RAC – Introduction

RAC stands for Real Application cluster. It is a clustering solution from Oracle that
ensures high availability of databases by providing instance failover, media failover
features. If any of the system(node) is not available, then service remains availble 
from other nodes.

So what is required to implement RAC –

1. Shared Disk System
2. Clusterware Software
3. Cluster Interconnect

Shared Disk System

When it comes Shared Disk System, you can use SAN or NAS using either NFS or iSCSI.
Once you have your Shared Disk System, the next thing you have to choose a way to
setup the disks –

1.) RAW volumes
2.) Cluster File System
3.) ASM

Oracle provides an entire software solution, including everything from disk management
with Oracle Automatic Storage Management (ASM) to data management with Oracle
Database and Oracle RAC.

So, what all this means – Just go with ASM.

Ok – Now you have a Shared system and decided to setup disks using ASM.

ClusterWare Software

Oracle Clusterware is the software, which enables the nodes to communicate with each
other, and forms the cluster and makes the nodes act as single logical server. Oracle
Clusterware is run by Cluster Ready Services (CRS) using two key components –
voting disk to record node membership information and Oracle Cluster Registry (OCR)
to record cluster configuration information.

Both must reside on shared storage.

Cluster Interconnect

Interconnects can play a vital role in high-performance computing (HPC) clusters
because they are designed to provide interprocess communication among various
individual servers thereby allowing compute nodes to behave as a single large

There are a number of Interconnect technologies available today including Gigabit
Ethernet, 10Gigabit Ethernet, Myrinet, SCI, Quadrics, and InfiniBand. Choosing the
correct one for a particular application is an important and often expensive decision.
Gigabit Ethernet, which uses the industry -standard TCP/IP protocol, is currently the
most popular choice. However, although TCP/IP is the most widely used protocol in
networking today, it is not suitable for most HPC cluster applications because of its
high CPU overhead and high latency. Specialized cluster interconnects, such as
finiBand,Myricom Myrinet, Quadrics QsNet, and Dolphin SCI, provide mechanisms to
pass messages in a more efficient manner than Gigabit Ethernet, thereby providing
lower CPU utilization and lower latency than the more popular Gigabit Ethernet
interconnect. However, most of the preceding interconnects use proprietary
technologies to achieve high performance.

This is all about a brief introduction of RAC – Let’s see explanation to the components
that we come across like OCR, Voting Disk, ASM etc., in next post.

This entry was posted in Oracle - RAC. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s