Using Globus

From GMU Office Of Research
Jump to navigation Jump to search

Globus is a research data management service. It provides tools and an easy-to-use interface to facilitate parallel, load-balanced, fault tolerant data transfer ranging from Megabytes to Petabytes in volume securely. The Globus connect portal can be used to transfer data across different clusters such as XSEDE, ARGO, NCAR etc. It can also be used to transfer data between a personal desktop/laptop and a cluster such as ARGO.


Logging in to Globus

Go to the Globus.org website and click on the login button. It will take you to the login page where a drop-down menu is available. In the search field search for George Mason University. Hit the continue button and it will redirect to the page of Mason Federated Services. Log in using your netID and password. If this is the first time you are logging into you account, you will be asked a number of questions (such as accepting terms and conditions of use) on several pages. Once logged in, the File Manager page will be displayed. Here you can select desired source and destination systems for file transfers.

Login netid.png Login netid 2nd.png


Transferring data between two clusters

Data transfer between clusters

In the Collection text box, you will enter the name of one the clusters that you will transfer files between. If one of the sources or destination is ARGO, you will need to enter "gmu#ARGO", and then select either gmu#ARGO-1.ORC or gmu#ARGO-2.ORC (it doesn't matter which one). The first time you do this, you will need to enter your username and password. You will be shown you home directory. You can navigate to another directory using the GUI, or you can enter a new path into the Path text box.

The define the other endpoint of the file transfer, you will need to select the Panels button at the top of the window showing two boxes side-by-side. This will display a second set of Collection and Path text boxes on the right side of the window. You will use these to define the other end of the transfer. For example, let's suppose you wanted to transfer files to or from an XSEDE or NCAR cluster. All XSEDE clusters have the term "XSEDE" in the name, and the same is true for NCAR clusters. Enter one of these terms into the second Collection box to perform a search, and then select the specific cluster you are interested in. Again, you will need to enter your username and password the first time you do this.

On the "to" side of the transfer, navigate to the directory where you want to place the file(s), and on the "from" side navigate to and select the file(s) or folder(s) that you want to transfer. Then select the Start button at the bottom of the "from" side to begin the transfer. In the example figure, a folder named LAI_4_16_rows_45 directory from NCAR GLADE system is selected to be transferred to the user's home directory on the Argo system. Data transfer rate and progress can be monitored from the activity tab. Also, Globus will notify you of the successful transfer status or any failure status through an email.


Transferring data between a personal desktop/laptop and a cluster

If one of the sources or destination is a personal computer, Globus Connect Personal software will be required to run from the personal computer to establish a secure connection to Globus system. To download Globus Connect Personal, look at the bottom of the landing page, globus.org, click on the Globus Connect link and it will take to the Globus Connect page where a download link for the Globus Connect Personal can be found. There will be a field "Display Name". Give your personal computer an identifier that you will remember. Later, the personal computer will be selected as one of the data nodes using this identifier/name. Download Globus Connect Personal for your platform. Extract the zip file and, from a terminal window, run the Globus Connect Personal executable. When run for the first time it will ask for security code. Copy and paste the setup up key as per instruction on the page https://docs.globus.org/how-to/globus-connect-personal-linux/. Once setup is done, execute Globus Connect Personal as shown in the picture. Click on connect. After a successful connection, 'connected' status will appear. Now go to the Manage Data page. The rest of the data transfer procedure is identical to the transfer from cluster to cluster method as described above. This time, one of the endpoints will be the personal computer. In one of the endpoint fields, if searched, the given identifier for the personal computer will show up and can be selected as the endpoint. The other endpoint will be the chosen cluster. After initiation of the transfer, data transfer rate and progress can be monitored from the activity tab. Also, Globus will notify you of the successful transfer status or any failure status through an email. Remember that Globus Connect Personal must be running on the personal computer whenever data is being transferred through globus.

Data transfer between a cluster and a personal computer

Globus connect personal download 2nd.png Globuspersonal setup1.png Globuspersonal setup2.pngGlobuconnect connected.png