Subscriber Count

    457

Subscribe2

Pages

Notes 1: Clustering Basics

What is High availability?

  • High availability is a solution to provide Fault tolerance, Disaster recovery, Disaster tolerance for business critical applications
  • It can be for Middletier, Database Tiers or the OS’s itself.
  • Common solutions for High availability are, Cluster, Cold Failover, Hot Failover.
    What is maximum availability?
  • Disaster Recovery, in other words tolerance and business continuity even after natural disasters, physical failures etc.

What is scalability?

  • Extension of the environments, ex: adding Application nodes, databases nodes etc.
  • Largely depends on the design ex:- hardware, logical coding etc.

What is cluster?

  • Cluster is a group of interconnected nodes (physical servers) that acts like a single Server. In other words clustering can be viewed logically as method for enabling multiple standalone servers to work together as a co-ordinated unit called a cluster.
  • Traditionally clustering has been used to scaleup systems, to speed up system, to survive failures.
    scale up Achieved by adding more nodes
    speed up Achieved by sharing workload
    survive Achieved by automatic failovers

What is Grid Computing with Clusters?

  • Low commodity hardware components are networked together to achieve increase computing power capacity. On demand scalability is provided by supplying additional nodes and distributing the workload to available machines

What are types of Clustering?

Hardware based clustering:-

  • IBM HACMP for example provides the servers with a clustering hardware component called coupling “I just know only one.”

Software based clustering:-

  • OS Level Clustering:- Sun clusters, Veritas Cluster, HP Service Guard, HP-Tru Cluster, Microsoft Clusterware
  • Database/Application Clustering:- Oracle Clusterware or Oracle RAC or Oracle CRS or Oracle Grid Infrastructure

What are the benefits of the Cluster?

  • Availability , in terms of
  • Scalability , in terms of extending your nodes to permissible level
  • Reliability, in terms Tolerance of faults
  • Affordability, Low commodity servers can be clustered
  • Transparency, in terms of transparent failovers capabilities
  • Increase in throughput , in terms increasing the processing capability and computing
  • What are common components of the software based Clustering? (Any clusterware)

    Clusterware (Software): To provide the single view representation of all nodes.

    • Veritas Cluster, Sun Cluster
    • Oracle Clusterware

    High speed interconnect: To provide inter cluster communication

    • Acts as a private interconnect
    • provides bandwidth scalability
    • Used for High availability of nodes
    • High speed low latency network solution like Gigabit, Memory, Hyper fabric etc.

    Shared Disk Subsystem: To provide common disk for all nodes to serve for a single databases/application etc.

    • Network based storage with capability of sharing the disks usually by cluster filesystems
    • NAS, SAN, NFS, ASM
      What kind of resources that Clusterware can provide high availability or failover?
    • OS
    • Databases/Applications
    • Storage
    • Network
      Next:- Oracle Clusterware solution & its components