Adding Hyper-V Cluster to VMM fails with code 13805

Hi all,

have SCVMM RTM.

have a 2 node Hyper-V cluster, both nodes are fully patched.

When trying to the cluster into SCVMM I get an error: 13805

"Unable to contact cluster service on SCLU680.xxx"

Recommende action: Ensure that the provided name is a cluster name or the computer name of a node in the cluster, and that the cluster is listed in Active Directory Domain Services. For a newly created cluster, you might need to wait for the cluster name to be added to Active Directory Domain Services.

Strange thing is that I was successfull 2 weeks earlier in adding this cluster to the beta release of SCVMM ...

For completeness: I also check the " Reassociate host with thsi virtual machine manager server" ...

Anyone an idea ?

BR, Peter

November 10th, 2008 5:34pm

Hi Peter,

Can you check that the cluster service is running on both nodes? Also, can you verify that Firewall exceptions areopen for "Failover Clusters" and "Failover Cluster Management" for the nodes?

Thanks,
Travis

Free Windows Admin Tool Kit Click here and download it now
November 11th, 2008 4:36am

I'm having the exact same problem, and it's slowly driving me batty. A couple of datapoints:

- Cluster services is running on both nodes.

- Problem also happens if I have a one-node cluster.

- Both firewall exceptions are open.

- Also happens with the firewall turned off completely

This is with VMM2008 RTM and 2k8 RTM, and a 2k3 DC (if that matters in the slightest)

Help?

November 11th, 2008 5:31am

VMM2k8 does not manage WS2k3 (Virtual Server) clusters directly. To manage these clusters you will need to add them as standalone hosts and will manage them as such. Currently, VMM supports full management only for WS2k8 Hyper-V clusters.

Travis

Free Windows Admin Tool Kit Click here and download it now
November 11th, 2008 5:55am

If you'll reread, you'll see that only my DC (which isn't in a VM or anything) is 2k3, everything else is 2k8 (w/ Hyper-V). I only mentioned the DC because the error message had the "you may have to wait for the cluster to be registered..." bit in it.

November 11th, 2008 6:05am

Oh okay, my mistake. Are you using the Add host wizard or Powershell to add the cluster? In the wizard, does entering in a node name resolve to the cluster? A message should appear asking if you want to add a failover cluster.

Could you run the Discover-Cluster commandlet in Powershell and show what is returned:

get-vmmserver <VMM Server>
$creds = Get-Credential
Discover-Cluster <ClusterName> -Credential $creds

Lastly, could you then get traces so that I can better see what is happening? Tracing directions: http://blogs.technet.com/chengw/archive/2008/05/08/how-to-collect-scvmm-traces.aspx

Thanks,
Travis

Free Windows Admin Tool Kit Click here and download it now
November 11th, 2008 7:36am

Hokay. Lets see.

I'm using the add host wizard to add the hosts.

Trying to run the powershell command sequence you suggest above gives effectively the same error:

Discover-Cluster : Unable to contact cluster service on J-VM.mydomain.com. (Error ID: 13805)

Back to the VMM GUI, entering the second node name (J-VM02) results in it resolving to the cluster. Entering the first node name (J-VM01) actually gives an error:

Access has been denied while contacting the server J-VM01.mydomain.com.

1. Verify that the specified user account has administrative privileges on J-VM01.mydomain.com.

2. Verify that DCOM access, launch, and activation permissions are enabled on J-VM01.mydomain.com for the Administrators group. Use dcomcnfg.exe to modify permissions, and then try the operation again.

ID: 406

I actually did pull up dcomcnfg to check the permissions. The error doesn't say which objects to check the permissions on, but I checked the permissions on "Virtual Machine Management" and "Virtual Machine Worker Process" and both seemed to be correct for group Administrators.

Entering the second node name (J-VM02) has it resolve the cluster. If I tell it to continue and accept all the defaults from there, it starts the host adding job, whichimmediately fails with:

Error (13805)

Unable to contact cluster service on J-VM.mydomain.com

Recommended Action

Ensure that the provided name is a cluster name or the computer name of a node in the cluster, and that the cluster is listed in Active Directory Domain Services. For a newly created cluster, you might need to wait for the cluster name to be added to Active Directory Domain Services.

