
network
• The standby site can have standby DNS that is activated on a
disaster event
• Or, production hostnames are resolved to local IP
address in standby site via /etc/hosts file
Configure DR Environment
• Install and Configure Primary Site
• Binaries and Data should be on a replicated partition
• Do not distribute services between servers that are not clustered
• If hostnames are to be different in standby site, setup
hostname aliases in Standby Site
• Install and Configure Standby Site
• Setup Database Replication
• Backup & restore is ok
• Setup Disk Replication
• Backup & restore is ok
• Enable Standby Site
Install and Configure Standby Site
• Install and configure in both production & standby site
HyperionHome for
each server is in a
separate directory;
it should be a
network share or
mount point in the
servers
Note: BI+ repository and
Essbase App directories
should be in a common
share across machines,
e.g. under /data
Setup Hostname Aliases
• If the hostnames have to be different between the two sites and there is no
separate DNS for standby site, setup alias for production site servers in the
standby site in the following way.
/etc/hosts entries in standby site;
order of server entries is important
10.2.1.111
10.2.1.112
10.2.1.113
10.2.1.114
...
server1 drserver1
server2 drserver2
server3 drserver3
dbserver drdbserver
Database Recommendations
• Use DB Host Name alias on Standby Site
• Data Guard Configuration for data repositories
• For planned configuration changes, force database
synchronization with Oracle Data Guard
• A backup & restore procedure will also work
Setup Disk Replication
• Enable disk replication
• A backup & restore procedure will also work
Enable Replication
Enabling Standby Site
• Break mirroring between primary and standby.
• Run Crash Recovery Procedure for each application
• Recover OpenLDAP
• Recover Essbase
• Bring up the services on the standby hosts
Existing Documentation
• Oracle Fusion Middleware Disaster Recovery Guide
http://download.oracle.com/docs/cd/E10291_01/core.1013/e12297/toc.htm
• Oracle Data Guard
http://www.oracle.com/technology/deploy/availability/htdocs/DataGuardOverview.html
• Oracle Fusion Middleware DR Solution using NetApp
Storage
http://www.netapp.com/us/library/technical-reports/tr-3672.html