Copyright © TIBCO Software Inc. All Rights Reserved

Overview

A cluster is a group of machines that work together as a single system to ensure that applications and resources are available at all times. The machines are managed as a single system using cluster software, which provides a way to support fault-tolerance, high-availability, scalability, and so on.

The computers are physically connected by cables and programmatically connected by cluster software. These connections allow computers to use failover and load balancing, which is not possible with a stand-alone computer.

Figure 1 shows a failover cluster solution. The nodes in a cluster are configured with a shared drive. The cluster software manages access to the shared drive so that only applications on a single machine use a particular resource at any time.

The failover cluster solution primarily provides uninterrupted service in the event of a failure within the cluster. This solution is commonly used for database applications.

Applications that are deployed to a failover cluster can be managed under cluster software. Applications are under the control of cluster software, which ensures that only one application runs at once on one of the cluster machines. This kind of application is in active-passive mode.

Figure 1 Failover Cluster Solution

Cluster Aware and Cluster Unaware Applications

A cluster environment supports two types of applications: cluster-aware applications and cluster-unaware applications.

TIBCO applications are cluster-unaware. The applications can be run seamlessly in a cluster environment, but do not use any cluster features.

•	Cluster-aware applications run in the cluster and are aware if the cluster itself is running, or whether a resource needed by the application is available. A cluster-aware application is supported by a resource type that is assigned by the cluster software.

•	Cluster-unaware applications know nothing about the cluster environment and consequently cannot exchange data with cluster objects or be aware that they are running under cluster software control.

Cluster software manages cluster-unaware applications by using basic methods for failure detection and application shutdown. For example, the Microsoft Cluster Server solution manages cluster-unaware software as a generic application resource type.

High-availability Modes

High-availability applications can be categorized to different modes, hot standby, warm standby, and cold standby in a cluster.

When TIBCO applications are installed under a cluster, high-availability is best characterized as warm-standby.

•	Hot standby—Applications on the backup node take over with almost no downtime and may even approach zero downtime. Using hot standby, two processors use hardware checkpoints to verify synchronization after each CPU instruction.

•	Warm standby—A slight delay occurs when a backup application takes over for another application.

•	Cold standby—The delay is a little longer than warm standby. In cold standby, the cluster software can't detect an outage and applications must be restarted manually. One or more spare processors are set aside to be used if an in-service processor fails.

Basic Cluster Terminology

The following basic terminology is used in this document.

•	Node. The term used to refer to a server that is a member of a cluster.

•	Resource. A hardware or software component that exists in a cluster, such as a disk, an IP address, a network name or an application.

•	Group. A combination of resources that are managed as a unit of failover. Groups are also known as resource groups, cluster packages, or service groups.

•	Dependency. An alliance between two or more resources in the cluster architecture.

•	Failover or failback. The process of moving resources from one server to another. Failover can occur when one server experiences a failure of some sort or when an administrator initiates a proactive failover.

•	Active/Active. From a software perspective, this describes applications (or resources) that can exist as multiple instances in a cluster. This means that both nodes can be active servicing clients.

•	Active/Passive. This term describes applications that run as a single instance in a cluster. In general, this also means that one node typically is idle until a failover occurs.

•	Rolling Upgrade. A rolling upgrade allows you to upgrade the software on cluster nodes, one node at a time, so that the applications and resources on the cluster are always available even though the nodes being upgraded are not.

•	Shared storage. This refers to the external SCSI or fiber channel storage enclosure and the disks contained therein. Shared storage is a requirement for multi-node clusters. Although this storage is shared, only one node can access an external storage resource at any given time.

Copyright © TIBCO Software Inc. All Rights Reserved