Sending snmp trap on stack member failure

[Guiscardo]

Hello,

when operationg switches (3850) in a stack and a stackmember fails, I'd like like the stackmaster (or the remaining Switch) to send an snmp trap, reporting the failure of the stack member.

Currently I have the problem, that in my test environment I only get a notification that something was removed, when I reload a stack member on purpuse.

If I simply pull the plug from a stack member to simulate a failure, nothing happens.

It is only when the switch has rebooted that I get messages, that something was added.

Please note that I do not want to use a network manager to activly "ask" the switches if something has chanced.

So is there a possibility to get the remaining switches to actively report when a stackmember goes down?

Kind Regards

EDIT: "snmp-server enable traps stackwise" appears in running config.

[Georg Pauwen]

Hello,

what happens if you explicitly enable the traps for:

snmp-server enable traps stackwise member-removed

You could also try to use an EEM script to manually enable an SNMP trap when a stack member is removed. Try the script below, the OID is for 'cswStackMemberRemoved'. 

snmp-server enable traps event-manager
!
event manager applet SNMP_STACK_MEMBER_REMOVED
event snmp oid 1.3.6.1.4.1.9.9.500.1.2.2.1.1 get-type exact entry-op eq entry-val "1" exit-op eq exit-val "2" poll-interval 5
action 1.0 snmp-trap strdata "Stack Member Removed"

[Guiscardo]

Hello, snmp-server enable traps stackwise member-removed is something that I already tried. To no avail I might add. I tried the script you suggeted, however the behaviour of the network manager stays the same. When I pull the plug from the swich and put it back in, I only get Traps telling me that FRUs were inserted. For some reason I fail to fathom no traps are send that the switch was powered down or anything was removed.

When I issue a "reload slot 2" I get a

FRU with name 111.111.111.111, description .1.3.6.1.4.1.9.9.117.2.0.4, physical index 2053 was removed

When the power plug is pulled nothing similar is generated.

[marce1000]

>...Please note that I do not want to use a network manager to activly "ask" the switches if something has chanced.

 From a monitoring viewpoint albeit with tools such as NAGIOS , Zabbix or others I still think this is the best solution. 1) Who does what when the stackmaster itself fails ? 2) 'Failure' as such is not uniquely defined can  also be due to external circumstances such as   power failure (for an individual stack member), albeit external power or not (e.g.). 

Therefore in our network I do active monitoring of the stack-members at various places with NAGIOS.

M.