none
SCVMM Access denied contacting server

    Question

  •  

    Hi,


     I’ve installed SCVMM 2008 RTM. The server hosting this is hyper-v enabled and part of a 2 node cluster for failover. As soon as I try and make this server a host I get the error ‘Access has been denied while contacting server XXX, .... check administrative permissions, check DCOM permissions’. I’ve done both these things and I can’t find a problem, both servers are part of the same AD. If I add the other host (part of the failover cluster) it gets a little further as it detects its part of a fail over cluster and it will try to the other nodes. This then fails when contacting the cluster, possibly this is also access related?


    Does anyone know, if installing SCVMM 2008 on a hyper-v enabled server is supported? I have added another server with no problems, any suggestions on how to fix this would be greatly appreciated,

    Wednesday, October 22, 2008 3:56 PM

Answers

  • Fix from Microsoft :-

     

    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.

     

    In my environment, SCVMM, Hyper-v and all the cluster resources were on the node I was trying to add into SCVMM. Once I moved the resources over as described above it worked perfectly

    Monday, November 03, 2008 2:01 PM
  • Tim,

     

    Not sure if I get your environment right. Here is what I understand:

    1) You installed SCVMM 2008 RTM version.

    2) You have a two-node failover cluster.

    3) You tried to add one of the cluster nodes and it failed.

    4) If you tried to add the other node, it proceeded further, but it failed later when it tried to add the second node.

     

    Did you install SCVMM server inside of a VM that is hosted by the cluster? Is the VM (where the SCVMM is installed) physically running on the node that you are unable to add?

     

    Could you please confirm or re-state your environment?

     

    Also, have you checked the firewall settings on the node that you could not add? Here are the ports we need open for SCVMM to function:

    http://blogs.technet.com/chengw/archive/2008/06/30/scvmm-and-network-ports-we-use-for-communication.aspx

     

     

    If the adding still fails, we'll need you to collect the VMM Server side trace to understand why. Here is how to collect the traces: http://blogs.technet.com/chengw/archive/2008/05/08/how-to-collect-scvmm-traces.aspx

     

    Thanks,

    Cheng

    http://blogs.technet.com/chengw
    Wednesday, October 22, 2008 7:03 PM

All replies

  • Tim,

     

    I assume the user account you specified while trying to add the cluster (using either of the nodes) is admin on both nodes of the cluster. Are you able to complete the add host wizard or you are seeing the error even before that? Can you please send the exact error message? Couple of other things to ensure are - clussvc is running on every node of the cluster, cluster name resource itself is online, AD replication is complete for the cluster name in case this is a newly created cluster.

     

    Having said that, you might want to rethink about installing VMM server on one of the cluster nodes. Generally the reason one creates cluster is to achieve high availability in case of hardware/software failure of the nodes. In your case, if the node running VMM server fails, the HAVMs running on that node will fail over to other nodes in the cluster but VMM server itself will be down - which means running VMM jobs will be lost and you will lose visibility to your virtual infrastructure thru VMM until the failed node and VMM server comes back up.

     

    Hope this helps.

     

    Regards,

    Shon

    Wednesday, October 22, 2008 6:20 PM
  • Tim,

     

    Not sure if I get your environment right. Here is what I understand:

    1) You installed SCVMM 2008 RTM version.

    2) You have a two-node failover cluster.

    3) You tried to add one of the cluster nodes and it failed.

    4) If you tried to add the other node, it proceeded further, but it failed later when it tried to add the second node.

     

    Did you install SCVMM server inside of a VM that is hosted by the cluster? Is the VM (where the SCVMM is installed) physically running on the node that you are unable to add?

     

    Could you please confirm or re-state your environment?

     

    Also, have you checked the firewall settings on the node that you could not add? Here are the ports we need open for SCVMM to function:

    http://blogs.technet.com/chengw/archive/2008/06/30/scvmm-and-network-ports-we-use-for-communication.aspx

     

     

    If the adding still fails, we'll need you to collect the VMM Server side trace to understand why. Here is how to collect the traces: http://blogs.technet.com/chengw/archive/2008/05/08/how-to-collect-scvmm-traces.aspx

     

    Thanks,

    Cheng

    http://blogs.technet.com/chengw
    Wednesday, October 22, 2008 7:03 PM
  • Hi Cheng,

                    Yes the environment is as you describe. The SCVMM server / console are installed on one of the cluster nodes, not in a VM. It’s installed on the node that has the DCOM error straight away rather than the one that gets a little further. I’ve double checked permissions and am sure these are correct, I will enable logging and retry this. The firewall is off on both nodes.

    These boxes have the RTM of hyper-v and I’m aware there is an update around for 6 cores, do you think it would be worth installing this or would SCVMM and older / incorrect version as an issue.

    Regards,

     

    Thursday, October 23, 2008 11:05 AM
  • Tim,

     

    The patches from Hyper-V would only turn your host into "Needs Attention" state, but they should not block you from adding them into SCVMM. If you want to be safe, please feel free to install all Hyerp-V patches.

     

    And, yes, Shon and I will need traces to understand what happened in your environment.

     

    Thanks!
    Cheng

    Thursday, October 23, 2008 5:41 PM
  • Hi Cheng / Shon,

     

    I've sent the logs to Cheng's email address. I am now getting a little further, in that SCVMM does not error immediately when I select one of the nodes. I did apply KB951308 and KB956589. However this still fails trying to contact the cluster service with Error 13805, you can see this in the logs. I’ve double checked permissions and they seem good, the cluster name is in AD. SCVMM is running using the local system account, should I try with my DA account?

     

    Regards,

     

    Thursday, October 30, 2008 10:49 AM
  • Fix from Microsoft :-

     

    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.

     

    In my environment, SCVMM, Hyper-v and all the cluster resources were on the node I was trying to add into SCVMM. Once I moved the resources over as described above it worked perfectly

    Monday, November 03, 2008 2:01 PM
  • I just ran into this problem also and the "cluster res" command told me that the cluster group "Cluster Group" was owned by my VMM server (it is a physical Hyper-V server which is part of many in a cluster). I ran the following command to move the group to another host (cluster group "Cluster Group" /move:NewClusterHost) and it worked right away.

    Maybe this is a bug which should be fixed in a further update release?

    Reuv
    Sunday, March 07, 2010 7:03 AM
  • I tried using "cluster res" in an elevated command prompt.  It showed both machines in my host cluster: host1 and host2.  However, it showed host1 as the owner of the cluster.   I was attempting to add host2 to SCVMM.  SCVMM resides on host2 currently.  I got it  working when I added host1 instead, because it was the "owner" host for the fail-over cluster.

    Short Answer: Try adding the "Owner" of the host cluster that is listed in "cluster res", when this is run via an elevated command prompt.

    Tuesday, September 11, 2012 7:19 PM