2 Replies Latest reply on Mar 17, 2016 12:27 PM by jemertz

    High CPU on one of several remote pollers


      This weekend we started getting high CPU on one of our remote pollers. We have migrated most of the polling tasks to other pollers, restarted the processes, and rebooted the system. There are only 289 objects being polled by this poller at this time.

      High CPU users include:

      SWJobEngineSvc2.exe (This one stays high in the list even when Solarwinds processes are stopped.)




      Our other pollers are working fine, although this did start shortly after our root SSL certificate expired. Could one device being polled by this, which may have an expired SSL certificate installed somehow be the cause, and if so how would I go about tracking it down? It seems like a stretch to even suggest such a cause, but the timing is close enough that it seems worth investigating.and


      NPM version is 11.0.1 and it has been stable for at least 6 months until this started.

        • Re: High CPU on one of several remote pollers

          We migrated ALL polling off the server and CPU remained at 100%
          Then we stopped Solarwinds altogether.
          Solarwinds Controller Polling Collector was hung. Force killed
          Brought the Solarwinds services back up, still with 0 devices being polled and the CPU returned to 100% utilization.


          Why would Orion use any significant amount of CPU when it has no devices to poll? Our other pollers are at 6-11k elements and few are running over 30%.

            • Re: High CPU on one of several remote pollers

              We found the solution, though not the exact root cause.
              Installing all Windows updates and rebooting (again; it wasn't the reboot alone that solved it!) finally allowed us to start the Orion processes without pegging the CPU. Our assumption is that there was a performance patch included, although we have not checked each update's effects.

              1 of 1 people found this helpful