2.1.1rc2

[eschutho]

The release candidate:
https://dist.apache.org/repos/dist/dev/superset/2.1.1rc2/

Git tag for the release:
https://github.com/apache/superset/tree/2.1.1rc2

The Change Log for the release:
https://github.com/apache/superset/blob/2.1.1rc2/CHANGELOG.md

The Updating instructions for the release:
https://github.com/apache/superset/blob/2.1.1rc2/UPDATING.md

Public keys are available at:
https://www.apache.org/dist/superset/KEYS

[eschutho]

Please add any testing results that you find in the discussion below.

[sfirke]

I'm experiencing a regression that is a little hard to pin down. I copied my Superset metadata DB and created a new instance on the copy using the 2.1.1rc2 image.

A dashboard native filter is failing to load its first value by default. It's set up like this:

It fails to load any value:

This causes the dashboard to fail to load, with a spinning Superset logo. If I refresh the dashboard, it will now load, but the charts using that filter are broken because that value is referenced in a Jinja virtual dataset. Everything works if I specify a filter value manually, or change the native filter to use a specific value as default.

This behavior worked in 2.1.0.

UPDATE: I seem to have replicated this bug in the 2.1.1rc2 ephemeral environment, feel free to log in and take a look. And the bug is not present in 3.0.0rc2; there my existing dashboard works correctly.

[sfirke]

I might have gotten it replicated in the 2.1.1rc2 ephemeral, without Jinja. See this dashboard, you might need to log in first:

Let me know if there's a reason that filter config actually should be failing to select the first value by default?

[eschutho]

Thanks @sfirke! It looks like the vote for rc2 didn't pass, so we'll see if we can get a fix for this or revert the PR for rc3.

[eschutho]

Hi @sfirke, I git bisected with your dashboard (thank you!) and found the missing PR, so I cherried it in and built a new ephemeral. It seems to be working, if you want to try it out: http://54.202.87.4:8080/superset/dashboard/11/?native_filters_key=ALp-Gi1w5p7plsXY_Djux9tCxT1l6FluyeRKGFVNnBkjVJ9lWb0eAi90cInTXBhn

[sfirke]

Yes this did it! I applied the docker hub image pr-24224-dev to my test environment and that fixed this bug 🙌

[sfirke]

Awesome! I wonder if we can easily generate a docker image that I can plug into my own test environment to test there. I think if you create a "don't merge this" dummy PR it will build an image and send to Docker hub?

[eschutho]

Yeah, I think you can use this PR for the 2.1.1 release: #24224

[michael-s-molina]

2.1.1 RC3 discussion is out.