6 Replies Latest reply on Apr 10, 2013 10:02 AM by ddavis44

    SolarWinds LEM locking up daily...

    msterling

      We have recently upgraded SW LEM, and almost daily the web portal will lock up, and my filters will stop logging. When I exit out of it and reopen I am unable to login, and I get a "Unable to connect to Manager: Manager is offline or missing a valid certificate". This happens so often is it very frustrating and the only fix we have found so far is to reboot the TriGeo box. Has anyone else experienced this problem?

        • Re: SolarWinds LEM locking up daily...
          michoudi

          I've had similar experiences which seemed to start around November 2012 but it seemed to occur about once a week. One thing you can try is opening a CLI session to your Manager and restart the Manager service, it should be quicker than a reboot.

           

          Technical support seems to have fixed the problem or at least alleviated it for the time being, still waiting to be sure. I believe the issue was attributed to performance issues caused by too many logs being sent to the wrong Tool/Connectors and logs with no associated Tool/Connector at all. After changing some Connectors, reducing some noise alerts and upgrading Connectors my experience has been much better recently.

          1 of 1 people found this helpful
          • Re: SolarWinds LEM locking up daily...
            gtr_sw

            We have been having the same issue. How big is the environment you are trying to monitor? We have a large infrastructure and we were getting upwards of 2,000 to 5,000 events per minute, an issue a SolarWinds tech said can cause the console to crash. He pointed us to this article:

            SolarWinds Knowledge Base :: LEM Manager crashes after receiving a high number of alerts from Windows 7 or Windows Server 2008

             

            We added the GPO and it has helped significantly.

             

            As for the filters not showing up, try installing the desktop console instead of using the web console. If that is not an option, make sure Flash is updated. I had issues seeing widgets and filters in the web console (thinking LEM was crashing) while other colleagues of mine could clearly see the everything on their desktops.

            • Re: SolarWinds LEM locking up daily...
              msterling

              Restarting the Manager service is definitely the way to solve this problem (for the short term). With our implementation the culprits seem to have been 4-5 internal SolarWinds rules firing almost non-stop (due to the lack of configuration, someone used the default settings). Once I disabled these rules, the lock up seems to have stopped (at least for 2 days). I will update this thread if it stops entirely.

              • Re: SolarWinds LEM locking up daily...
                networkjr

                I have experienced this on several occasions in the past...the LEM was becoming overwhelmed and appears to stall or go offline. In almost every circumstance I was able to determine that the LEM was becoming overwhelmed do to the number of devices that depend on the manager for log normalization. My answer may seem academic, but ensure that you allocated enough resources to your virtual environment to meet this resource demand.  

                • Re: SolarWinds LEM locking up daily...
                  martindl76

                  I would "as networkjr" suggested, check your VM resource allocations. The new upgrade is now compatible with unlimited resource allocation. This was not the case with the previous version where you had to set strict limits. It sounds like a resource usage problem.

                  • Re: SolarWinds LEM locking up daily...
                    ddavis44


                    I also had this issue, we are monitoring over a hundred very busy servers and network devices combined and had 12GB memory allocated to our vm appliance. I opened a case and was told to increase the memory to 16GB because of the heavy load, it seems to have fixed the problem.This is also in line with networkjr's recommendation.