Providing a minimal reproducible pipeline

When you create a new support request, we need to be able to reproduce your issue on our infrastructure. To aid our efforts, we ask that you provide us with a minimal reproducible pipeline that will allow us to identify a solution to your issue in a reduced timeframe. The more information you provide upfront, the less additional information we will need to request from you to start our debugging process. Once your issue has been reproduced on our infrastructure, we can advise on the expected behavior and the steps you need to take for the solution. 

A minimal reproducible pipeline may consist of one or more of the following components:

• For Nextflow pipelines, include:

  • As few processes as possible

    • Ideally, provide us with the process information that you are experiencing issues with, rather than your entire production pipeline

  • Any required Nextflow configuration

    • Include all the configuration variables that are relevant to the problematic processes in your nextflow.config

  • Access to the containers used by the pipeline, if they aren’t already public

    • To rule out any issues with your containers, we need to run the pipeline using the same containers used in your environment

  • A minimal test dataset that can be used as input to the pipeline

    • With some exceptions, we generally do not need your entire production dataset to be able to reproduce your pipeline issue 

  • The exact command and parameters used to run the pipeline

    • To rule out the run parameters as the cause of your issue

  • Specific details about the compute environment and infrastructure (if applicable)

    • Any relevant information about your environment will allow us to replicate your pipeline in the same setting that produced your issue

  • Any other relevant information

    • Cloud provider constraints, software dependencies, etc.
      
      

• Other useful information:

  • Nextflow version

  • Tower version

  • Tower CLI version and exact command (if applicable)

  • Java version

You can either attach the files listed above to your initial email, the issue (if already created) in our support portal, or (ideally) via a Git repo that contains the code. This will make it easier for us to clone and run the pipeline.