none
Strange issue when using WCF with HPC (No new client connection in 0 milliseconds) RRS feed

  • Question

  • When I'm running my WCF-based task (service on each node + console-based client app) on HPC server cluster, the 2 jobs is created - one of them is WCF broker, and another is workers. Task works well, I'm getting all needed data from each node, but when it finishes, the strange statuses are shown in HPC cluster manager (Jobs->finished).

    The broker job has the state 'Finished' with an error message "No new client connection in 0 milliseconds, exiting", and the worker job has also this state but error message says "The task's job has been canceled. Please see the Job details for additional information.". In client application Session.Dispose() is called.

    Can you suggest how to get rid of these strange problems?

    Thanks,
    Max
    Friday, March 27, 2009 5:22 PM

Answers

  • If the job status if 'Finished', it means overall it's good.

    The task message you see "No new client connection in 0 milliseconds, exiting", which means the broker exits because the only client (your client) has disconnected so the broker decides to die, is normal.

    I'm afraid you can't get rid of those messages but I hope the job status get you enough information.

    • Marked as answer by McSIMM Wednesday, May 6, 2009 8:44 AM
    Thursday, April 2, 2009 1:34 AM