Conversation
|
|
||
| ## [Unreleased] | ||
|
|
||
| ## [23.11.0] - 2023-11-23 |
There was a problem hiding this comment.
I have to admit this is a bit confusing to me. What do you think of removing it again?
Also we should add entries for this PR
| gcsfs==2023.9.2 | ||
| geomet==0.2.1.post1 | ||
| gevent==23.9.1 | ||
| gevent==22.10.2 |
| graphql-core==3.2.3 | ||
| graphviz==0.20.1 | ||
| greenlet==2.0.2 | ||
| greenlet==3.0.2 |
There was a problem hiding this comment.
I'm no Python guy but I assume we need to be careful here making a breaking change with a Airflow dependency.
https://github.com/python-greenlet/greenlet/blob/master/CHANGES.rst#300a1-2023-06-21 sounds like Python 2.12 addition and 2.7, 3.5 and 3.6 removal as well as C++ bump.
WDYT of pulling in 3.0.2 only for the experimental ARM builds to not disturb the "production" x86 image?
Maybe we could also back-port the fixes (whatever they are) to greenlet 2.0.2
Options:
- Merge as-is
- Pull in
3.0.2only for the experimental ARM builds to not disturb the "production" x86 image - Back-port the fixes (whatever they are) to greenlet 2.0.2
- Try to bump
greenletupstream in Airflow. This does not work forgeventhowever, as this PR downgrades it. This is more of a long-term solution, as we have a dependency on upstream releases.
Arch meeting 2024-01-03 outcome:
- No
- Maybe, but only as long as arm support is experimental
- No
- Yeah, at least raise an issue in Airflow
| gcsfs==2023.9.2 | ||
| geomet==0.2.1.post1 | ||
| gevent==23.9.1 | ||
| gevent==22.10.2 |
There was a problem hiding this comment.
Because... although they claim it gevent 23.9.1 is not building together with greenlet 3.0.2 on arm64.
There was a problem hiding this comment.
Fwiw I'm with Sebastian and Malte. We need to be careful with changing dependency versions
| && pushd /stackable && curl --fail -L https://repo.stackable.tech/repository/packages/statsd_exporter/statsd_exporter-${STATSD_EXPORTER}.linux-${ARCH}.tar.gz | tar -xzC /stackable && ln -s statsd_exporter-${STATSD_EXPORTER}.linux-${ARCH}/statsd_exporter statsd_exporter && popd | ||
|
|
There was a problem hiding this comment.
Not a fan of push and pop d. Can we just use "curl -o .." or workdir or at least use wrapping (" && \") for better readability?
Co-authored-by: Sebastian Bernauer <sebastian.bernauer@stackable.de>
Co-authored-by: Sebastian Bernauer <sebastian.bernauer@stackable.de>
Co-authored-by: Sebastian Bernauer <sebastian.bernauer@stackable.de>
Co-authored-by: Sebastian Bernauer <sebastian.bernauer@stackable.de>
Co-authored-by: Sebastian Bernauer <sebastian.bernauer@stackable.de>
Co-authored-by: Sebastian Bernauer <sebastian.bernauer@stackable.de>
|
outdated, closing it. |
Description
This PR fixes some problems with the arm build of our products.
Notable mentions:
Definition of Done Checklist