Sign in
Microsoft.com
 
Microsoft
 
 
 

locked
clusrun exits with proxy error RRS feed

 > 

  • Question

  • Question
    Sign in to vote
    0
    Sign in to vote

    I have a couple of nodes in an HPC2008R2 that can not communicate with the master via clusrun commands. One example is below. The error happens both with the GUI and the CLI version

    can you give me any idea what services are used here? I checked the event logs both on the systems affected and athe master node - but could not detect any problem. Re-installation of the affected compute nodes does not help. The nodes report as "HEALTH OK" and can be reached via remote desktop. I can also remotely query their WMI interfaces (powershell, get-wmi... comand)

     

    C:\Windows\system32>clusrun /nodes:emswlad13 dir
Enter the password for 'MSHPC\mhebenst' to connect to 'emswhpcm1':
Remember this password? (Y/N)y
Command has failed on node EMSWLAD13.
Message:Unable to open standard output file on node EMSWLAD13:Exception 
    
'Failed to create standard output file '\\EMSWHPCM1\pipe\Hpc\Proxy\505\EMSWLAD13',
    
 error code 1231' reported creating the task. 
    
-------------------------- EMSWLAD13 returns 1 --------------------------

     

    thanks for any help

     

    Michael

    • Moved by parmita mehtaModerator Tuesday, October 19, 2010 7:14 PM (From:Windows HPC Server Deployment, Management, and Administration)
    Thursday, October 14, 2010 9:34 PM

Answers

  • Question
    Sign in to vote
    0
    Sign in to vote

    Hi Michael,

    Error message 1231 means that The network location cannot be reached.

    This means that the compute node is not able to write to the location on the headnode it wants. This might be a firewall or a name resolution issue.

     A good test would be to login on the compute node having the problem  as the user who was running clusrun and then try to see if you can write to a share on the headnode from that compute node.

    See if you can write a file to \\EMSWHPCM1\CCpSpoolDir\ from  the compute node.

    If you cannot write to a share on the headnode, you might want to do a nslookup for the headnode on the compute node to see that the server name gets resolved correctly.

    sayantan

    Wednesday, October 20, 2010 8:00 PM
    Moderator