Setup
We first consider a simple Gladier Flow that transfers data from an instrument computer to an analysis computer, and then runs an analysis function on the analyasis computer. The flow thus comprises two steps:
Transfer: Copy data to the analysis computer.
Compute: Run the analysis function on the data copied in the first step.
The first step involves a Globus Transfer action, and the second a Globus Compute action. In order for your computer to be accessible by Globus, it needs to be running a Globus collection. See the Globus Connect Personal section below. For the second part, you need to set up a Compute Endpoint to tell Globus where it should execute functions. See the Comptue Endpoint section below for details.
Globus Connect Personal
To run tools like gladier_tools.globus.Transfer, we need access to a Globus collection.
Follow the Globus Connect Personal instructions
to set up your own Globus Connect Personal
endpoint and configure a collection, then navigate to
Globus Web App collections to see details about the
collections to which you have access.
Note
Your Globus endpoint should be accessible on the same machine you install your FuncX endpoint.
Globus Compute Endpoint
We also need a Compute Endpoint to run remote functions. Follow the Compute Endpoint instructions to set up your own endpoint. Once an endpoint has been configured, you can:
run
globus-compute-endpoint listto list all endpoints to which you have accessrun
globus-compute-endpoint start <endpoint-name>to start an endpoint