As far as tracing goes, I don't see an obvious way to attach a file or send to you privately, so I'm going to pick out what seems like the relevant bits in the trace and just post those, in the name of brevity. Full trace available on request.

[2404] 0964.1380::11/11-16:11:59.313#09: DiscoverCluster.cs(87): Starting discovery for cluster J-VM.mydomain.com
[4612] 1204.1298::11/11-16:11:59.313#21:ClientCache.cs(749): Waiting on 1 objects...
[2404] 0964.100C::11/11-16:11:59.328#21:EventClientConnection.cs(239): Server session uuid:e875e18b-e3cd-4966-bd27-d4530bb32263;id=1 - Sending 1 events
[4612] 1204.01A8::11/11-16:11:59.328#21:Callback.cs(50): Client uuid:7089dcaf-72e0-46b5-b2de-83a592f184bb;id=1 - 1 events
[4612] 1204.01A8::11/11-16:11:59.328#21:ClientCache.cs(599): New object: abe6449b-4e39-4e03-b5ac-e6d0af522229
[4612] 1204.01A8::11/11-16:11:59.328#21:Callback.cs(53): Client uuid:7089dcaf-72e0-46b5-b2de-83a592f184bb;id=1 - events processed
[2404] 0964.1380::11/11-16:11:59.531#09:ClusterHelper.cs(473): Error getting cluster details for J-VM.mydomain.com
[2404] 0964.1380::11/11-16:11:59.531#09:ClusterHelper.cs(473): System.Management.ManagementException: User credentials cannot be used for local connections
[2404] at System.Management.ManagementException.ThrowWithExtendedInfo(ManagementStatus errorCode)
[2404] at System.Management.ManagementScope.InitializeGuts(Object o)
[2404] at System.Management.ManagementScope.Initialize()
[2404] at Microsoft.VirtualManager.DB.Adhc.ClusterHelper.GetClusterObjects(String scopeString, String className, String computerName)
[2404] 0964.1380::11/11-16:11:59.531#19:Task.cs(246): Task failed error ClusterDoesNotExist (13805) with exception Microsoft.VirtualManager.Utils.CarmineException: Unable to contact cluster service on J-VM.mydomain.com.
[2404] Ensure that the provided name is a cluster name or the computer name of a node in the cluster, and that the cluster is listed in Active Directory Domain Services. For a newly created cluster, you might need to wait for the cluster name to be added to Active Directory Domain Services.
[2404] at Microsoft.VirtualManager.Engine.Adhc.DiscoverClusterTask.DiscoverNewCluster()
[2404] at Microsoft.VirtualManager.Engine.Adhc.DiscoverClusterTask.RunSubtask()
[2404] at Microsoft.VirtualManager.Engine.TaskRepository.SubtaskBase.Run()
[2404] at Microsoft.VirtualManager.Engine.TaskRepository.Task`1.SubtaskRun(Object state)
[2404] *** Carmine error was: ClusterDoesNotExist (13805)
[2404] *** J-VM.mydomain.com ** |TaskID=ABE6449B-4E39-4E03-B5AC-E6D0AF522229

So it does correctly detect that J-VM02 is part of the J-VM cluster. The interesting parts seem to be..

System.Management.ManagementException: User credentials cannot be used for local connections

... and ...

Task failed error ClusterDoesNotExist (13805) with exception Microsoft.VirtualManager.Utils.CarmineException: Unable to contact cluster service on J-VM.mydomain.com.Ensure that the provided name is a cluster name or the computer name of a node in the cluster, and that the cluster is listed in Active Directory Domain Services. For a newly created cluster, you might need to wait for the cluster name to be added to Active Directory Domain Services.


The first of those almost implies to me that one shouldn't be running the VMM master server on one of the machines in the cluster -- is that actually the case? I suppose I wouldn't be surprised if that were the case, but I don't remember seeing that documented anywhere, either.

