Forum Discussion

Kelemvor's avatar
Icon for Expert rankExpert
2 years ago

Is there any way to monitor a Windows Cluster?


We have a bunch of Windows clusters in our environment (Hyper-V, SQL, etc).  All of these are setup so things can migrate between the cluster nodes to keep things always running.  Unfortunately, LM handles clusters as if they are an Active/Passive failover pair.  In this type of setup, one server is the Active server doing all the work and the other server is the Passive server that just sits there in case the Active server has issues.  Because of this, LM pulls information on what it considers the Active server, and it only monitors the information about what’s running on that particular server.  It ignores all the other servers which makes monitoring a Windows cluster relatively useless.  When I add a Hyper-V cluster into LM, it sees that Server1 is the Owner server and downloads all the VMs running on that server to monitor them.  It does nothing with Server2, 3, 4, 5, etc even though they all have VMs running on them and are all active all the time.

I contacted LM support and was told that LM can't monitor clusters with more than one active node so there's nothing we can do.  I'm wondering if anyone else has any suggestions how to handle this?

Our main issue is that on our SQL clusters, we can have database instances running on all of the servers in the cluster.  These instances have Windows Services that go along with them.  If DB Instance1 is on SQLServer1, everything is fine.  If Instance1 moves to SQLServer2, then we get alerts because now the Windows services associated with Instance1 stops running on SQLServer1.  The Instance1 service is now running on SQLServer2 because it moved, but LM doesn't see that so we get tons of alerts any time something moves from one server to another.  That's what we're trying to make not happen.

Hopefully that makes sense.  If anyone else has run into this and found a solution, let me know.


1 Reply

  • 2 minutes ago, Kelemvor said:

    LM can't monitor clusters with more than one active node so there's nothing we can do

    Sounds like the support guy taking the easy answer. Or, their answer was, "it doesn't do it out of the box and support is not pro services, so I can't build that for you," which also doesn't solve your problem.

    These could be limitations of the OOTB datasources, in which case, someone will need to build new ones.