4 Replies Latest reply on Nov 13, 2017 11:07 AM by Mark Roberts

    Second Polling Engine and a Fallback

    Kryptic

      Just curious what expected behavior is with a second polling engine.  I have about half of our nodes being monitored by our second polling engine.  I've noticed that when I drop my secondary polling engine (either shut down services or kill the server) my nodes on that server don't report down.  Will the primary SW instance pick up polling for those devices?  Or, did I just not wait long enough to see those nodes "break"?

        • Re: Second Polling Engine and a Fallback
          d09h

          Your polling engine that you have taken down can't write information to the database if you have shut down relevant services or turned the polling server off.  The database won't 'know' status of nodes that are no longer being polled.

           

           

          You have some options for high availability.

           

           

          Here is one link:  https://www.solarwinds.com/high-availability

            • Re: Second Polling Engine and a Fallback
              Kryptic

              Thanks for the info.  I guess I'm less concerned about getting high availability and more wondering what the expected behavior is.  Should(n't) Solarwinds report those nodes that are on the other polling engine as down or unavailable?

                • Re: Second Polling Engine and a Fallback
                  cha-chang!

                  Just to add to d09h's response.  What you are seeing is expected.  Nodes report to the poller you assign it and do not auto-magically find another poller. 

                  • Re: Second Polling Engine and a Fallback
                    d09h

                    http://www.solarwinds.com/documentation/en/flarehelp/orionplatform/content/core-how-orion-works-sw1625.htm

                     

                    relevant part: 

                    The polling engine requests real-time statistics (through the selected polling method, such as SNMP or WMI) from monitored devices in your environment.

                    Polled data is further processed and stored in the SolarWinds Orion database.  (In the scenario you described, poller can't/ won't populate database.)

                     

                     

                    How Orion Platform products work

                    Orion Platform products monitor the health and performance of your network through ICMP, SNMP, WMI, API, and Syslog communication and data collection.

                    A simple centralized Orion Platform deployment includes at least two servers:

                    • Primary server where you install your Orion Platform products. The primary server includes the Main Polling Engine and the Orion Web Console.
                    • A separate server to install the SolarWinds Orion database.

                    The polling engine requests real-time statistics (through the selected polling method, such as SNMP or WMI) from monitored devices in your environment.

                    Polled data is further processed and stored in the SolarWinds Orion database.

                    Data is requested from the database and displayed in Orion Web Console.

                    See Orion Scalability Engine guidelines for information about scaling you Orion Platform product.

                • Re: Second Polling Engine and a Fallback
                  Mark Roberts

                  What you want to do is create/enable the default Orion Polling Engine down alert, so that you can be alerted to the fact that this APE has gone offline and is no longer polling the devices assigned to it.

                   

                  Orion should not put those devices in a down state, as that is not the condition that exists, as if it did receivers of the alerts will begin working the alerts on the basis that the devices are down, when they are not (in all likelihood)