The second (ClusterDoesNotExist), J-VM gets registered in DNS and I can see the computer object for it in AD, so I'm assuming it's correctly registered in AD and the like.

I think that covers most everything. I'll try moving my VMM server off of the cluster and see if that makes things happier, in the meantime. For me, this is somewhat sub-optimal (since I want clustering for quick migration more than redundancy), but if that's what is going on, at least I'll know.

Thanks much!

November 11th, 2008 7:45pm

On any node in the cluster, can you run cluster res command in an elevated command window and see if cluster name is owned by the same machine that has VMM installed? That might the one of the reasons for hitting this error. If that is the case, you can run cluster group cluster group /move command to move the cluster name to some other node in the cluster and then retry cluster addition to see if it works.

Hope this helps.

Regards,

Shon

Free Windows Admin Tool Kit Click here and download it now
November 11th, 2008 11:03pm

Travis,

I tried the above:

PS C:\Windows\System32> Discover-Cluster sclu680.xxx -Credential $
creds
Discover-Cluster : Unable to contact cluster service on sclu680.xxx. (Error ID: 13805)


To restart the job, run the following command:
PS> Restart-Job -Job (Get-VMMServer axscvmm.xxx | Get-Job | where
{ $_.ID -eq "{fbdec20a-0e14-4748-9c74-437907c0233e}"})
At line:1 char:17
+ Discover-Cluster <<<< sclu680.xxx -Credential $creds
PS C:\Windows\System32> Discover-Cluster sdca680.xxx -Credential $
creds


Name : sclu680.xxx
ClusterNodes : {SDCA680.xxx, SDCA681.xxx}
HAFileServers : {}
OperatingSystem : Windows Server 2008 Enterprise
Stack : HyperV
OperatingSystemVersion : 6.0.6001
ServerConnection : Microsoft.SystemCenter.VirtualMachineManager.Remoting.
ServerConnection
ID : d5940424-93bf-48a5-af2e-8dce79b08811
MarkedForDeletion : False
IsFullyCached : True

So when I use cluster name, it fails with same error but when I use node name, it works. Even using the other physical node name it works. IN my situation SCVMM is not installed on any of the clusternodes...

Peter

November 12th, 2008 3:17pm

An extract from the debugging session I did:

