Friday, August 3, 2012


lets discuss about Clustering

First and foremost question is WHAT IS CLUSTERING?

Microsoft Sql server clustering is collection of two or more physical servers with having an identical access to the shared disk that offers the disk resources to store the database files.

By naming convention these servers are referred as 'Nodes'.These nodes will maintain the communication between one another via private network, by sending heartbeat signal between each other.If one node fails to send the signal in the cluster then automatically the secondary server will take the ownership of all the responsibilities and application services depend on the primary node to maintain the high availability.This process refers as 'Fail over'.

There is an option to make this process manually. Manual fail over is very beneficial to make the upgrades and patches to the server or some other physical server maintenance is required.

Sql Server clusters are exclusively designed for high-availability databases to eliminate the down time in association of any hardware failure.A group of services are hosted on cluster node are referred as resource groups. Those groups are resides on shared disk in the form of SAN disk.Each resource group will be mapped to a logical drive that is physically hosted on the shared disk array and also have its own associated IP address and network name.

In windows server 2003 w.r.t Sql server 2005 enterprise edition is limited to 8 nodes in a single cluster.while considering the sql server 2008 the ability is doubled i.e 16 nodes.

No comments: