fix: wheel corruption when running parallel tox processes #3667
+4
−12
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Hard links share the same inode/data blocks as the original file.
When parallel tox processes rebuild a wheel, the original file gets
overwritten, corrupting what other processes are reading via their
hard-linked session views.
Using shutil.copyfile() creates an independent copy. On modern
filesystems (APFS, Btrfs, XFS), this automatically uses copy-on-write
(reflink) which is nearly as fast as hard linking.
Fixes wheel corruption errors ("Could not find EOCD") when running
multiple tox processes in parallel.
Signed-off-by: Bernát Gábor <bgabor8@bloomberg.net>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When running multiple tox processes in parallel (e.g.,
tox -p autoor multiple CI jobs), users encounter wheel corruption errors:Root Cause
The session view mechanism used hard links (
os.link()) to create per-session copies of built wheels. Hard links share the same underlying file data (inode).Example of the problem:
.tox/.tmp/package/1/pkg-1.0.whl(hard link to original)Even though numbered directories (
.tmp/package/1/,.tmp/package/2/) prevent path collisions, all hard links point to the same inode, so modifying the original corrupts all views.Solution
Replace
os.link()withshutil.copyfile(), which creates an independent copy. On modern copy-on-write filesystems (APFS, Btrfs, XFS with reflink), this automatically uses reflink, making it nearly as fast as hard linking while maintaining data independence.