SNMP trap - action based on status.

[kofeyh]

Right, firstly, I have SNMP trap handling working via the perl script located on the wiki.

I've amended it a bit, but otherwise it will send each SNMP trap, against identified hosts, as they occur.

At present, I have a set of triggers defined like this (note, I use regular expressions to simplify triggers):

({Template_SNMP_traps:snmptraps.regexp(@snmp_trap_ down_redundant_event)}=1) &({Template_SNMP_traps:snmptraps.nodata(180)}=0 )

So, in short, a trap is received, then after 180 seconds (3 minutes) the trigger is reset (PROBLEM > OK). This means that even though a problem may still exist, it's status is set as OK; this isn't ideal, but seems to be the usual method.

Otherwise, the behaviour (as I understand it) will mean the problem status will remain indefinitely if snmptraps.nodata isn't set.

What I would like to have is a flow as follows:

TRAP > PROBLEM > TRAP > OK.

In other words, for traps where both a problem trap and an ok trap are sent, I'd like the "ok" trap to act exactly as an agent alert would be, in that a PROBLEM is closed by a subsequent OK.

For example: VMware can send a trap if something generates an Alarm, then a second trap when the Alarm ceases. Typically this is done by setting a string like OldStatus.0: Green, NewStatus.0 Red; this is then reverse when the alarm silences.

I'd like, as in above, to have the first trap cause a trigger to go into PROBLEM, and the second to set it back to OK.

Has anyone managed to figure out how to do this? Is it actually possible?

I get the impression this isn't currently possible.. suggestions?

[yanpei]

I want to control up and down of a ethernet port of cisco router 7609 with snmp ,how can i do?

how to use snmp set command?

thanks.

yanpei

[oct]

Hello,

I'm experiencing a similar problem. In my case, I monitor an APC UPS that sends a trap when on battery event is generated and a new trap when power is restored. I configured a trigger like this:

{Plantilla_APC_UPS_SNMP:snmptraps.str(upsOnBattery )}=1 | {Plantilla_APC_UPS_SNMP:snmptraps.str(powerRestore d)}=0

Although at first it seemd to work, I noticed that, If the UPS sends a trap different from upsOnBattery and from powerRestored (for example, "self-testPassed") the status of the trigger goes to problem. I have to add a restriction at the "powerRestored" part of the OR operation. Maybe something lite:

{Plantilla_APC_UPS_SNMP:snmptraps.str(upsOnBattery )}=1 | ({Plantilla_APC_UPS_SNMP:snmptraps.str(powerRestor ed)}=0 & {Plantilla_APC_UPS_SNMP:snmptraps.str(upsOnBattery ,7200)}=1)

This way, will change the status of the trigger onlyu if in the last 2 hours, a battery event was received). I have to test it.