>
>
> "Add Support for Python 3.8 (#8836
> <https://github.com/apache/airflow/pull/8836>)(#8823
> <https://github.com/apache/airflow/pull/8823>)" is in the changelog but
> seems like *1.10.11-python3.8* tag is missing in DockerHub?
>

We do not support Python 3.8 in 1.10.* line:

https://github.com/apache/airflow/blob/master/README.md#requirements

J.



> On Sun, 12 Jul 2020 at 16:22, Jarek Potiuk <[email protected]>
> wrote:
>
> > FYI -> the images for 1.10.11 are in DockerHub now. They also include the
> > "latest" tag - which means that anyone running `*docker pull
> > apache/airflow*` will get the python3.6 variant of 1.10.11 version of
> > image. Also *apache/airflow:1.10.11* points to 3.6 variant.
> >
> > J.
> >
> > [image: Screenshot from 2020-07-12 10-19-23.png]
> >
> > On Sat, Jul 11, 2020 at 8:43 PM Jarek Potiuk <[email protected]>
> > wrote:
> >
> >> Hello Everyone,
> >>
> >> Before we agree on release cadence of the Charts I have a proposal for
> >> Prod images. They have not been released yesterday with 1.10.11.
> >>
> >> The relevant Dockerfile changes are part of the officially voted and
> >> released sources and those sources are enough to build the image and
> push
> >> it to the registry. In fact - I am already doing so.
> >>
> >> I think this is perfectly fine with ASF release policy that we release
> >> the images without extra voting in such case (it's built using the
> released
> >> sources):
> >> http://www.apache.org/legal/release-policy.html#compiled-packages
> >>
> >> I think we should continue this way and do not introduce separate
> cadence
> >> to release the docker images - unless we decide to make an ad-hoc
> release
> >> at some point.
> >>
> >> I believe for the image (unlike for the Helm Chart), such ad-hoc
> releases
> >> are not really possible. We will test the future Helm chart against
> >> those released versions of images, and we should maintain backward
> >> compatibility.
> >>
> >> With my proposal (which I am going to talk about next week at my
> >> "Production Image talk" and ask for feedback from the users) that if
> >> someone wants to customize the image, they should rebuild the image on
> >> their own using either stable (if cautions) or master (if adventurous)
> >> Dockerfile. This way they can get the image much better size-optimized.
> >> With the latest Breeze release, it's super-easy to build your own image
> >> using our Dockerfile - with your own dependencies both on Python. apt
> dev
> >> and apt runtime dependencies.
> >>
> >> J.
> >>
> >> --
> >>
> >> Jarek Potiuk
> >> Polidea <https://www.polidea.com/> | Principal Software Engineer
> >>
> >> M: +48 660 796 129 <+48660796129>
> >> [image: Polidea] <https://www.polidea.com/>
> >>
> >>
> >
> > --
> >
> > Jarek Potiuk
> > Polidea <https://www.polidea.com/> | Principal Software Engineer
> >
> > M: +48 660 796 129 <+48660796129>
> > [image: Polidea] <https://www.polidea.com/>
> >
> >
>


-- 

Jarek Potiuk
Polidea <https://www.polidea.com/> | Principal Software Engineer

M: +48 660 796 129 <+48660796129>
[image: Polidea] <https://www.polidea.com/>

Reply via email to