HP manual Arbitration for Data Integrity in Serviceguard Clusters

Page 3

Arbitration for Data Integrity in Serviceguard Clusters

Arbitration for Data Integrity in Serviceguard Clusters

 

Clustering is an architecture of interconnected servers that allows

 

multiple hosts to run the same applications, permitting the individual

 

systems to be up or down. Applications move easily between systems,

 

accessing the same shared data from different nodes at different times.

 

The goal is to provide high availability for the application and the data

 

without endangering the integrity of the data. Particularly, the cluster

 

manager software must ensure that when an application is running and

 

processing data on one node in the cluster, the same application does not

 

inappropriately start up on another node and begin processing the same

 

data at the same time. Cluster software must protect data integrity at all

 

costs.

 

Different clustering solutions have adopted various methods for ensuring

 

the safety of data. Some solutions rely solely on redundant membership

 

links to ensure data integrity. Others, including the Serviceguard

 

product family, use a process called arbitration to keep more than one

 

incarnation of a cluster from running and starting up a new instance of

 

an application. The Serviceguard team has always agreed that definitive

 

arbitration should be used in high availability systems that run

 

mission-critical applications.

 

This paper describes some basic cluster membership concepts, then

 

discusses some different ways of determining cluster membership.

 

Several types of arbitration are shown with example configurations that

 

illustrate each type. Then the arbitration methods used by Serviceguard

 

and related products are presented in some detail.

 

 

NOTE

Some of the arbitration methods mentioned in this paper are available

 

only on particular platforms or with specific types of configuration. Other

 

methods, while possible, have not been fully tested on all types of

 

systems. For details about support of particular kinds of arbitration,

 

refer to your platform’s Configuration Guide, available through your HP

 

representative.

 

 

3

Image 3
Contents Manufacturing Part Number B3936-90078 July Arbitration For Data Integrity Serviceguard ClustersLegal Notices Arbitration for Data Integrity in Serviceguard Clusters Membership Cluster Membership ConceptsCluster Membership Concepts Split-Brain QuorumTie-Breaking To Arbitrate or Not to Arbitrate No Arbitration-Multiple PathsMultiple Heartbeat Failures Single Node Failure No Arbitration-Multiple MediaAdditional Multiple Paths with Different Media Multiple Paths with Different MediaNo Arbitration-Risks How Serviceguard Uses Arbitration Cluster StartupStartup and Re-Formation Dynamic Cluster Re-Formation Cluster Quorum and Cluster LockingCluster Lock No Cluster Lock Lock Requirements Lock Disk Operation Use of a Lock Disk as the Cluster LockDual Cluster Lock Single Cluster LockUse of a Lock LUN as the Cluster Lock Lock LUN Operation Oot IrrorQuorum Server Operation Use of a Quorum Server as the Cluster LockRunning the Quorum Server Setting up the Quorum ServerQuorum Server Status and State Specifying a Quorum ServerViewing Quorum Server System Data Viewing Quorum Server Status and StateUse of Arbitrator Node Use of Arbitrator NodesArbitration in Disaster-Tolerant Clusters Extended Distance ClustersMetropolitan Clusters Quorum Server Arbitrator NodesUse of Dual Lock Disks in Extended Distance Clusters Continental ClustersDisk area is not mirrored Arbitration for Data Integrity in Serviceguard Clusters Arbitration Advantages Disadvantages Mode SummaryComparison of Different Arbitration Methods Arbitration for Data Integrity in Serviceguard Clusters Summary
Related manuals
Manual 407 pages 39.81 Kb