-
Notifications
You must be signed in to change notification settings - Fork 0
Globus Setup and Usage
Globus is a service for the fast, simple, reliable, and secure transfer of files between clusters. Compute Canada recommends using this program for the transfer of files between your wsl environment and the Compute Canada node. This guide is for the installation and usage using a Linux environment.
- Navigate to your home directory in wsl.
wsl
cd ~
- Download Globus Connect Personal.
wget https://downloads.globus.org/globus-connect-personal/linux/stable/globusconnectpersonal-latest.tgz
- Extract the the files from the downloaded tarball.
tar -xvf globusconnectpersonal-latest.tgz
- Navigate to the globusconnectpersonal directory.
cd globusconnectpersonal-x.y.z
-
x.y.z,
dictate your version number.
- Run Globus Connect Personal for the first time.
./globusconnectpersonal
-
You will be given a series of text prompts for relevant information.
-
One of the text prompts will be to give your personal endpoint a unique name.
-
When you generate your personal endpoint, it will display the UUID. Make sure to copy this information down as it will be used later.
- Go to the Compute Canada Globus Webpage: (linked here).
- Navigate to the dropdown menu for organizations. Select
Compute Cnada
organization and click continue. - Sign in using your Compute Canada username and password.
- In the left sidebar, select
FILE MANAGER
. - In the
COLLECTION
search bar input your UUID and find your endpoint. - In the
~/globusconnectpersonal-x.y.z
directory and enter the following command:
./globusconnectpersonal -start
- This should allow you to access your files.
- If it doesn’t, close your terminal window and renavigate to the
~/globusconnectpersonal-x.y.z
directory and reinput the command. Refresh the list of your endpoint in the file manager and it should display the contents of yourwsl
directories
-
In
FILE MANAGER
on the right sidebar, select theTransfer or Sync to…
option. This will open a second window where you may search for a second endpoint to transfer your files to and from your personal endpoint. -
Search for and select the Compute Canada cluster you will be running your programs in. Below is a list of the endpoint names, based on the respective cluster:
- computecanada#beluga-dtn
- computecanada#cedar-dtn
- computecanada#graham-globus
- computecanada#niagra
- Select the files you want to transfer and click the
Start
button below the collection name. - Your files should now be transferred into the Compute Canada cluster.
Instructions on how to use Globus: https://docs.globus.org/how-to/ Compute Canada’s documentation on Globus: https://docs.alliancecan.ca/wiki/Globus
- 1.0.1 (October 10 2022): Adjusted the code blocks.
- 1.0.0 (August 31 2022): Initial draft of document.
Should we have some kind of footer?
General Guides
- Nomenclature: Data Files; Scripts
- Script Writing
- Organizing Data and Analysis
- Working in GitHub
Useful Links