You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@airflow.apache.org by Jarek Potiuk <ja...@potiuk.com> on 2025/01/12 11:02:59 UTC

[ANNOUNCE] Yanking Celery provider 3.9.0

Hello here,

While working on backporting "pull_request_target" workflow to 2.10, I have
found that 3.9.0 version of provider is pretty much completely broken on
Airflow 2 - you can see those failures here:

https://github.com/apache/airflow/actions/runs/12731808997/job/35486973244?pr=45527

This has already been fixed in main in
https://github.com/apache/airflow/pull/45255 (thanks Bugra)   as response
to this issue https://github.com/apache/airflow/issues/45364 - but 3.9.1
with the fix will be released in the next wave.

Since We do not have Airflow 3 yet, and 3.9.0 just does not work on Airflow
2, yanking it is the best thing to do.

J.

Re: [ANNOUNCE] Yanking Celery provider 3.9.0

Posted by Amogh Desai <am...@gmail.com>.
Yeah those tests are really nice and thanks for isolating them.

Makes debugging and re running them much easier than before and also
resources are saved.

Thanks & Regards,
Amogh Desai


On Sun, Jan 12, 2025 at 4:37 PM Jarek Potiuk <ja...@potiuk.com> wrote:

> BTW. It's really cool we have those e2e k8s tests (and for all the python
> versions and combinations of executors and k8s versions). Even if they take
> quite some time to run and they are somewhat flaky sometimes - they are the
> last resort to detect and find things like that :) .
>
> J.
>
> On Sun, Jan 12, 2025 at 12:02 PM Jarek Potiuk <ja...@potiuk.com> wrote:
>
> > Hello here,
> >
> > While working on backporting "pull_request_target" workflow to 2.10, I
> > have found that 3.9.0 version of provider is pretty much completely
> broken
> > on Airflow 2 - you can see those failures here:
> >
> >
> >
> https://github.com/apache/airflow/actions/runs/12731808997/job/35486973244?pr=45527
> >
> > This has already been fixed in main in
> > https://github.com/apache/airflow/pull/45255 (thanks Bugra)   as
> response
> > to this issue https://github.com/apache/airflow/issues/45364 - but 3.9.1
> > with the fix will be released in the next wave.
> >
> > Since We do not have Airflow 3 yet, and 3.9.0 just does not work on
> > Airflow 2, yanking it is the best thing to do.
> >
> > J.
> >
> >
>

Re: [ANNOUNCE] Yanking Celery provider 3.9.0

Posted by Jarek Potiuk <ja...@potiuk.com>.
BTW. It's really cool we have those e2e k8s tests (and for all the python
versions and combinations of executors and k8s versions). Even if they take
quite some time to run and they are somewhat flaky sometimes - they are the
last resort to detect and find things like that :) .

J.

On Sun, Jan 12, 2025 at 12:02 PM Jarek Potiuk <ja...@potiuk.com> wrote:

> Hello here,
>
> While working on backporting "pull_request_target" workflow to 2.10, I
> have found that 3.9.0 version of provider is pretty much completely broken
> on Airflow 2 - you can see those failures here:
>
>
> https://github.com/apache/airflow/actions/runs/12731808997/job/35486973244?pr=45527
>
> This has already been fixed in main in
> https://github.com/apache/airflow/pull/45255 (thanks Bugra)   as response
> to this issue https://github.com/apache/airflow/issues/45364 - but 3.9.1
> with the fix will be released in the next wave.
>
> Since We do not have Airflow 3 yet, and 3.9.0 just does not work on
> Airflow 2, yanking it is the best thing to do.
>
> J.
>
>