Test Altair Accelerator

You may wish to run a few simple jobs to confirm that the system is running properly.

  1. Submit a simple job with the following commands:
    % nc run sleep 30
    Resources= linux
    Env      = SNAPSHOT(vnc_logs/envjohn36829.csh)
    Command  = sleep 30
    Logfile  = vnc_logs/20020602/105731.27873
    JobId    = 00002539
    nc: message: Scheduled jobs: 1       Total estimated time: 0s
    • The JobId in this example is 2539. The leading zeros are insignificant.
    • The environment used for the job execution is a snapshot of the current environment, stored in the file vnc_logs/envjohn36829.csh
    • The default resource list for the job is linux which, in this example, is the platform from which the job was submitted. Later, you will learn how to control the resources assigned to a job.
    Note: If the following dialog appears, see Qualify the Working Directory of Jobs.
      The current directory '/export/scratch/john' may not be a valid directory path for all hosts in the cluster.
      If the job runs on a host other than the local node, and the current directory is not shared on that host, unpredictable results may occur.
        (1)   Continue.
              By choosing this option, you assert that the path
              is valid everywhere.  If it is not, the job is likely
              to fail, because the remote host cannot reach
              the current directory.
              This option causes the creation of a flag file
              called   .vnc    which has the purpose of
              avoiding the repetition of this question for this
              directory and its subdirectories
        (2)   Abort.
              Please ask your Accelerator administrator to
              change the equiv.tcl file to define the rules that
              give a logical name to the current directory.
       Please reply: [1,2] >
  2. Check the status of the job:
    % nc list
    00002539 Done    sleep 10
    % nc summary
    ...output omitted
    % nc info 2539
    Id,User,Group   00002539,john,users
    Environment     SNAPSHOT(vnc_logs/envjohn36829.csh)
    Directory       ${HOMES}/john
    Command         sleep 10
    Status          Running
    Host         alpaca
    Resources    linux
    QueueTime    0s
    RunningTime  9s
  3. Check the output of the job, which in this case is empty, since it is the output of sleep.
    % nc info -l 2539
    Log file is: '${HOMES}/john/vnc_logs/20020602/105731.27873'
  4. Rerun the job.
    The first time Accelerator notices that the job is already Done (=VALID) so it does not run it again. The second time, the option -f forces Accelerator to rerun the job even if it is already Done.
    % nc rerun 2539
    nc rerun 00002539
    nc: message: Not rerun: 00002539
    % nc rerun -f 2539
    nc: message: Job 00002539 is already VALID.
    nc: message: Scheduled jobs: 1       Total estimated time: 15s
  5. Stop the running job. The job will fail.
    % nc stop 2539
  6. Forget the job:
    % nc forget 2539
    nc: message: Forgetting 1 jobs
  7. Get a quick report:
    % nc summary
    Accelerator Summary For User john
    TOTAL JOBS        105      Duration: 4m17s
    Failed             53
    Queued             50
    Idle                2
    JOBS   GROUP    TOOL        WAITING FOR...
    50   pi       vtclsh     'hpux#1 '