-
-
Notifications
You must be signed in to change notification settings - Fork 177
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
BUG: some failing builds on windows with micromamba #2101
Comments
Also:
I wonder if there's a lingering open handle somewhere 🤔 |
I think that's a red herring honestly, or rather just a symptom of a previous problem. And the fact that the exact same recipe builds if reverting the infra changes is an argument in favour of that IMO. |
Yea, I think the same infra change is behind the two things. I had to add a bit of a hack to reuse the cache in this line, so maybe that's messing with the permissions somehow. |
Very interesting findings in conda-forge/arrow-cpp-feedstock#1432. If I configure the Windows pipeline so micromamba and conda-build don't use the same cache, it works. This is not ideal because conda will have to redownload some files for the conda-build envs. If I make them share the cache in some way (either by moving the micromamba cache over to conda's location as currently done in conda-smithy, or by configuring micromamba to use conda's location from the beginning), we hit the error in this issue. It looks like micromamba is linking things in a way that prevents conda / conda-build from reusing those cached files. |
Since we switched to micromamba, I'm seeing some rare but reproducible failures for particularly heavy builds.
They look like
seen here:
The problem is 100% reproducible (in arrow), and after undoing the micromamba-changes on arrow, the build passed again (see here).
CC @jaimergp @jjerphan @JohanMabille
The text was updated successfully, but these errors were encountered: