Generating a config change trap

Faye Ly faye____pedestalnetworks.com
Wed Sep 4 19:19:49 CEST 2002


Frank,

Thank you for the prompt response.  Actually
the trap is generated per SET PDU sent to the
master.  Can you elaborate on the 'poor design'
part?  Our subagents don't live in another box
nor card so from the network manager's view, it
is a single agent.

Thanks.

-faye

-----Original Message-----
From: Frank Fock [mailto:Frank.Fock____t-online.de]
Sent: Wednesday, September 04, 2002 9:47 AM
To: Faye Ly
Cc: agentpp-dl____agentpp.com
Subject: Re: Generating a config change trap


Faye,

Unless the trap should be generated for every SET PDU
sent to the master (which I would not consider this
a good design), putting the trap trigger code in the
master is not helpful.

I would suggest putting it in the
MibLeaf::commit_set_request or
MibLeaf::cleanup_set_request method (it is easier to
put it into commit_set_request but may be you have
requirements so that the latter one fits better).

Please note, do not use NotificationOriginator in an
AgentX++ subagent, use SubAgentXMib::notify instead.

Hope this helps.

Best regards,
Frank

Faye Ly wrote:
> Frank,
> 
> I have a question for you, I need to generate a SNMP
> v2c trap that is triggered per "SNMP set PDU request" is
> completed successfully.  This is to notify the network
> manager that something has changed so it knows if it
> needs to re-fresh it's database.
> 
> Logically it will be best if this trap is generated by
> the master agent as it keeps all the states of a single
> PDU that may or may not be traversing multiple sub-agents.  
> But I need your advice on this.
> 
> Your help will be greatly appreciated.
> 
> -faye
> 






More information about the AGENTPP mailing list