[SNMP4J] V2 Trap Packet Fragmentation

[Amol Rao]

Thanks for the response, Frank. I was bit surprised by NPE since we
are getting it only on one system. I fixed the SNMP4J code in our
environment for Null check and the problem went away.

We are investigating our code in more detail and initial investigation
reveals that there is SNMP get request coming almost at the same time,
which could be causing the NPE here, its not clear how the PDUs (trap
and get) are clashing at this point though.

Isnt it safe to check Null at line 499?

Thanks.

On Wed, Apr 18, 2012 at 6:50 PM, Frank Fock <fock at agentpp.com> wrote:
> Hi,
>
> Another user pointed me to the fact that you
> probalby using SNMP4J 1.x. In that case
> line 499 uses the variable binding list Vector
> of the PDU. If you get a NPE there, then
> you have set the list to null. You will have
> to fix that.
>
> Best regards,
> Frank
>
> Am 17.04.2012 23:20, schrieb Frank Fock:
>> Hi,
>>
>> A NullPointerException at that location can be caused
>> only by a NULL request ID. How you managed to
>> be NULL - I don't know.
>> Normally, SNMP4J takes care to set it. So you must have
>> done something special.
>>
>> A SNMP packet fragementation does not exists on UDP
>> level in the way you described it (two packets with their
>> own header).
>>
>> Best regards,
>> Frank
>>
>> Am 17.04.2012 00:12, schrieb Amol Rao:
>>> ...any thoughts? anyone?
>>>
>>> On Mon, Apr 16, 2012 at 4:09 PM, Amol Rao<xamolrao at gmail.com>   wrote:
>>>> I see the following exception when trap is being sent out. Any pointers?
>>>>
>>>>
>>>> java.lang.NullPointerException
>>>>          at org.snmp4j.PDU.getBERPayloadLengthPDU(PDU.java:499)
>>>>          at org.snmp4j.PDU.getBERLength(PDU.java:436)
>>>>          at org.snmp4j.mp.MPv2c.prepareOutgoingMessage(MPv2c.java:118)
>>>>          at org.snmp4j.MessageDispatcherImpl.sendPdu(MessageDispatcherImpl.java:444)
>>>>          at org.snmp4j.Snmp.sendMessage(Snmp.java:1078)
>>>>          at org.snmp4j.Snmp.send(Snmp.java:878)
>>>>          at org.snmp4j.Snmp.send(Snmp.java:871)
>>>>          at org.snmp4j.Snmp.send(Snmp.java:836)
>>>>          at org.snmp4j.Snmp.notify(Snmp.java:780)
>>>>
>>>>
>>>> Cheers,
>>>> Amol
>>>>
>>>> On Mon, Apr 16, 2012 at 1:05 PM, Amol Rao<xamolrao at gmail.com>   wrote:
>>>>> Hi All:
>>>>>    Just wanted to reach out and see if anyone encountered this situation.
>>>>>
>>>>> I am sending out an SNMP trap (~1K size) from an agent system. When
>>>>> manager receives the trap, its observe that its divided into multiple
>>>>> traps and each trap carries one variable binding - which leads to
>>>>> conclusion that the packets are getting fragmented at the SNMP level,
>>>>> I wonder why this happens.
>>>>>
>>>>> I did wireshark and realized that there is no UDP level fragmentation.
>>>>>
>>>>> Both source (trap generation) and Destination (trap receiver) uses
>>>>> snmp4j-1.11.jar but since on the receiving side, I have logic that
>>>>> expects the trap to come in as one UDP packets, the whole thing breaks
>>>>> apart. Moreover, I observed that its happening on only one system, the
>>>>> same code works on another identical system.
>>>>>
>>>>> Any pointers as to why this could be happening and what is the way around?
>>>>>
>>>>> I am pretty sure that we are not crossing UDP packet limit, not that
>>>>> it would matter at SNMP layer.
>>>>>
>>>>> Thanks.
>>>>>
>>>>> Cheers,
>>>>> Amol
>>> _______________________________________________
>>> SNMP4J mailing list
>>> SNMP4J at agentpp.org
>>> http://lists.agentpp.org/mailman/listinfo/snmp4j
>
> --
> ---
> AGENT++
> Maximilian-Kolbe-Str. 10
> 73257 Koengen, Germany
> https://agentpp.com
> Phone: +49 7024 8688230
> Fax:   +49 7024 8688231
>
> _______________________________________________
> SNMP4J mailing list
> SNMP4J at agentpp.org
> http://lists.agentpp.org/mailman/listinfo/snmp4j