Ad Widget

Collapse

snmp always gets network error for a particulair OID

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • Toonces
    Junior Member
    • Apr 2014
    • 6

    #1

    snmp always gets network error for a particulair OID

    I am trying to create an snmp monitoring item in zabbix to pull data for a specific OID that so we can monitor sessions across a weblogic application. We have been monitoring this using Zenoss for a few years with this setup and I know the complete OID I need to get data from.

    I created the item in Zabbix but always get "SNMP agent item "luminis_total_sessions" on host "XXXXX" failed: first network error, wait for 15 seconds" followed by another error 15 seconds later and the data is never retrieved. This occurs over and over.


    If I run snmpwalk or snmpget on the zabbix server where it is trying to get the data, it works for me every time. I just can't get zabbix to get the data.

    snmpget -r 0 -xxxxx -v2c -On IP.ADDRESS 1.3.6.1.4.1.2021.8.1.101.8
    .1.3.6.1.4.1.2021.8.1.101.8 = STRING: 318

    snmpwalk -r 0 -xxxxx -v2c -On IP.ADDRESS 1.3.6.1.4.1.2021.8.1.101.8
    .1.3.6.1.4.1.2021.8.1.101.8 = STRING: 390


    Any suggestions? by all rights it looks like I should be able to pull the data.

    Thanks!
  • aib
    Senior Member
    • Jan 2014
    • 1615

    #2
    We need some additional information.
    1) Screenshot of Host description
    2) Screenshot of Item description
    3) The value of Timeout= from zabbix_server.conf
    4) Piece of zabbix_server.log with the error message.

    Just from top of my head you can:
    - wrongly describe Host -> SNMP interface
    - forgot to write macros {$SNMP_COMMUNITY}
    - incorrectly use SNMP OID
    - incorrectly setup key settings for OID
    - use default Timeout=3 instead of more usefull Timeout=15 or even =30
    - etc.

    Give me more data - I will give you more correct answer
    Sincerely yours,
    Aleksey

    Comment

    • Toonces
      Junior Member
      • Apr 2014
      • 6

      #3
      Originally posted by aib
      We need some additional information.
      1) Screenshot of Host description
      2) Screenshot of Item description
      3) The value of Timeout= from zabbix_server.conf
      4) Piece of zabbix_server.log with the error message.

      Just from top of my head you can:
      - wrongly describe Host -> SNMP interface
      - forgot to write macros {$SNMP_COMMUNITY}
      - incorrectly use SNMP OID
      - incorrectly setup key settings for OID
      - use default Timeout=3 instead of more usefull Timeout=15 or even =30
      - etc.

      Give me more data - I will give you more correct answer
      Thank You for the quick reply. I checked the easiest thing first and just changed the Timeout value from 5 to 30. I will gather other data as it gets it with the new Timeout value and report back.

      Best Regards,

      Erik

      Comment

      • Toonces
        Junior Member
        • Apr 2014
        • 6

        #4
        Timeout = 30

        Log:
        2338:20140410:110007.541 SNMP agent item "luminis_total_sessions" on host "l4res-p.pima.edu" failed: first network error, wait for 15 seconds
        2349:20140410:110101.912 temporarily disabling SNMP agent checks on host "l4res-p.pima.edu": host unavailable
        2349:20140410:110202.014 enabling SNMP agent checks on host "l4res-p.pima.edu": host became available

        {SNMP_COMMUNITY} is a global variable and works to get all LLD items.

        In the item list I get Not Supported with the error: l4res-p.pima.edu: [-2] Name or service not known

        Thank You again,

        Erik
        Attached Files

        Comment

        • aib
          Senior Member
          • Jan 2014
          • 1615

          #5
          Sorry about that - my eyes so old that I cannot see properly your images.

          When you said {SNMP_COMMUNITY} - did you mean {$SNMP_COMMUNITY} or something else?

          One more change in zabbix_server.conf :
          - check the graph "Zabbix Data Gathering Process busy"
          - if any of pollers busy for 100% - increase the amount of StartPollers in config and restart server
          Sincerely yours,
          Aleksey

          Comment

          • Toonces
            Junior Member
            • Apr 2014
            • 6

            #6
            I apologize for the size of the images. They look very large on my system but I can barely see the posted ones. Perhaps a resizing issue?

            You spotted my issue with the incorrect global variable for SNMP_COMMUNITY. I looked it over a hundred times and didn't catch that I had entered it incorrectly on that particular item. It should indeed be {$SNMP_COMMUNITY} and was not. I checked the graph and found unreachable pollers occasionally hit 100% so I increased the number to 5.

            Thank You for the help!!

            Erik

            Comment

            • aib
              Senior Member
              • Jan 2014
              • 1615

              #7
              Awesome!
              Now you will have different questions , I guess.
              Sincerely yours,
              Aleksey

              Comment

              Working...