Re[2]: variables in traps

2004-07-19 Thread Norio Nakamoto
David, Thank you for your answer. I read through RFC 1157 and I understand taht Trap-PDU defines variable bindings as 'VarBindList' exactly same as Get/Response-PDU, So it should be implmented in the same way. I will follow this for our agent implementation. Norio Nakamoto On Sat, 17 Jul 2004 0

Re: variables in traps

2004-07-17 Thread David T. Perkins
HI, How SNMP messages are formed are defined in RFC 3416 - SNMPv2c and SNMPv3 RFC 1157 - SNMPv1 Notice that both have a varBindList returned, just like a response for a GET. It's not an idea or convention, it's the way the protocol is defined. Now, there has been people that have not unders

Re: variables in traps

2004-07-16 Thread Norio Nakamoto
One more thing. Can you tell me which RFC gives this idea or convention? On Fri, 16 Jul 2004 18:14:25 -0700 "David T. Perkins" <[EMAIL PROTECTED]> san wrote: > HI, > > Variables in notifications (TRAPv1, TRAPv2, INFORM) always have instance > sub-ids. If they are scalars, it's always 0. For col

Re: variables in traps

2004-07-16 Thread Norio Nakamoto
David, Thanks for such a quick answer. So you should treat the variables in both notifications and reqeust/ response with the same manner. Is that correct? Then the agent I mentioned seems right. Norio Nakamoto On Fri, 16 Jul 2004 18:14:25 -0700 "David T. Perkins" <[EMAIL PROTECTED]> san wrot

variables in traps

2004-07-16 Thread David T. Perkins
HI, Variables in notifications (TRAPv1, TRAPv2, INFORM) always have instance sub-ids. If they are scalars, it's always 0. For columnar, you have to look at the INDEX or AUGMENTS clause of the row definition. A well written DESCRIPTION clause will make it obvious which instance is returned. You NEV