SNMP Reference Guide

28 Introduction
SNMP agents implement a hash/digest mechanism to prevent unauthorized SNMP Set operations.
One limitation of this practice is that only server administrator-developed SNMP management
applications have the capability to support the hash/digest mechanism.
NOTE: The default SNMP agent configuration usually includes a SNMP community name such as public. For
security reasons, change the SNMP community names from their default values. For information about changing
SNMP community names, see the Dell OpenManage Server Administrator User’s Guide. This guide is available on
the Dell Support website at support.dell.com and on the Dell PowerEdge Installation and Server Management and
Dell PowerEdge Documentation CDs. For additional guidelines, see the Securing an SNMP Environment article,
dated May 2003, in the Dell Power Solutions magazine. This magazine is also available at
www.dell.com/powersolutions.
NOTE: In Dell OpenManage Server Administrator version 5.2, SNMP Set operations are disabled by default. Server
Administrator provides support to enable or disable SNMP Set operations. For more information on enabling or
disabling SNMP Set operations in Server Administrator, see the Dell OpenManage Server Administrator User’s
Guide or the Dell OpenManage Server Administrator Command Line Interface User's Guide.
Management Actions
Management actions can be performed using the SNMP Set command. These actions can consist of
configuring a phone number for the system’s owner, rebooting a system, or changing the asset tag of the
system. See the previous section, "SNMP Security," for limitations on Set operations.
SNMP Traps
SNMP is frequently used to monitor systems for fault conditions such as temperature violations,
hard drive failures, and so on. Management applications can monitor for these conditions by polling the
appropriate OIDs with the Get command and analyzing the returned data. This method has its
drawbacks. If it is done frequently, significant amounts of network bandwidth can be consumed. If it is
done infrequently, the response to the fault condition may not occur in a timely fashion. SNMP traps
avoid these limitations of the polling method.
An SNMP trap is an asynchronous event indicating that something significant has occurred. This is
analogous to a pager receiving an important message, except that the SNMP trap frequently contains all
the information needed to diagnose a fault.
Two drawbacks to SNMP traps are that they are sent using UDP, which is not a guaranteed delivery
mechanism, and that they are not acknowledged by the receiver.
An SNMP trap message contains the trap’s enterprise OID, the agent IP address, a generic trap ID, the
specific trap ID, a time stamp, and zero or more variable bindings (varbinds). The combination of an
enterprise OID and a specific trap ID uniquely identifies each Server Administrator-defined trap.
A varbind consists of an OID and its value and provides additional information about the trap.
In order for a management system to receive SNMP traps from a managed system, the node must be
configured to send traps to the management system. Trap destination configuration is dependent on the
operating system. When this configuration is done, a management application on the management
system can wait for traps and act on them when received.
For a list of traps supported by the server administrator SNMP subagent, see "Traps."