00005739195.61203003[1828] 0724.0544::11/12-12:50:06.438#09:ClusterHelper.cs(467): Error getting cluster details for sclu680.xxx
00005740195.68386841[1828] 0724.0544::11/12-12:50:06.501#09:ClusterHelper.cs(467): System.Runtime.InteropServices.COMException (0x800706BA): The RPC server is unavailable. (Exception from HRESULT: 0x800706BA)
00005741195.68386841[1828] at System.Runtime.InteropServices.Marshal.ThrowExceptionForHRInternal(Int32 errorCode, IntPtr errorInfo)
00005742195.68386841[1828] at System.Management.ManagementScope.InitializeGuts(Object o)
00005743195.68386841[1828] at System.Management.ManagementScope.Initialize()
00005744195.68386841[1828] at Microsoft.VirtualManager.DB.Adhc.ClusterHelper.GetClusterObjects(String scopeString, String className, String computerName)
00005745195.68627930[1828] 0724.0544::11/12-12:50:06.516#19:Task.cs(246): Task failed error ClusterDoesNotExist (13805) with exception Microsoft.VirtualManager.Utils.CarmineException: Unable to contact cluster service on sclu680.xxx.
00005746195.68627930[1828] Ensure that the provided name is a cluster name or the computer name of a node in the cluster, and that the cluster is listed in Active Directory Domain Services. For a newly created cluster, you might need to wait for the cluster name to be added to Active Directory Domain Services.
00005747195.68627930[1828] at Microsoft.VirtualManager.Engine.Adhc.DiscoverClusterTask.DiscoverNewCluster()
00005748195.68627930[1828] at Microsoft.VirtualManager.Engine.Adhc.DiscoverClusterTask.RunSubtask()
00005749195.68627930[1828] at Microsoft.VirtualManager.Engine.TaskRepository.SubtaskBase.Run()
00005750195.68627930[1828] at Microsoft.VirtualManager.Engine.TaskRepository.Task`1.SubtaskRun(Object state)
00005751195.68627930[1828] *** Carmine error was: ClusterDoesNotExist (13805)
00005752195.68627930[1828] *** sclu680.xxx ** |TaskID=B4011BA5-D647-4931-94A5-2D6D979C5E68
00005753195.68728638[1828] 0724.0544::11/12-12:50:06.516#04:CarmineObjectLock.cs(358): Releasing all locks for task b4011ba5-d647-4931-94a5-2d6d979c5e68|TaskID=B4011BA5-D647-4931-94A5-2D6D979C5E68
00005754195.68937683[1828] 0724.0544::11/12-12:50:06.516#19:RunningTaskDbWriter.cs(464): MarkTaskComplete --> TaskResult: ClusterDoesNotExist (13805)|TaskID=B4011BA5-D647-4931-94A5-2D6D979C5E68
00005755195.69744873[1828] 0724.0544::11/12-12:50:06.516#25Tongue TiedqmSession.cs(92): Getting Instance for session Engine
00005756195.69754028[1828] 0724.0544::11/12-12:50:06.516#25Tongue TiedqmSession.cs(96): Found session Engine
00005757195.69766235[1828] 0724.0544::11/12-12:50:06.516#25Tongue TiedqmSession.cs(360): AddToStream - session:Engine datapointBig SmileATAID_VMM_BBOS_NUMBER_OF_FAILED_TASKS
00005758195.70579529[1828] 0724.0544::11/12-12:50:06.532#21:EventManager.cs(94): PublishUpdate b4011ba5-d647-4931-94a5-2d6d979c5e68
00005759195.70602417[1828] 0724.0544::11/12-12:50:06.532#20SurprisebjectAuthorizationManager.cs(94): Accessibile=True, Profile=Administrator, Object Type=Job, ID=b4011ba5-d647-4931-94a5-2d6d979c5e68
00005760195.70610046[1828] 0724.0544::11/12-12:50:06.532#21:EventClientConnection.cs(142): Publish uuid:b502c363-0605-477a-b790-e28fae3d7e11;id=1 Job Update
00005761195.71710205[1828] 0724.0544::11/12-12:50:06.548#21:EventClientConnection.cs(239): Server session uuid:b502c363-0605-477a-b790-e28fae3d7e11;id=1 - Sending 1 events
00005762195.72132874[3964] 0F7C.0B40::11/12-12:50:06.548#21:Callback.cs(53): Client uuid:6a5f84e0-73c5-4789-833f-4a9d5d592fb9;id=1 - events processed
00005763195.73225403[3964] 0F7C.0504::11/12-12:50:06.563#18Tongue TiederverConnection.cs(1125): Unknown exception received in PerformOperation
00005764195.74642944[3964] 0F7C.0504::11/12-12:50:06.563#09:MonadUtils.cs(234): Creating new error record from exception
00005765195.74699402[3964] 0F7C.0504::11/12-12:50:06.563#09:MonadUtils.cs(234): Microsoft.VirtualManager.Utils.CarmineException: Unable to contact cluster service on sclu680.xxx.
00005766195.74699402[3964] Ensure that the provided name is a cluster name or the computer name of a node in the cluster, and that the cluster is listed in Active Directory Domain Services. For a newly created cluster, you might need to wait for the cluster name to be added to Active Directory Domain Services.
00005767195.74699402[3964] at Microsoft.SystemCenter.VirtualMachineManager.TaskHelper.ThrowOnFailure()
00005768195.74699402[3964] at Microsoft.SystemCenter.VirtualMachineManager.Cmdlets.AddHostClusterCmdlet.StartTask()
00005769195.74699402[3964] at Microsoft.SystemCenter.VirtualMachineManager.Cmdlets.TaskCmdlet`1.Process()
00005770195.74699402[3964] at Microsoft.SystemCenter.VirtualMachineManager.Cmdlets.CmdletBase.ProcessRecord()
00005771195.74699402[3964] *** Carmine error was: ClusterDoesNotExist (13805)
00005772195.74699402[3964] *** sclu680.xxx **

So it simply does not find the Virtual cluster name in AD.

