The Intranet Management System

Managing SNMP traps

IMS supplies a means of monitoring the occurrence of SNMP traps, which are system notifications triggered by a wide range of events, including the following special conditions:

SNMP startup: generated whenever the SNMP server process is started up (see snmpd(1Msnmp))
request authentication failure: generated on failure of a request for data (caused, for example, by the use of an incorrect community string when formulating a request)

For remote node traps to be handled by IMS on the local system, the originating nodes must have been configured to send traps to the local system. This is done using the ``SNMP Configuration Manager''.

For more details of SNMP traps, see ``Traps''.

A log of received trap notifications can be viewed by selecting the Main screen Alarms Trap Alarm Log menu option.

At the top of the resulting screen are buttons (Log now and Log after next reboot) that permit the activation and deactivation of trap logging. Clicking on the first causes trap logging to take immediate effect: the second delays logging until after the next system reboot.

The following is a typical log file entry:

   Community: public.
   Enterprise: experimental.1.42.42.42.42.
   Agent-addr: 128.114.144.103.
   Authentication failure trap.
   Time Ticks: 0.
    
   Name: iso.1.1.1
   Value: NULL

A scroll bar is supplied for manually browsing through the trap entries in the log file, while specific entries may be found using the search facility, which is equipped with Next and Previous buttons.

Reacting to ``down'' nodes

Whenever IMS detects a ``down'' node on the local subnet, that is, the node has failed to react to ping(1Mtcp), a warning mail is sent to the user. This mail identifies the node, and gives the date and time of the failure. This default behavior can be modified, however, using the Main screen Alarms Node Status Alarm menu option.

A screen is displayed that contains some explanatory text and the default action to be taken, as follows:

   echo $status | mailx -s "Intranet Manager:$node is down" user

Overwriting this prompt line causes IMS to execute the given command whenever a node is found to be down. The command will be executed exactly once.