All changes pertaining to the Kubernetes charts for Hubs deployed for the UW JupyterHub for Teaching and Learning service, as well as ancillary projects are documented here. This service is based on the Zero to JupyterHub with Kubernetes project. Changelog conventions from the Keep A Changelog project.

Note: changes to Charts apply only to Hubs created on or after release date.

[2025-Winter] - 2024-11-25

Changed

• Using v2.5.6 of RTTL-admin application
• Target Kubernetes version now 1.30.5
• Updated usersync & identity pods to use Google Artifact Registry
• Default user notebook images updated to 2.7.0, with the following changes:
  • Uses Quay.io for base images
  • JupyterLab v4.2.0
  • Jupyter Notebook v7.2.0
  • Python v3.11.9
  • R v4.3.3

Added

• Added libncurses5 and zip OS packages to SciPy Notebook image (the Tellurium package was also added but later removed due to dependency conflicts).
• Added lmodern OS packages in RStudio Notebook image.
• Added CI/CD for Notebook images.

Fixed

• Fixed issue with user pod timeouts on first launch on a new node by increasing KubeSpawner HTTP timeout for user pods to 60 seconds.
• Fixed issue with syncing Canvas courses containing 'Pending' users (eg, invited but not yet accepted).
• Fixed RStudio help pane links bug (see: https://github.com/jupyterhub/jupyter-rsession-proxy/issues/156)

[2024-Spring] - 2024-03-05

Changed

• Using v2.5.4 of RTTL-admin application
• Move from deprecated 'jupyter-labhub' spawner to 'jupyterhub-singleuser'.
• Updated RStudio to v2023.12.1 Build 402

Fixed

• Fixed an issue with image streaming from Artifact Registry required switching to rebuilt jupyter-docker-stacks images without any WORKDIR statements (which can cause duplicate layers to be created). This is expected to only be a temporary measure.
• Fixed download and R package install issues in the RStudio console.

[2024-Winter] - 2023-12-19

Changed

• Upgraded to JupyterHub chart 3.2.1
• Upgraded to JupyterHub version 4.0.2
• Target Kubernetes version now 1.27.8

[2023-Autumn-breakfix] - 2023-12-21

Changed

• Using v2.5.3 of RTTL-admin application

Fixed

• Fixed an issue with user logins related to breaking changes in JupyterHub OAuthenticator which was preventing user login.

[2023-Summer] - 2023-06-09

Changed

• Using v2.5.2 of RTTL-admin application
• Switch PYTHONPATH setter to use python version in users .bash_profile instead

[2023-Spring-update] - 2023-03-20

Changed

• Using v2.5.1 of RTTL-admin application
• Append '~/.local/bin' to PATH for locally installed packages
• Create PYTHONPATH env variable with '~/.local/lib/python3.9/site-packages' to resolve first-time import errors with local pip installs.
• Removed 'sticky_bit' fix on RStudio deployments, as the underlying issues with Google image streaming appears to be resolved.

[2023-Spring] - 2023-03-02

Changed

• Using v2.5.0 of RTTL-admin application
• Upgraded to JupyterHub chart 2.0.0
• Upgraded to JupyterHub version 3.0.0
• Increased default body size for sending from the browser to the Juptyer server to 100MB.
• Chart updates to support Kubernetes 1.25 deprecations.

[2022-Autumn] - 2022-10-11

Changed

• Using v2.4.0 of RTTL-admin application.
• Change image pull policy to 'Always' for images with 'latest' tag.

Added

• Support for hub-shared NFS volumes

[2021-Winter] - 2021-11-23

Changed

• Using v2.3.3 of RTTL-admin application
• Upgraded to JupyterHub chart 1.1.3
• Upgraded to JupyterHub version 1.4.2
• Notebooks upgraded to JupyterLab version 3.2.3
• Added support for jupyter.rttl.uw.edu URLs

Notes:

• nbgitpuller links now working see: https://jupyterhub.github.io/nbgitpuller/link

[2021-Summer] - 2021-06-09

Added

• Support for RStudio Server Open Source notebooks via rsession-proxy
• RStudio version 1.2.5042

Changed

• Using v2.3.2 of RTTL-admin application:
• Upgraded to JupyterHub chart 0.11.1
• Upgraded to JupyterHub version 1.3.0
• Upgrades and fixes for supported notebook images, as well as improved documentation. See notebook changelogs for more details.
• Retention period for user data disks extended to one year after end of quarter.

[Initial release] - 2020-01-28

Added

• JupyterHub for teaching and learning deployed using JupyterHub chart 0.9.0-beta.4
• Using v2.2.0 of RTTL-admin application:
• Using JupyterHub chart 0.9.0-beta.4
• JupyterHub version 1.1.0
• User notebooks automatically created in Google Cloud Kubernetes cluster with up to 4GB RAM and 4 CPUs. 5GB user home directory as standard.
• Integrates with UW Canvas for user authentication using OAuth 2.0.
• Create and remove hub users automatically based on course roster in UW Canvas.
• Supported notebook images:
  • datascience (extends official base image)
  • r (extends official base image)
  • scipy (extends official base image)
  • tensorflow (extends official base image)
• All images include nbgitpuller