Should (can) we move domain specific hubs here?

[jhamman]

In the weekly Pangeo developers call today, this topic came up. Should we move the domain specific deployments (atmos, ocean, polar, hydroshare, etc.) to this repo? We seem to be a bit fragmented in our development and maintenance of our various jupyterhubs and this could be a way to centralize some of our knowledge/resources.

cc @dsludwig @rabernat @raphaeldussin @NicWayand @bartnijssen @rsignell-usgs

[rabernat]

Yes!

[rabernat]

What should live here? Here is one example

• deployments / dev.pangeo.io / config
  • common
  • develop
  • prod
  • staging
• deployments / ocean.pangeo.io / config
  • common
  • staging
  • prod
• deployments / atmos.pangeo.io / config
  • common
  • staging
  • prod
• deployments / polar.pangeo.io / config
  • common
  • staging
  • prod

Charts:

• pangeo-deploy: currently this is a custom helm chart, but it would be ideal to instead link to an independent, up-to-date helm chart (see discussion in https://github.com/pangeo-data/pangeo/issues/511)

[rabernat]

The advantage of unifying these clusters into a single repo would be that the maintainers can more easily see what the other admins are doing.

[guillaumeeb]

I imagine that the CI magic will need some modifications to avoid restarting other clusters at every commit.

This will also make commit history and PR a little messy. But that should be manageable.

[rabernat]

I imagine that the CI magic will need some modifications to avoid restarting other clusters at every commit.

That is all handled automatically by hubploy: https://github.com/yuvipanda/hubploy

This will also make commit history and PR a little messy. But that should be manageable.

It will let the cluster managers see what the others are doing and encourage collaboration / interaction. Much like we already have on this repo.

[guillaumeeb]

Nice, I did not take the time to dive into hubploy!

And yes, for the second point, I agree that the benefits seem to exceed any drawback.

[rabernat]

Here's how I propose we handle this. Let's schedule a virtual hackathon, in which the people involved all block out a couple of hours uninterrupted to refactor things in a coordinated fashion.

The earliest I could do this would be the first week of February (when I'm finally back in New York).

If you're interested in participating, could you please indicate your availability here: https://www.when2meet.com/?7432306-tRQmH

[arokem]

Can I add my PanNeuro (learning2learn.pangeo.io) deployment configuration here too?