Hi, > You'll notice that everything is polling ok at a 10s interval until > there is a time drift at 2011-11-04 19:32:38,500, from this point on, > all the snmp results return null, which I assume are timeouts. Looking > at the logs, it appears that the request should return a "Not in time > window" exception. Also, the pcap does show a pdu report of > "1.3.6.1.6.3.15.1.1.2.0" for the not in time window exception. Let me > know if you need further information or see something that I'm doing > wrong. I'll see if I can debug further as well.
so now I can see the reason: While on the agent side only 10 seconds have passed, your desktop time increases from 19:21:38,317 to 19:32:38,500 (11 minutes!). So SNMP4j drops all responses of the agent, as these are 11 minutes old, which is more than the 150 seconds time window. Sorry, but SNMPv3 standard does not allow to resynchronize the engine time in this case. My suggestion would be: Create a new thread that continuously sleeps for five seconds and checks if this time drift has happend. If there was a time drift, clear the complete time table. >From the link you mailed before, I would have expected that the desktop time is going too slow. What might happen is that this time drift occurs when windows synchronizes its time with a time server. If you reduce the time between two time synchronizations, so that the time drift is below two minutes, SNMPv3 could handle this by itself and you would not need the additional thread. Regards, Jochen _______________________________________________ SNMP4J mailing list SNMP4J@agentpp.org http://lists.agentpp.org/mailman/listinfo/snmp4j
