Test Configuration

Warning

This is an experimental feature right now that may change drastically in the future. So always revisit this documentation after upgrading OnDemand and attempting to test your cluster configuration.

We currently provide an experimental rake task in the dashboard for testing your resource manager configuration in the cluster configuration files.

  1. For all rake tasks we will need to be in the dashboard’s root directory (this will change in the future):

    cd /var/www/ood/apps/sys/dashboard
    
  2. We will now list all available tasks that we can run:

    source /opt/ood/ondemand/enable
    bin/rake -T test:jobs
    # rake test:jobs           # Test all clusters
    # rake test:jobs:cluster1  # Test the cluster: cluster1
    # rake test:jobs:cluster2  # Test the cluster: cluster2
    

    This list is dynamically generated from all the available cluster configuration files that reside under /etc/ood/config/clusters.d/*.yml.

  3. For now let us test just a single cluster (in our example we use cluster1, so you will need to replace it with the name of the cluster you configured):

    Run this command:

    sudo su $USER -c 'source /opt/ood/ondemand/enable; bin/rake test:jobs:cluster1 RAILS_ENV=production'
    

    Results

    # [sudo] password for user:
    # Rails Error: Unable to access log file. Please ensure that /var/www/ood/apps/sys/dashboard/log/production.log exists and is writable (ie, make it writable for user and group: chmod 0664 /var/www/ood/apps/sys/dashboard/log/production.log). The log level has been raised to WARN and the output directed to STDERR until the problem is fixed.
    # mkdir -p /home/user/test_jobs
    # Testing cluster 'cluster1'...
    # Submitting job...
    # [2018-04-24 10:15:32 -0400 ]  INFO "execve = [{\"PBS_DEFAULT\"=>\"oak-batch.osc.edu\", \"LD_LIBRARY_PATH\"=>\"/opt/torque/lib64:/opt/rh/rh-nodejs6/root/usr/lib64:/opt/rh/rh-ruby24/root/usr/lib64\"}, \"/opt/torque/bin/qsub\", \"-N\", \"test_jobs_cluster1\", \"-S\", \"/bin/bash\", \"-o\", \"/users/appl/jnicklas/test_jobs/output_cluster1_2018-04-24T10:15:32-04:00.log\", \"-l\", \"walltime=00:01:00\", \"-j\", \"oe\"]"
    # Got job id '10820525.oak-batch.osc.edu'
    # Job has status of queued
    # Job has status of queued
    # Job has status of queued
    # Job has status of queued
    # Job has status of completed
    # Test for 'cluster1' PASSED!
    # Finished testing cluster 'cluster1'
    

    Please ignore the Rails Error: message as this is just a warning that doesn’t affect your OnDemand installation in any way. We are currently tracking this issue in GitHub at OSC/dashboard#364.

    Tip

    We actually launch the rake task with sudo to best mimic the environment that the OnDemand applications are run under.

    You can run the rake task as the current user, but it may lead to a false positive as your user environment may have the correct libraries and paths loaded in it that may not necessarily exist in the cleaner OnDemand application environment.

    This creates and submits a batch job that echo’s a defined string. It then pings the batch server every 5 seconds until the job is completed. Finally, it parses the output file looking for the defined string. The test passes if it can find the string in the output file.

    If something fails at any point in the chain, then the test fails. This may require you to make edits to the corresponding cluster configuration file under /etc/ood/config/clusters.d/ and run the test again.

    Note

    If your job fails to submit because you need to supply more submission arguments, e.g., a queue, memory requirements, an account, etc. You can provide these command line arguments as a string with the environment variable SUBMIT_ARGS as:

    Run this command:

    sudo su $USER -c 'source /opt/ood/ondemand/enable; bin/rake test:jobs:cluster1 RAILS_ENV=production SUBMIT_ARGS="-A myaccount"'
    

    Note that the SUBMIT_ARGS="..." is defined at the end of the command.