[SNMP4J] Can you help me work it out?

[Frank Fock]

Setting engineBoots to zero violates the SNMPv3
standard. A resync of times is then not possible
without manual intervention.
Please do not ask why (it is a security requirement).
This is a really frequently asked question!

What is the manufacturer of the device that resets
its engine boots counter to zero?

Frank

On 06.02.2011 13:23, ning zheng wrote:
> Frank,
> Thanks for your quick response. let me descript it more details. I am a 
> new hand for SNMPV3. sorry for bother you.
> 
> In this case, A is an element in network, B is the manager server. B 
> will GET/SET on A, A will send snmptrap to B.
> sometimes A is restarted, then it send the snmptrap to B with 
> engineBoots and engineTime zero to do re-sync.
> Normally, after A get the right engineBoots and engineTime from B, this 
> error(NotInTimeWindow) will not happen, but occasional. it will enter 
> the error flow I listed in previous mail.
> 
> Also, I attached the message flow file(NotInTimeWindow.zip) for this 
> error, you can unzip it and open with Wireshark(Filter:snmp).
> A: 188.116.247.62
> B: 172.18.1.21
> would you like check it? Thanks+++!
> 
> Best Regards,
> Stephen
> 
>  > Date: Sun, 6 Feb 2011 12:03:05 +0100
>  > From: fock at agentpp.com
>  > To: snmp4j at agentpp.org
>  > Subject: Re: [SNMP4J] Can you help me work it out?
>  >
>  > Hi,
>  >
>  > For a TRAP or NOTIFICATION PDU, the sender is
>  > authoritative. Thus, A should include its
>  > own engine ID, boots and time.
>  >
>  > If your description is correct, then
>  > NET-SNMP has a bug (which I do not
>  > think in this case).
>  >
>  > Thus, please check your implementation/configuration.
>  >
>  > Best regards,
>  > Frank
>  >
>  > On 06.02.2011 10:16, ning zheng wrote:
>  > >
>  > > We use SNMP4J in our SNMP management application and we've recently 
> run into an issue
>  > > The issue is as below: (A: net-snmp B:SNMP4j)
>  > >
>  > > 1. A-->B: A send a snmptrap to B with engineBoots and engineTime 
> values with zero. -- send trap (engineId set but engineBoots and 
> engineTime set to 0)
>  > > 2. B-->A: NotInTimeWindow Report (engineId, engineBoots, and 
> engineTime set)
>  > > 3. A-->B: A re-send the snmptrap with right engineId, engineBoots, 
> and engineTime from B -- send trap (engineId, engineBoots and engineTime 
> set)
>  > > 4. B-->A: NotInTimeWindow Report (engineId, engineBoots, and 
> engineTime set)
>  > > 5. Repeat Step 3 and 4
>  > >
>  > > My question is why B always report NotInTimeWindow(in 4,6,8) even A 
> has set the right engineId, engineBoots and engineTime.
>  > > I check the RFC3414 about NotInTimeWindow, but not found the root 
> cause, seems the trap from A should be accepted. Is there a SNMP4j bug?
>  > >
>  > > Thanks very much for any feedback!
>  > > Stephen
>  > >
>  > > _______________________________________________
>  > > SNMP4J mailing list
>  > > SNMP4J at agentpp.org
>  > > http://lists.agentpp.org/mailman/listinfo/snmp4j
>  >
>  > --
>  > AGENT++
>  > http://www.agentpp.com
>  > http://www.snmp4j.com
>  > http://www.mibexplorer.com
>  > http://www.mibdesigner.com
>  >
>  > _______________________________________________
>  > SNMP4J mailing list
>  > SNMP4J at agentpp.org
>  > http://lists.agentpp.org/mailman/listinfo/snmp4j

-- 
AGENT++
http://www.agentpp.com
http://www.snmp4j.com
http://www.mibexplorer.com
http://www.mibdesigner.com