Web interfaces for Puhti and Mahti

Intro

The web interfaces for Puhti and Mahti at www.puhti.csc.fi and www.mahti.csc.fi can be used to access the supercomputers using only a web browser. A web interface for LUMI is also available at www.lumi.csc.fi (see the LUMI documentation for more details).

Features available in both the Puhti and Mahti web interfaces:

• View, download, upload and move files between Allas, the supercomputer and your local computer
• Open a shell on the login node
• Open a persistent shell on a compute node
• View running batch jobs
• View disk quotas and project status
• Launch interactive apps and connect to them directly from the browser:
  • Desktop with apps such as Maestro and VMD
  • Julia-Jupyter
  • Jupyter
  • Jupyter for courses: An interactive Jupyter session specifically for courses
  • MLflow
  • TensorBoard
  • Visual Studio Code

Apps available in Puhti only:

• Accelerated visualization with applications:
  • Blender
  • COMSOL
  • ParaView
  • VMD
• RStudio
• MATLAB

Everything still runs directly on the supercomputers, giving you all the resources and power of a supercomputer, but easily accessible using only a web browser.

Connecting

Using a web browser, go to www.puhti.csc.fi or www.mahti.csc.fi. On the landing page, click on "Log in" and select an appropriate authentication provider. When logging in using your CSC user account, select CSC as the authentication provider and use the same username and password you use when connecting with ssh.  

After successful authentication, you will see the dashboard.  

From here you can browse your files on the supercomputer, start a shell, view running jobs or start one of the many available applications. The dashboard also contains some important system information.

Available features

Shell

The shell apps can be found under Pinned apps or on the top navbar under the Tools section. There are two different shells.

The Login node shell launches a normal linux shell on one of the login nodes. Any command that is running when the login shell browser tab is closed will stop. Note that the same rules apply here as during a normal ssh session. Login nodes are only for light pre/postprocessing (see Usage policy).

The Compute node shell launches a persistent shell on a compute node for heavier commands that should not be run on login nodes. The persistent shell will keep running even if you close your browser or lose internet connection.

Files

The file browser can be opened using the Files section on the top navbar (this displays a list of all project disk areas), or using the shortcut to the home folder at the bottom of the front page. In the file browser you can upload/download files, create new files and directories, or open a shell in the current directory.

Note

Keep the tab with the file browser open while the file transfer is in progress to ensure that it completes successfully. Uploaded files will overwrite existing files with the same name without prompting. Currently the maximum size for individual file uploads is 10GB

Clicking on a file will open it in view only mode, for more options like editing, renaming and deleting, use the button with three dots next to the filename.

The file browser comes with a basic text editor. Some important notes on that:

• If no changes have been made, the save button is grayed out.
• There is no save-as feature
• If a read-only file is opened no indication will be given to the user but no changes will be applied

Using Allas

In the web interface, the Allas object storage service can also be accessed using the file browser.

To configure authentication for Allas it is recommended that you use the Cloud storage configuration app available in the web interface. Once you open the app, you will be prompted to enter your CSC password. After you have authenticated using your password, you will be able to create both S3 and Swift connections, also known as remotes, to Allas. The remotes are only valid for a single project, but you can create remotes for all of your projects. The created remotes will be visible in the Files dropdown in the navbar and in the file browser.

Note

The Swift and S3 protocols are not fully compatible with each other, particularly with files larger than 5 GB. For more details about the differences between the protocols, see Allas protocols.

Additionally, LUMI-O is also supported for use through the file browser and can be configured by running allas-conf as allas-conf --lumi. After running the command, the web interface server must be restarted, which can be done by clicking Restart web server in the Help menu in top right section of the navbar. Once the server has been restarted, the lumi-o remote will be available in the Files dropdown in the navbar and in the file browser.

Configured remotes that are not accessible, for example, due to expired authentication or network connection issues, are not be visible in the Files dropdown menu.

The file browser works the same way when accessing Allas as it does when accessing the shared filesystem on the supercomputer. Note that uploading large files from your local computer to Allas is currently not recommended due to technical limitations.

Using IDA

The IDA storage service can also be used, although some key features, such as moving data from the staging area to the frozen area, are only possible though the IDA WWW-interface. To use IDA in the Mahti web interface, it must first be configured for use with Rclone on a Mahti login node as follows.

module load allas
rclone config

In the Rclone configuration interface, create a new remote with the following options:

1. Storage: WebDAV (#45)
2. URL: https://ida.fairdata.fi/remote.php/webdav/
3. Vendor: Nextcloud (#1)
4. Username: Your CSC username
5. Password: Login to the IDA WWW-interface, go to the settings in the top right corner. Go to the Security tab and create a new app password. Copy the password and paste it in the Rclone configuration interface.
6. Bearer token: Leave empty
7. Advanced config: No

After completing the Rclone configuration, the web interface server must be restarted, which can be done by clicking Restart web server in the Help menu in top right section of the navbar in the web interface. IDA can then be accessed in the file browser, where you will be able to upload, download, transfer and edit files in the staging area and view and download files in the frozen area.

Active jobs

Recent and running batch jobs can be viewed using the Jobs section on the top navbar and selecting Active jobs. Here you can view the current status of the job and what kind of resources were requested. Deleting a running job will cancel the job.

In the future it will become possible to submit batch jobs through the web interface, but for now the recommended way is to launch the jobs using sbatch from the shell.

Interactive apps

Interactive apps are programs that can be launched and run on the compute nodes and provide a web interface. These are apps such as Jupyter Notebook, RStudio and Visual Studio Code.

If the interactive app does not start or does not work as expected you can delete the session and try to launch the app again.

Launching an interactive app

The interactive apps can be found in the navigation bar under Apps, or on My Interactive Sessions page. After selecting an interactive app from the list you will be presented with a form to configure the session. After submitting the app form, and the Slurm job for the app has finished queuing, the app will be started and you will be able to connect to the application on the My Interactive Sessions page.

For a list of applications and specific instructions see apps.

Partitions and resources

Only some of the partitions on the system are available for use in the web interfaces. Some apps have a more limited set of partitions available than other apps.

In the Puhti web interface, the interactive, small, test, gpu and gputest partitions are available. Selecting the gpu or gputest partition will allocate a Nvidia V100 GPU.

In the Mahti web interface, the interactive and gpusmall partitions are available. Selecting the gpusmall partition will allocate a split Nvidia A100 GPU (a100_1g.5g). For more details about the split GPUs, see the Mahti partitions page.

Project view

Using the project view under the Tools section on the top navbar, you can view current disk and project billing unit quotas on the supercomputers. For more information see project-view.

---

Last update: March 5, 2024

Do you find this page useful?

Yes

Thank you for your feedback!

No

Thank you! Your feedback is appreciated.

Please let us know how we could improve this page by contacting us or by creating an issue in GitHub. You can also edit this page yourself and contribute your improvements by creating a pull request.