F5 MP fails to resolve the right host name for the SCOM Health Service in an MS Cluster

and also this

 

 

Global Critical: 0 : [03/02/2010 11:24:23]Unhandled global exception: System.InvalidOperationException: Cannot open Service Control Manager on computer 'RMS Cluster Group'. This operation might require other privileges. ---> System.ComponentModel.Win32Exception: The RPC server is unavailable --- End of inner exception stack trace --- at System.ServiceProcess.ServiceController.GetDataBaseHandleWithAccess(String machineName, Int32 serviceControlManaqerAccess) at System.ServiceProcess.ServiceController.GetDataBaseHandleWithConnectAccess() at System.ServiceProcess.ServiceController.GenerateNames() at System.ServiceProcess.ServiceController.get_ServiceName() at System.ServiceProcess.ServiceController.GenerateStatus() at System.ServiceProcess.ServiceController.get_Status() at F5Networks.ManagementPack.DataAccess.OperationsManager.DataSource.DataSourceConnector`2.Connect() at F5Networks.ManagementPack.DataAccess.OperationsManager.DataSource.ManagedDataSourceConnection`3.CoreConnect() at F5Networks.ManagementPack.DataAccess.ManagedConnections.ManagedConnectionBase`1._DoConnectWithRecovery() at F5Networks.ManagementPack.DataAccess.ManagedConnections.ManagedConnectionBase`1.Connect() at F5Networks.ManagementPack.DataAccess.ManagedConnections.ConnectionPoolBase`2._CreatePoolMember() at F5Networks.ManagementPack.DataAccess.ManagedConnections.ConnectionPoolBase`2._InitializeCache() at F5Networks.ManagementPack.DataAccess.ManagedConnections.ConnectionPoolBase`2.Initialize(UInt32 initialOperationSize) at F5Networks.ManagementPack.Services.DeviceMonitor._CreateDataSourceConnection() at F5Networks.ManagementPack.Services.DeviceMonitor._InitializeDataStore() at F5Networks.ManagementPack.Services.DeviceMonitor._CreateDiscoveryManager() at F5Networks.ManagementPack.Services.DeviceMonitor.OnAsyncServiceStart() at F5Networks.Services.WorkerServiceBase._OnAsyncServiceStartHelper()

 

 

and we getting F5 Events error 201 in the eventlogs

 

 

thanks

 

Bao

Hi Julian

 

 

we got this during the install as well

 

 

<03-02-2010 03:02:29> FatalError: The DataSource and ConditionDetection modules required for the F5 Monitoring Service to run could not be loaded by Operations Manager monitoring host. Check the Operations Manager Event Log for errors and manually start the F5 Monitoring Service once any issues have been resolved.

 

 

 

RMS is Windows 2003 R2 X64

 

 

regards

 

Bao

Bao,

 

 

For licensing, you will need a key for both nodes.

 

 

We also had a similar issue during setup on W2K8 RMS cluster:

 

<02-03-2010 11:17:20> FatalError: The DataSource and ConditionDetection modules required for the F5 Monitoring Service to run could not be loaded by Operations Manager monitoring host. Check the Operations Manager Event Log for errors and manually start the F5 Monitoring Service once any issues have been resolved.

 

 

This has been solved with help of Julian:

 

1.Remove the F5 Monitoring Service from the RMS cluster group resources (if it’s still there).

 

2.Uninstall the F5 Management Pack from both cluster nodes.

 

3.Download the latest F5 MP build (v1.6.2.258) from F5 Dev Central: http://devcentral.f5.com/Default.aspx?tabid=217

 

4.Install the F5 MP on the current active cluster node. VERY IMPORTANT: towards the end of the installation, after licensing / accept / OK, where the install screen seem to be stuck (spinning), wait up to 1-2 minutes, and if you still have the setup running, let it run, but open up the Services Control Manager and try to manually start the F5 Monitoring Service. After a few moments, following the service start, the setup should complete successfully. This is a known issue and we’re trying to fix it in our next release. If even after the manual start of the F5 Monitoring Service the setup won’t complete successfully, you should stop here and let us know.

 

5.VERY IMPORTANT: DO NOT add the F5 Monitoring Service to the clustered resources yet and DO NOT fail over to the other cluster node yet.

 

6.Install the F5MP on the other cluster node, and if you experience the same problem with the last screen (after licensing / accept / OK) hung (spinning), stop here and let us know.

 

7.Add the F5 Monitoring Service as a cluster resource (following our DevCentral link) to the RMS cluster group and bring the F5 Monitoring Service online. If it doesn’t come up green, stop here and let us know.

 

8.If you got to this point without errors, and your RMS cluster shows all green, you should be OK now, so you can go ahead and discover a device on the active cluster node.

 

 

Hope it can help.

 

 

Best Regards,

 

 

Vincent

 

I've added a new thread on the current issue for better tracking. As Vincent pointed out, he had similar issues while deploying the F5 Management Pack in on an RMS cluster and they could be worked around until a certain point. Thanks to the significant feedback from you guys, some of these issues have already been resolved in our F5 MP/setup (and will make it into our next release of the F5 MP).

 

 

For you Bao, the problem is that when the F5 MP is trying to resolve the host name of the [virtual] host (i.e. cluster group) running the clustered SCOM Health Service, which is actually the "RMS Cluster Group", it will get this right, but then when trying to query the Health Service status through the Service Control Manager (SCM), it will still use the "RMS Cluster Group" host name, which is obviously wrong. It should rather query the SCM through the FQDN/NetBIOS name of the server hosting the cluster. Apparently this has been working as expected under Windows Server 2008 / R2 but is producing the error that Bao is hitting, on Windows Server 2003. We're looking into fixing this issue as well.

 

 

Unfortunately there's not much we can do to circumvent the problem at this point, so Bao, you'll have to wait for the new F5 MP release which will address this issue. We are working hard to get this new release out the door ASAP and will let you know when this happens.

 

 

Thank you for your support and patience!

 

Julian

Bao,

 

 

We created a fix for the specific problem you were encountering (failing to correctly resolve the network name of the RMS Cluster Group on a Windows Server 2003 MS Cluster). The reason why we didn't hit this problem in the first place was because in our test environment the name of the RMS cluster group was the same with the Network Name associated with the RMS cluster. And this is always true with Windows Server 2008 clustering, where apparently the cluster group name is the actual network name of the cluster group (in our case the RMS Cluster).

 

 

Just to make sure the fix would cover your environment as well (as there might be some variables that we haven't thought about), can you please check the following, and you can send the results directly to managementpack(at)f5.com:

 

 

1. What is the value for the following registry key: HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Operations Manager\3.0 on any of the RMS Cluster nodes?

 

2. What is the name of the MS Cluster hosting the RMS cluster group?

 

3. What is the name of the Network Name resource associated with the RMS Cluster Group?

 

4. What is the host name shown in your SCOM Management Console's Tools > Connect window?

 

 

For 2 & 3 you would go to the Cluster Administration console and get the values from there. For 3 select the RMS Cluster Group and then in the details panel select the "Network Name" resource, right-click > Properties > Parameters tab > Name.

 

 

Providing the answers for these questions would help us better understand your SCOM RMS Cluster environment and possibly assist us with addressing the possible configuration scenarios with RMS clustering.

 

 

Thanks!

 

Julian

 

Bao,

 

 

We released v1.7.0.390 of the F5 MP, which should fix your issues. Please let us know if you have any problems with deploying the F5 MP in your RMS cluster.

 

 

Thanks!

 

Julian