[SNMP4J] Received response that cannot be matched to any outstanding request

[Walker, Rob]

Frank,

I have tried the suggested adjustments.  I'm still receiving these
messages.  How do you recommend investigating further?  Could it be due
to the nature of UDP?  Do you recommend throttling back the traffic by
dispatching these queries in batches rather than all at once?  And one
more question: What is the difference between cloning the PDU verses
constructing a new PDU.  Could this code be causing a problems? 

this.pduf = new DefaultPDUFactory();

for each of the 2000+ hosts


Tbanks!

Robert Walker 

-----Original Message-----
From: Frank Fock [mailto:fock at agentpp.com] 
Sent: Friday, October 27, 2006 5:06 PM
To: Walker, Rob
Cc: snmp4j at agentpp.org
Subject: Re: [SNMP4J] Received response that cannot be matched to any
outstanding request

Rob,

To get a clean picture, you should set the retries value
to zero and then the timeout value to a large value
(let's say 5min.). If you then still get those messages,
you need to further investigate them.

You design sounds pretty feasible.

Best regards,
Frank

Walker, Rob wrote:
> 
> Frank,
> 
> Thank you for the response.  I have adjusted the timeout and retry
> values, however I'm still dropping packets.  Perhaps I'm not
> implementing correctly.  I'm constructing a single Snmp session and
> TransportMapping.  I then loop through 2000+ instances of a class that
> contains its own PDU and Target.  I'm using the TableUtils getTable to
> walk particular trees.  I'm utilizing the Executor framework
> CompletionService to collect data in parallel. 
> 
> Once again Thank You for the reponse,
> 
> Robert Walker  
> 
> 
> 
> 
> Hi Rob,
> 
> The UDP buffer size is increased on the operating system
> level. So the first choice is to increase the timeout
> and then the retries of the Target instance you use
> to send the request.
> 
> Best regards,
> Frank
> 
> Walker, Rob wrote:
>> I'm receiving:  Received response that cannot be matched to any
>> outstanding request.
>>
>>  
>>
>> A possible fix to this problem was suggested by. 
>>
>>  
>>
>> Have you tried to increase the timeout value?
>> Also consider UDP buffer sizes. That is, when
>> sending many requests in short time, then there
>> could be UDP buffer overflows.
>>  
>> Another common error is reusing PDUs between
>> concurrent requests without cloning the PDU/request
>> ID.
>>  
>>
>>  
>>
>>  
>>
>> How do you increase the UPD buffer size?
>>
>>  
>>
>> I'm using this code snippet to construct the Transport Mapping;
>>
>>  
>>
>>         AbstractTransportMapping transport = null;
>>
>>         
>>
>>         try {
>>
>>             transport = new DefaultUdpTransportMapping();
>>
>>  
>>
>>         } catch (IOException e) {
>>
>>             // TODO
>>
>>         }
>>
>>  
>>
>> Thanks in advance,
>>
>>  
>>
>> Robert Walker
>>
>>  
>>
>>  
>>
>> _______________________________________________
>> SNMP4J mailing list
>> SNMP4J at agentpp.org
>> http://lists.agentpp.org/mailman/listinfo/snmp4j
> 

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