8.1 Server Clustering

Server clustering involves several functional components, including failover, load sharing, and common data access. As long as each server is tied to its own storage via parallel SCSI cabling, you cannot efficiently combine the resources of multiple servers. Separating servers from storage, and then reconnecting them on a high-speed network topology, enables storage to be accessed independently. This, in turn, facilitates deployment of storage for data-centric applications and eliminates the access constraints of the server-centric model. If any server can provide a route to the same data, no server is indispensable.

As shown in Figure 8-1, common access to storage resources is supplied by the SAN interconnection. For servers on the same SAN to be clustered, however, cluster software must be run on each server to coordinate and monitor server-to-server communications.

Figure 8-1. Dual-pathed four-server cluster on a SAN

graphics/08fig01.gif

Failover for clustered servers can scale from 2 to more than 30 servers in a single configuration. Failover strategies are vendor-specific but generally require mutual monitoring of server status via a heartbeat protocol. The heartbeat protocol sends keepalive messages between the servers as well as notification of status or required action. The heartbeat can be run in-band over the Fibre Channel or Gigabit Ethernet SAN infrastructure, or out-of-band over a Fast Ethernet connection. In some implementations, a failover algorithm triggers on the loss of an individual application or discrete server component (for example, failure of a host bus adapter or LAN interface). Other servers in the cluster assume the tasks of an individual server that has suffered a failure. If multiple servers are running the same application, you need additional middleware to ensure that servers do not overwrite shared data.

In the example given in Figure 8-1, the SAN infrastructure supplements the high availability of the server cluster by providing dual data paths through the SAN. Each server is provisioned with two SAN interfaces, each running to a separate SAN switch. Similarly, the storage targets are dual-ported to each data path. Although failure of an individual server may trigger failover to one of the surviving servers, failure of an individual SAN switch or link will still provide an alternative path to storage data.

Coordinating the activity of multiple servers and safeguarding file access are not trivial tasks, and that is reflected in the cost of clustering solutions. Compared with the cost of lost business due to down time, however, high availability via clustering is easily justified for mission-critical networks. In addition to clustering capability, integration of dual or multiple pathing into the operating system further facilitates deployment of server cluster solutions.



Designing Storage Area Networks(c) A Practical Reference for Implementing Fibre Channel and IP SANs
Designing Storage Area Networks: A Practical Reference for Implementing Fibre Channel and IP SANs (2nd Edition)
ISBN: 0321136500
EAN: 2147483647
Year: 2003
Pages: 171
Authors: Tom Clark

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net