Like this article? We recommend
Services
The primary purpose of the management server is to monitor and manage the
cluster environment. The majority of the management server functions are beyond
the normal management services associated with the Sun Cluster 3.0 software
environment. You can supply these services to other systems as computing
resources in the management server allow.
The management server has enough CPU power and memory resources to
consolidate logging messages from the cluster nodes and to implement the
following services:
- Sun Management Center software
- SunPlex™ software
- Sun StorEdge Component Manager™ software
- Network Time Protocol (NTP)
- AnswerBook2™ software
- Sun Ray™ hardware
The Sun Cluster software cluster control panel (ccp) graphical user interface
(GUI) is installed on the management server to enable easy access to the cluster
node consoles. This GUI enables the SA to execute commands concurrently on all
cluster nodes; however, it does not provide a management interface. The ccp
works with a terminal concentrator (TC) to provide a seamless connection to
serial port consoles. See the following figure for details.
FIGURE 1 Connections Between the Management Server and Cluster Nodes
For increased availability, the management server includes a second internal
SCSI disk to mirror the boot drive. Mirroring using the Solaris Volume Manager
software takes place automatically during the on-site configuration.
To eliminate the need for a keyboard, mouse, and monitor, a TC connects the
cluster nodes to the console port of the management server, thus providing
console access to all cluster nodes.
Configuring the Management Server
You can configure the management server as a central repository for logs and
messages (informational, warning, and error) for each cluster node; this
configuration is similar to Sun Fire™ server domain logging by the System
Service Processor.
Modify the
/etc/syslog.conf files on each cluster node to route
system messages to the management server's SYSLOG service. This routing
allows easy correlation of cluster node events over time to improve cluster
management. In addition, you can review cluster node events on the management
server when a cluster node is down.
It is important to note that the management server itself does not constitute
a critical component in the operation of a cluster. All cluster operations
continue without impact if the management server fails. The default
configuration for Sun Cluster 3.0 software has no dependency on it. In case of a
management server failure, implement an alternative, operational procedure to
access the system messages and console devices of servers to maintain
serviceability of the cluster components.
The Sun Management Center software has a log file filter that allows you to
define patterns of messages that should be highlighted or ignored, and to
quickly identify key messages in the midst of more routine informational
messages. This log file filter enables easy analysis of the correlating log
files.
Every aspect of managing, securing, planning, and debugging a network
involves determining when events happen. Time is the critical element that
allows an event on one network node to be mapped to a corresponding event on
another. In many cases, these challenges can be overcome by the enterprise
deployment of the NTP service. The management server includes the
xntpd(1M) daemon, which is bundled with the Solaris OE software to
provide time synchronization services to all cluster nodes.
){kind=link}