[RAWL-3481] Use 8 zstd workers and compression level 3 for create-layer#122
Merged
Conversation
a9dccc9 to
f050f09
Compare
Flush zstd writer between files to bound memory with multi-threaded compression. Without periodic flushes, multi-threaded zstd buffers input asynchronously, causing memory usage to grow up to the full uncompressed layer size. Declare cpu:8 execution_requirements for zstd OCI layer actions. ocitool create-layer spawns 8 zstd worker threads internally. Without declaring this cost, Bazel over-subscribes the machine by scheduling too many concurrent zstd layer builds. Add conditional execution_requirements so Bazel's scheduler accounts for the actual CPU usage (8 for zstd, default 1 for gzip).
f050f09 to
17e8031
Compare
bcmyers
approved these changes
Apr 30, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
cpu:8execution requirement on zstd layer actions in Bazel so the scheduler reserves cores accordingly