Globus Command Line Interface (CLI) provides an interface to Globus services from the shell and can be incorporated into scripts or used interactively. The CLI allows users to transfer files and manipulate directory structures on a remote endpoint without use of the web interface.
Connecting to Globus on the Command Line
To establish a Globus connection, a user must connect their HCC account to their Globus account. To do so, follow these steps:
- Connect to the desired HCC cluster.
Load the Globus CLI module:
- Use the command `globus login` to start the authorization procedure. A web address will be displayed on screen. Copy and paste this URL into your browser.
- If you are not already logged into Globus, do so now.
- Label the connection if desired, then click Allow to grant Globus the permissions outlined on the screen.
At this point, you can verify you are logged in:
If you have logged in, you will see your email address displayed. If not, the message "No login information available." will be displayed.
To terminate your Globus connection:
Activating a Globus Endpoint
To find available endpoints:
Where search_term is replaced with the keyword you wish to search for. For example:
will display any endpoint with "hcc" in it's Display Name.
To use the endpoint, you can refer it it by it's UUID number. To activate an endpoint, use the command `globus endpoint activate`:
Copy the given URL and paste it into the address bar of your web browser. If you are not already logged into the Globus website, you will be prompted to do so. Once you are logged in, you need to click the `Activate Now` button to activate the endpoint.
Once an endpoint is activated, it will remain activate for 7 days. You can now transfer and manipulate files on the remote endpoint.
To make it easier to use, we recommend saving the UUID number as a bash variable to make the commands easier to use. For example, we will continue to use the above endpoint (Tusker) by assigning its UUID code to the variable `tusker` as follows:
This command must be repeated upon each new login or terminal session unless you save these in your environmental variables. If you do not wish to do this step, you can proceed by placing the correct UUID in place of whenever you see `$tusker`.
Globus CLI supports the commands `mkdir`, `rename`, and `ls`. All globus commands follow the format `globus <command> <endpoint>:<file_path>`. We can list the files on the remote endpoint with the command `globus ls`. To list the files in the home directory on the remote endpoint, we would use the following command:
To make a directory on the remote endpoint, we would use the `globus mkdir` command. For example, to make a folder in the users work directory on Tusker, we would use the following command:
To rename files on the remote endpoint, we can use the `globus rename` command. To rename the test file we just created above, we would use the command:
Single Item Transfers
All transfers must take place between Globus endpoints. Even if you are transferring from an endpoint that you are already connected to, that endpoint must be activated in Globus. Here, we are transferring between Crane and Tusker. We have activated the Crane endpoint and saved its UUID to the variable $crane as we did for $tusker above.
To transfer files, we use the command `globus transfer`. The format of this command is `globus transfer <endpoint1>:<file_path> <endpoint2>:<file_path>`. For example, here we are transferring the file `testfile.txt` from the home directory on Crane to the home directory on Tusker:
You can then check the status of a transfer, or delete it all together, using the given Task ID:
To transfer entire directories, simply specify a directory in the file path as opposed to an individual file. Below, we are transferring the `output` directory from the home directory on Crane to the home directory on Tusker:
For additional details and information on other features of the Globus CLI, visit Command Line Interface (CLI) Examples in the Globus documentation.