Closed jhamman closed 5 years ago
I would love to test out some hopefully-improved "base" environments for pangeo somewhere. The current "base" pangeo environment is quite fragile, with a pool of five channels to pull from and numerous pinned package versions.
Instead, the environment can use conda-forge except for a few specific packages (e.g. intake, holoviews), which removes the need for pinning. We have a version with the same packages as the current pangeo.pydata.org, and a version with pyviz included
These containers end up being about 0.5GB larger than the existing containers (3GB->3.5GB):
REPOSITORY TAG IMAGE ID CREATED SIZE
rsignell/pangeo-gcs-notebook-geoviews 2018-10-22 aa5ac36bbc7c 14 hours ago 3.497 GB
rsignell/pangeo-gcs-notebook 2018-10-22 3f849ba56d19 17 hours ago 3.39 GB
pangeo/notebook ede11f6 70ce76c45be3 4 weeks ago 2.944 GB
Would one approach be for @dsludwig (or someone) to create a dev.pangeo.io
and then folks submit PRs to the Dockerfile
/ environment.yaml
and associated config to try different experiments?
A new dev instance setup here: https://github.com/pangeo-data/dev.pangeo.io-deploy
The branch develop
deploys to dev.pangeo.io
, staging
deploys to staging.pangeo.io
and prod
deploys to hub.pangeo.io
.
@rsignell-usgs you have write privileges to that repository
On Tue, Oct 23, 2018 at 6:27 PM Derek Ludwig notifications@github.com wrote:
A new dev instance setup here: https://github.com/pangeo-data/dev.pangeo.io-deploy The branch develop deploys to dev.pangeo.io, stagingdeploys to staging.pangeo.ioandproddeploys tohub.pangeo.io`.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/pangeo-data/pangeo/issues/441#issuecomment-432444701, or mute the thread https://github.com/notifications/unsubscribe-auth/AASszIb-rML3B2yCjisx5GlEifgUeJgbks5un5ftgaJpZM4X0U6f .
This is done so I'm going to close.
We are now creating many pangeos. This is great! But unlike when we all shared a single JupyterHub, we don't have a centralized place to test out new functionality.
This has come up in conversation with @yuvipanda in #401 and @rsignell-usgs in #145 (among other places).
My thought is that a dev/demo cluster would be of use to those of us actively developing/testing generic pangeo topics. It may also help us centralize our development whilst democratizing our deployments.
Thoughts from concerned parties (@dsludwig, @yuvipanda, @mrocklin, @rsignell-usgs, @rabernat)?