snmp stops reponding

rlaskyrlasky
edited May 2011 in InfraSensing Sensors
I have a hardware version 2 temperature sensor on firmware 1.20, and I am trying to monitor by polling via snmp. It works fine for about an hour, and then it stops responding. The web interface still works, but no reply from snmp polling.



I saw on another post (http://community.serverscheck.com/display_topic_threads.asp?ForumID=8&TopicID=1640&PagePosition=1) that there was a version 1.30 firmware that would address the problem, but the sensor will not boot back up after I try to update the firmware and I have to hard reset it and it goes back to 1.20



Any ideas on how to keep this thing responding to snmp?

Comments

  • AdministratorAdministrator
    Firmware 1.3 was for hardware release 1 and not hardware release 2.



    What is the frequency polling set for the sensor? Every minute?



    What monitoring software are you using? If not the ServersCheck Monitoring Software, then are you SNMP sessions properly closed after each request?



    We quickly ran a test on a temperature sensor with hardware version 2 sensor gateway polling through SNMP some 205000 times the sensor in 22 minutes. We'll keep it running for a while to see if it hangs like you describe.
  • AdministratorAdministrator
    Running now for a bit more than 2 hours straight with 1663278 SNMP polls performed.



    Here is the Perl code used to create the test:

    use Net::SNMP;



    my $SNMP_oid=".1.3.6.1.4.1.17095.3.6.0";

    my $SNMP_port=161;

    my $SNMP_host="10.0.0.34";

    my $SNMP_community="public";



    print "# ".localtime()." Start tests from SensorGateway on $SNMP_host - $SNMP_communityn";



    my $i=0;



    while (1>0) {

    my ($session, $error) = Net::SNMP->session(

     -hostname => $SNMP_host,

     -community => $SNMP_community,

     -port => $SNMP_port

    );



    if (!defined($session)) {

     print "# ".localtime()." SNMP failedn";



    }

    else {



    print "# $i ".localtime()." Connected to SensorGatewayn";



    my $result = $session->get_request(

    -varbindlist => [$SNMP_oid]

    );



    if (!defined($result) or $result->{$SNMP_oid} eq "") {

    print "# $i ".localtime()." Couldn't retrieve value for OIDn";

     $session->close;

     #exit;

    }

    else {

    my $SNMP_status="";

     $SNMP_status=$result->{$SNMP_oid} ;

     $SNMP_status=~ s/r//g;

     $SNMP_status=~ s/n//g;



     print "# $i ".localtime()." SNMP returned string: ".$SNMP_status." n";  

    }



    }

    $i++;

    $session->close;

    }
  • rlaskyrlasky
    I'm using nagios to do the monitoring. I used the nagios code that is on the serverscheck website with a few modifications. I was checking every 2 minutes.



    I will try your perl code, as I dont know if any of my mods may have made it so that the snmp session wasnt closed properly.



    I will let you know how it goes.
  • rlaskyrlasky
    Well, still no good, even using the perl code above directly. It says up for maybe 2 hours, then SNMP stops responding.



    Maybe I just got a bad unit? Is there a warranty?
  • AdministratorAdministrator
    It may indeed be a bad unit. There is a warranty - see link at bottom of our page.



    For a RMA contact support by going to the SUPPORT section and loga RMA request by referring to this thread.



    They will test your unit and run it in a network to test. It could always be a network issue too.
This discussion has been closed.