Red Hat CLUSTER SUITE FOR ENTERPRISE LINUX 5.1 Instrukcja Użytkownika Pobierz pdf (Strona 8)

Introduction to Linux Clustering

4 Clustering fundamentals

4.1 Basics

High-availability clustering is a complex topic, and it is important to fully understand key

concepts behind it.

The basics are simple:

● Each computer is called a node.

● Two or more nodes form a cluster.

● In the event of a failure of any one of the nodes, the remaining nodes will take up the

work being performed by the dead node.

What makes clustering complex, is how the cluster handles node failures, shared disk storage

and situations such as split-brain.

A cluster typically works as follows:

● All the nodes run the cluster management software (eg: Linux-HA or Redhat Cluster

Suite), which controls stuff such as heartbeats, application starting/stopping and

keeping quorum (more on this later).

● One of the nodes runs an administration application, that allows you to manually

add/remove nodes and provides the ability to manually move applications from one

node to another.

● In the event of a failure with a node, the other nodes fence it, which involves using a

hardware device like a managed power switch to physically turn the node off. This is

done to prevent the node from writing to any of the storage devices and corrupting

the data.

The other nodes then decide which node should run the applications that were on the

dead node, and one of the nodes will be chosen (depending on the config) and will

start up the application.

● All the cluster nodes have access to a central storage array (eg: a SAN or network

attached storage). This storage location runs a clustered filesystem which allows all

the nodes to read and write at the same time.

4.2 Important clustering components

4.2.1 FAILOVER

There are 3 types of failover methods that exist.

1 2 3 4 5 6 7 8 9 10 11 12 13 ... 32 33

Brak uwag

Red Hat CLUSTER SUITE FOR ENTERPRISE LINUX 5.1 Instrukcja Użytkownika Strona 8