3.5.1 Matrix Server Administration Guide
Chapter 1: Introduction 11
Copyright © 1999-2007 PolyServe, Inc. All rights reserved.
Matrix Server includes several built-in service monitors for monitoring
well-known network services. You can also configure custom monitors
for other services.
A device monitor is similar to a service monitor; however, it is designed
either to watch a part of a server such as a local disk drive or to monitor a
PSFS filesystem. A device monitor is assigned to one or more servers.
Matrix Server provides several built-in device monitors. You can also
define your own custom device monitors.
When a device monitor is assigned to a server, you can select the virtual
hosts that will be dependent on the device monitor. If a device monitor
indicates that a device is not functioning properly on the primary server,
Matrix Server transfers the dependent virtual host addresses from the
primary server to a backup server.
Event Notification
A notifier defines how the matrix handles state transition, error, warning,
and informational messages. You can configure a notifier with a
combination of events and originating matrix entities and then supply a
script that specifies the action the notifier should take. For example, you
could forward events to e-mail or to any other user-defined process.
Matrix Design Guidelines
Be sure to consider the following guidelines when planning the physical
configuration of your matrix.
Server Memory
Memory resources are consumed on each matrix server to manage the
state necessary to preserve the coherency of shared filesystems. For this
reason, the servers in the matrix should have approximately equal
amounts of physical memory. As a guideline, the ratio of memory
between the smallest and largest servers should not exceed 2. For
example, the smallest server could have 1 GB of memory and the largest
server 2 GB. If this ratio is exceeded (such as a 1 GB server and an 8 GB
server, with a ratio of 8), paging can increase on the smallest server to the
extent that overall matrix performance is significantly reduced.