1. Client: Oh! My
Server Failed what
to do Next ..?
Can Any One help
me?
Presented By:
Saurabh Soni
2. Down Time
Unplanned
Planned
Upgradation Disaster
•DB Upgrade •Fire
C
•OS Upgrade •Flood
A
•Middleware Upgrade •Earthquake
U
•Application Upgrade S
E
S
Planned Routine Operation Human Error
O
•Backup F •System Admin
•Security Management •Operator Error
D
•Performance Management •User Error
O
•DBA
W
N
Periodic Maintenance S/W & H/W Failure
T
I
•Software Patches •Middleware & Application
M
•Storage Maintenance •Operating System
E
•Initialization Parameters •Power Supply
•Schema Management •Database
•Operating System •Network
•Middleware •Memory
•Tap & Disk
3. Questions Arrives !!!!
Availability
Does Company do a Commitment of an availability of System?
Is Company’s profit tied to System Availability?
Do Human lives depend on System Availability?
Scalability
Cost
Downtime Cost for an Hour?
Recovery
Does Company need Disaster Recovery System?
Company’s Applications Viewed Directly
5. Oracle RAC Architecture
•Detail Example of two nodes:
•Each Node has its Own Table Space - Memory Area
- Own Log File
•If One node have More load then it automatically activates another Node.
6. Features
Improved Work load
Throughput Management
Scalability Over REAL
Transparency
Single Instance APPLICATION
Systems CLUSTER
Improved
High Availability Response Time
• Improved Throughput: because more node handles the situation very well.
• High Availability : Server available 99.99% if any Node Fails handle the Situation.
• Improved Response Time: Better response by RAC because Node Distribute Work.
• Transparency & scalability over single instance system that we can add more node
As we want to increase the performance of System.
7. Existing Infrastructure
• Big Cost
• Hard to Maintain
• Number of Computer
Resource
• Hard to align with
Business processes
• Availability
• Full Scalability
• Server Shared
Between Application
EMAIL • Automatic Workload
ERP
Management
CRM
DW
8. Work Load Management
Spare Node
Supply Chain
Node
Working Node
Test Situation : Normally RAC System With Normal Server Allocation.
Node act as a Server in the above Scenario.
Normally System has 1 spare Server (Wheel) RAC (CAR)
9. Work Load Management
Spare Node
Supply Chain
Node
Working Node
Here One Server Fails
Still Database Available.
Test Situation : In RAC System With Server Fails.
Fail Server is not working and removed from the RAC.
Fail Server (Puncher Wheel) Remove from RAC (CAR).
Spare server Working as Working Node and Handle the Node.
10. Comparison Of Technology
Shared Data Partitioned Data
Individual Instance
Multiple Instances
Individual Database
Shared Database
SQL Server
Oracle RAC
Shared Nothing Cluster.
Very High Level
Group of independent databases.
Shared disk
View of single database instances.
Shared cache
Working on Partitioned SQL Server back
Served by 1 or more database server
up server mechanism follows.
instances
Data Replication use if one Of the Node fails the
All instances working again the same
Other Backup device handle the situation.
copy of data
11. Benefits
• Expand the Scalability
• Full Transparency
• High Availability
• Buffer Cache Management via
Distributed Memory
• Row Locking & Concurrency
Control.
• Recovery Manager, Online
Backups, and Archiving.
• Fast Commits, Group Commits,
and Deferred Writes.
13. Cost Estimation
Which Factors are Major Affected:
1. Hardware and software cost
2. Supplier support and maintenance cost
3. Personnel to support and maintain the system cost
4. Capital spent or saved for a particular product(s) cost
14. REAL WORLD EXAMPLE WITH FACT
FIGURES !!!!!!
• Syndicate Bank implemented core banking with the Help
Of Oracle Application Architecture (RAC).
• Over 1700 Branches as of January 2008 impacting 19 million
customers over India.
• Oracle RAC provide on-line, real-time banking across
branches seamlessly.
• The clustered database environment has helped optimize
maximum data availability at reduced costs.
• Not only this, Syndicate Bank achieved a record last year
when 101 branches went live on a single day.
15. References
Books
Oracle Database 10g: High Availablity with RAC Flashback & Data Guard
by Matthew Hart and Scott Jesse ISBN:0072254289
WebSites
http://www.dbasupport.com/oracle/ora9i/rac.shtml
http://www.oracle.com/technology/products/database/clustering/pdf/managing_rac_rel2.pdf
http://lbdwww.epfl.ch/f/teaching/courses/oracle9i/rac.920/a96597/pshwarch.htm#14054
http://www.oracle.com/technology/products/database/clustering/pdf/twp_rac11g.pdf
www.oracle.com
http://www.miracleas.dk
http://press-releases.techwhack.com/18720-syndicate-bank
http://msdn.microsoft.com/en-us/library/aa479364.aspx
WhitePapers
http://www.oracle.com/technology/products/database/clustering/pdf/Oracle9iRAC
BusinessWhitePaper.pdf