UpgradeΒΆ

Here we suppose a user would like to upgrade to a version <version>. Here we have outlined the following steps. These docs outline two paths: upgrading qhub locally on your own terminal or via CI.

If you are deploying QHub on your terminal locally you will need to upgrade via pip.

pip install --upgrade qhub=<version>

If you are using CI you will need to modify the CI provider code to reflect the new CI version:

  • github-actions .github/workflows/qhub-ops.yaml change the line pip install qhub==<version>

  • gitlab-ci .gitlab-ci.yml change the line to pip install qhub==<version>

If you are using the default images being provided by QHub you will want to upgrade the images being used. As of v0.3.9 there are docker images built for each release. If you are using your own QHub images it may be a difficult process to upgrade the images. We strive to minimize the number of changes to the images.

...
default_images:
  jupyterhub: quansight/qhub-jupyterhub:v<version>
  jupyterlab: quansight/qhub-jupyterlab:v<version>
  dask_worker: quansight/qhub-dask-worker:v<version>
  dask_gateway: quansight/qhub-dask-gateway:v<version>
...
profiles:
  jupyterlab:
    - display_name: Small Instance
      ...
      kubespawner_override:
        ...
        image: quansight/qhub-jupyterlab:v<version>
    ...
  dask_worker:
    Small Worker:
      ...
      image: quansight/qhub-dask-worker:v<version>

We try to make QHub backwards compatible with configuration in qhub-config.yaml. There is a schema validator pydantic for QHub which will report any errors in the configuration file.

If you are deploying QHub on your terminal locally run:

qhub deploy --config qhub-config.yaml

If you are deploying via CI commit the two changes to qhub-config.yaml and the CI at the same time.