The virtual server object is present in AD in the default location...

Peter.

Free Windows Admin Tool Kit Click here and download it now
November 12th, 2008 4:01pm

Peter,

I have requested one of our developers Evan to look at the traces.

In the meantime, is the cluster addition working fine if you specify one of the node names in the Add Host wizard? When you do that, is the cluster getting added as a single unit or each node is getting added separately as a standalone host?

Thanks in advance!

Regards,

Shon

November 12th, 2008 11:09pm

Hi Shon,

When I use the node names, it detects (for both nodes) that the node is part of a cluster and then it always tries to add the whole cluster.... which fails every time with the mentioned error.

Thanks for the help on this !

Peter
Free Windows Admin Tool Kit Click here and download it now
November 12th, 2008 11:12pm

I see... so you are blocked on addition (not just on discovery). I will let Evan take a look... thanks for your patience in the meantime.

Regards,

Shon

November 13th, 2008 12:00am

Hi, Peter. Looks like we are failing to execute a WMI query on the cluster. Can you try a couple things for me? From the VMM server machine:

1. Ping sclu680.xxx

2. Run "wbemtest" from the start menu and click "Connect...". Enter \\sclu680.xxx\root\MSClusterfor the namespace, enter the credentials you are using to try to add the cluster in VMM, select the "Packet privacy" option in the lower right, and click "Connect".

3. Repeat with the following namespaces:

\\sclu680.xxx\root\cimv2

\\SDCA680.xxx\root\MSCluster

\\SDCA680.xxx\root\cimv2

\\SDCA681.xxx\root\MSCluster

\\SDCA681.xxx\root\cimv2

Please reply with any errors you get for these steps.

Thanks,

Evan


This posting is provided "AS IS" with no warranties, and confers no rights. User assumes all risks.

Free Windows Admin Tool Kit Click here and download it now
November 13th, 2008 12:33am

Hi Evan,

OUCH !! Deeply embarrassed here ...
Seems DNS for the virtual server was not up to date !! For some reason I cannot remember it seems changed from the intended IP value (that value was registered in DNS).
After updating DNS all is fine !!!
I can start playing now :-)

Thanks all for your help !!

Peter

November 13th, 2008 12:49am

Hi Evan,

OUCH !! Deeply embarrassed here ...
Seems DNS for the virtual server was not up to date !! For some reason I cannot remember it seems changed from the intended IP value (that value was registered in DNS).
After updating DNS all is fine !!!
I can start playing now :-)

Thanks all for your help !!

Peter

Free Windows Admin Tool Kit Click here and download it now
November 13th, 2008 12:49am

I also got this error and after some investigation, I noticed my cluster’s IP was offline due to an earlier network change I had done.  Fixing this, and bringing the cluster’s IP online allowed me to rerun the add-vmhostcluster job.

November 10th, 2010 10:13pm

Hi Peter,

I have Hyper-v 2012 failovercluster with 2 node. I want to integrate this cluster to SCVMM2012 R2 . When I try to import cluster from SCVMMConsole I obtain the following error:

Error (13805)
Unable to contact cluster service on (Cluster Name)

Recommended Action
Ensure that the provided name is a cluster name or the computer name of a node in the cluster, and that the cluster is listed in Active Directory Domain Services. For a newly created cluster, you might need to wait for the cluster name to be added to Active Directory Domain Services.

If I make wbemtest.exe for each node I connect without problem: 

\\host01\root\MSCluster

\\host01\root\cimv2

\\host02\root\MSCluster

\\host02\root\cimv2

If I try to connect to cluster name ( \\Cluster-Hyp\root\cimv2 ) I have the following error:

Number: 0x800706ba

Facility: win32

Description: The RPC server is Unavailable.

Please help

Free Windows Admin Tool Kit Click here and download it now
December 30th, 2013 2:52am

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics