Traps

Although SNMP is normally used in a synchronous manner in which systems poll each other periodically, it does include a mechanism for asynchronous events. In SNMP, these are referred to as ``traps.'' Traps are sent by agents to management stations to signal the occurrence of an asynchronous event.

Because traps are asynchronous messages signaling an event that may require attention, there are no responses from the management station to them. An agent may generate seven kinds of trap:

coldStart

Implies that the sending agent is reinitializing itself with significant changes in its configuration, for example, after a reboot.

warmStart

Implies that the sending agent is reinitializing itself with no changes in its configuration.

linkDown

Implies that a failure in one of the communication links of the sending agent.

linkUp

Implies that one of the sending agent's communication links has been restored to normal running status.

authenticationFailure

Implies that an instance of authentication failure has occurred at the sending agent.

egpNeighborLoss

Implies that an EGP peer relationship of the sending agent's EGP protocol has been lost.

enterpriseSpecific

Implies that an enterprise-specific event has occurred.

See also:

• ``Authentication''

• ``SNMP configuration''

• ``Configuring the SNMP ``trap'' group''