Release Process for QHub¶
Currently Qhub is not fully automated. This makes it especially important to manually check the features. This is a minimal set of features that are guaranteed with each release.
qhub deploy and
qhub destroy of QHub on all
[ ] Azure
[ ] Amazon Web Services
[ ] Digital Ocean
[ ] Google Cloud Platform
[ ] Existing Kubernetes Cluster
Check services of Qhub on each.
[ ] Login and launch jupyterlab notebook and able to run basic python calculations
[ ] Launch dask-gateway
able to scale to 2 workers
run basic dask calculation
view dask gateway dashboard
[ ] Conda-store environments are created and available in jupyterlab notebook as kernels (dask, dashboard)
[ ] Create basic panel dashboard
In order to create a release:
Make sure the
RELEASE.mdis up to date with change, bug fixes, and breaking changes and move all the
Upcoming Releaseinto a new section titled
Release <version> - <month>/<day>/<year>
__version__ = "<version>"
Commit these changes to
Finally create a Release on QHub. The tag should be
v<version>off of the branch
main. Use the
RELEASE.mdto get the title
Release <version> - <month>/<day>/<year>and set the text description to the
RELEASE.mdfor the given version. Click
If this worked a new version will be uploaded to pypi for QHub
This should also trigger the all appropriate Docker images to be built, tagged, and pushed up to DockerHub.