7 Replies Latest reply on Dec 11, 2012 11:09 AM by Sohail Bhamani

    10.4 rogue pollers

    rfletcher

      I just upgraded to 10.4 and discovered I have 3 cisco nodes reporting like they are apc ups's and are triggering alert on input/output voltage and battery health status. Has anyone else encountered an issue like this while upgrading?

       

      I have verified that the pollers are not assigned to the cisco nodes, yet the charts and displays continue to appear on the node details page.

        • Re: 10.4 rogue pollers
          Sohail Bhamani

          I think what you are seeing is not related to any rogue pollers but related to a new feature in NPM 10.4.  This new feature gets the network hardware health information including what you mentioned.  To disable this, you can click on list resources for the node in question and you should see a Network Hardware Health (or something similar) checkbox as the first object that can be checked.

           

          Sohail Bhamani

          Loop1 Systems

          http://www.loop1systems.com

          1 of 1 people found this helpful
            • Re: 10.4 rogue pollers
              RichardLetts

              This is going to be a general gripe... in large installations this sort of thing is impractical.

              Any time someone says 'you just click list resources and turn it off' when you have 1000 such things it just ain't gonna happen without someone suffering from RSI. In addition, for devices with over a thousand interfaces, doing a list resources on it just to turn OFF something is a little resource intensive.

                • Re: 10.4 rogue pollers
                  Sohail Bhamani

                  I agree completely.  I think the PM folks do know they need to enhance large environments abilities to do such an activity across a large number of devices without having to get lost in the minutia.  Hopefully we will see some of this as they keep pumping out the versions.

                   

                  Sohail Bhamani

                  Loop1 Systems

                  http://www.loop1systems.com

                    • Re: 10.4 rogue pollers
                      michal.hrncirik

                      Hi everyone,

                       

                      this HW health features has became interesting. We did not receive feedback like this during our beta or RC phase and we are working on feasible solution for everyone. The fact is that our HW health just returns statuses of sensors or HW health software of router, switch or other networking device. If device returns sensor status "warning" NPM shows that as Node warning state.

                      Some users suggest to make it configurable so you as the user may decide which sensor status will be represented as Node warning or Critical. Unfortunately this doesn't solve problem that Sohail has mentioned - large environments. You would have to define that for all your vendors/type of devices which doesn't seem to solve "easy to use problem".

                       

                      I have also another question, in case of rfletcher - are those statues false or they are providing correct value and you just do not want to pair HW health warning with Node/interface availability?

                       

                      thanks for sharing your thoughts.

                      Michal

                    • Re: 10.4 rogue pollers
                      Víťa Tauer

                      The 'mass disable' can be done via SQL.

                       

                      UPDATE APM_HardwareInfo SET IsDisabled=1

                       

                      should do the trick (disabling HWH on everything).

                    • Re: 10.4 rogue pollers
                      rfletcher

                      Thank you for your advice and help, actually turned out the pollers did get assigned to those nodes, however they were not showing via the web console but when I went into the nodes via Orion Universal Device Poller they were showing as assigned to the specific nodes.