First verify that traps are mapped; this is usually ok, as interface trap mapping is already set up in SPECTRUM for most devices. To check if they are not mapped, look for this event at the VNM model: 0x10801 Unknown alert from device <device> of type <type>. If a trap arrives but the device is not modeled, or the IP is incorrect, look for a 0x10802 event Trap Received from unknown SNMP device with IP <IP>. Reference the Event Configuration Files Guide pdf document for details on trap mapping.
By default, port traps will assert a yellow alarm at the device level. This can be changed by adjusting AssertLinkDownAlarm to True or False
You can set AlarmOnLinkDownTrap=0 never, 1 always, 2 Check Admin (usually the best setting)
When a Link down trap is received, SPECTRUM then does the following:
- Polls the port once to verify if it is down. If it is , it will then generate a Bad Link Detected alarm.
- Sets OutstandingLinkDownTrap=True, and polls the port REGARDLESS of polling settings (reference TS30276).
At the Fault Management view, you can set DisableTrapEvents=True or False
If you set DisableTrapEvents=True there will be no EVENTs (or alarms) from port status traps, which will reduce processing .
Note that in previous versions of SPECTRUM there was an attribute AlarmOnSpecificInterface=True -> This used AlarmOnLinkDownTrapIfTypes to set AlarmOnLinkDownTrap=True If AlarmOnSpecificInterface=False, this set AlarmOnLinkDownTrap=False and PollPortStatus=False. This no longer exists in SPECTRUM 8.1/9.0.
I am getting port status traps into SPECTRUM from my devices, but no alarms?
What are the key trap handling settings inside SPECTRUM other than the event mappings?
What settings can I check if I am not getting port alarms as expected when traps are received?
Also reference TS 30276 Ways to configure Port Status polling in SPECTRUM
(Legacy KB ID CNC TS30278 )