Tar and Transfer for collections with an associated flow policy

These examples demonstrate how to build flows that are meant to be used as the associated_flow_policy for GCS collections. They are variations of the Tar and Transfer examples that rather than get the source paths from the user input, they are parsed from the incoming globus-transfer-transfer#0.10 transfer data.

Each of these examples creates an archive file from the user’s files and transfers that archive to a destination. In one case the source data is already on the server running Globus Connect Server and Globus Compute, and in the other it is on a source collection owned by the end user.

Prerequisites

To run these examples, you must have a properly configured server and some local software installed.

You must have a co-located Globus Connect Server Collection and Globus Compute endpoint, either hosted on the same server or at least with access to a shared filesystem.

Globus Connect Server Collection: You can follow this guide for setting up a Globus Connect Server Collection to install Globus Connect Server and configure a collection.

For ease of use, we recommend using a Guest Collection.
Globus Compute Endpoint: This guide for setting up a Globus Compute Endpoint covers installation of the Globus Compute software.

This Compute endpoint must have read/write permissions on the same storage location where the Globus Connect Server collection is hosted.
Globus CLI: You will also need the Globus CLI installed (CLI installation docs).

Globus CLI documentation recommends installation with pipx, as in pipx install globus-cli.
Globus Compute SDK: You must have the globus-compute-sdk Python package available. We strongly recommend using a virtual environment for this installation; installing with pip install globus-compute-sdk.

You can follow the Globus Compute install documentation to install the Compute SDK client package in a virtualenv.