darshan-logs

Repository of example Darshan log files. Each is in a subdirectory with it's own README.md describing why that specific log is a notable example.

If a Darshan log is too large to store directly in the git repository, then the directory will contain a *.darshan.link file rather than a *.darshan file. The link file is just a text file containing the URL to the example in an externally hosted download site.

You can download all such files at once by running the download-large-logs.sh script at the top level of this repository. This script will also set the correct path/name for each log so that there is a corresponding *.darshan file alongside each *.darshan.link file in the repository.

Guidelines for submitting new log files

• Anyone can contribute logs for consideration by opening a GitHub pull request.
• Make sure that you have permission to publicly share the log (i.e., that it does not disclose any proprietary or private information).
  • The darshan-convert utility can be used to anonymize logs with the --obfuscate command line argument if needed.
• If the log file is intended to exercise a particular corner case for CI testing, then review existing logs to see if there are any that already posses the desired property.
  • A duplicate example may still be warranted to help isolate behavior or present a variation on how it is manifested, but it is helpful to be aware of the duplication.
• Include relevant details in a README.md file for the log (see existing logs for examples).
• If the file is greater than 100 MiB in size, then it should be hosted externally rather than added directly to the repository.
  • Use a publicly-accessible persistent URL.
  • Follow existing examples for how to constitute the log file in a local clone using the download-large-logs.sh script and a URL reference stored in a *.darshan.link file.

There is no formal convention for how to name logs or how to describe them in the README.md files for now. We may add taxonomies or indices in the future.