none
can't provision a preconfigured compute node apparently AFTER SP2

    質問

  • Time Message
    11/29/2011 3:27:17 PM The Management service encountered an error while performing a change on the node: Access by head node NY-SMBC-CM\NYBLSQLDEV01 to node NY-SMBC-CM\APPSER3 is denied. This can occur if you are adding workstation nodes that are not in the same Active Directory domain as the head node, and you did not specify the fully qualified domain name of the head node when installing HPC Pack on the workstations. Check the operation log in the Administration Console for more information.

    This worked when I was at SP1.  As you can see, the 2 machines are in the same domain.

    Where exactly is the "operation log in the Administration Console"?

     

    2011年11月29日 20:29

すべての返信

  • Hi,

    The log you are referring to is part of the cluster manager UI.

    Go to Node Management -> click on the name of the node (in your case NY-SMBC-CM\NYBLSQLDEV01 or  NY-SMBC-CM\APPSER3) -> In the lower part of the window you can find a tab named "operations"

     

    Hope that answers your question

    2011年12月1日 19:12
  • yes that's what I thought.

    Unfortunately, since the node didn't successfully provision ( Provision Log Tab ), there's nothing in the Operations tab.

    Note since my original note, I've also determined that adding SP3 does NOT change my situation ( on EITHER the head node OR the node to be provisioned ), and that upgrading WORKSTATION nodes results in the same problem.

    I've also discovered if I provision at SP1, then update to SP2, the Node Health -> Node Connectivity goes to "HPC Node Manager Service unreachable"

     

     

    2011年12月1日 21:47
  • Dear Whalerboy,

     

     Just looking for a bit of clarification here. Am I correct in presuming that

    1) you installed HPC SP2 on your head node

    2) you updated your workstation node to SP2

    3) When you installed SP2 the workstation node was pointed correctly at the head node (when you enter in the FQDN for the head node during setup)

    4) The node was visible in an unapproved state on your head node

    5) you assigned a workstation template to that node to bring it into the cluster.

     

    Then it was during step 5 that you got the error

    "The Management service encountered an error while performing a change on the node: Access by head node NY-SMBC-CM\NYBLSQLDEV01 to node NY-SMBC-CM\APPSER3 is denied. This can occur if you are adding workstation nodes that are not in the same Active Directory domain as the head node, and you did not specify the fully qualified domain name of the head node when installing HPC Pack on the workstations. Check the operation log in the Administration Console for more information."

     

    The correct order for updating a cluster is to patch the head node first and then patch the compute nodes.  If you patch compute (or workstation nodes) prior to the head node you can have communication issues.

     

    You could also be having issues with stale Active Directory connections -  one other option would be to do a soft rejoin to the domain and update your group policy on your workstation node and head node.

     

    Just some ideas - if you could clarify that I have the steps correct above (in your situation) then that would be able to help us troubleshoot more.

     

    Thanks,

    Mark

     

    2011年12月8日 18:19
  • just getting back to this ...

     

    Yes your summary is correct.

     

    In addition, I just re-installed the head node, and upgraded to SP3, and still have the same access denied problems trying to add compute nodes.

     

    Help!

    2012年1月25日 20:09