The High availability clusters are computer clusters which are implemented mainly for the purpose of improving the availability of the services. The high availability clustering detect hardware/software faults, and instantly restart the application on another system without requiring administrative intercession, this process is commonly known as a Failover process. It has functionality that provides failover from one node to another if the current node becomes unavailable, a failover cluster appears on the network as a normal application, but it has additional functionality that increases its availability in case one server fail.
 –>
The clustering software may configure the node before starting the application on it. To run the system smoothly it require appropriate filesystems to be imported and mounted and to set-up cluster rules that enable a common set of configurations also network hardware may have to be configured, and some supporting applications may need to be running as well.Diagram: HA Cluster network (2 node)
The purpose of HA clusters to utilize all available techniques to make the individual systems/shared infrastructure as reliable as possible. The high availability clustering should include the following;
* Build redundancy into a cluster.
* Redundant network connections so that a switchover or network interface failures do not result in network outages.
* Disk mirroring to avoid system crashes in case of failure of internal disks.
* SAN data connections.
* Redundant electrical power inputs.
The high availability clusters generally use a critical clustering in private network connection and monitor the status of each node in the cluster. Such clustering are often required for critical databases, business applications, file sharing on a network and customer services such as e-commerce websites.
September 24th, 2008 at 2:27 pm
[...] the best reliability possible, as most of the Hosting accounts come with a Clustered Environment. Clustering means that multiple servers back-up each other to provide un-interrupted running of your Managed [...]