Fix symlink file creation overhead #17478
Closed
Conversation
This file contains 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
coeuvre
changed the title
brentleyjones
approved these changes
Feb 13, 2023
|
@bazel-io flag |
bazel-io
added
the
potential release blocker
sgowroji
added
the
team-Remote-Exec
coeuvre
force-pushed
the
fix-symlink-check
branch
from
d78e9be to
90fdc67
Compare
|
@bazel-io fork 6.1.0 |
bazel-io
removed
the
potential release blocker
hvadehra
pushed a commit
that referenced
this pull request
Feb 14, 2023
The cost of symlink action scales with the size of input because Bazel re-calculates the digest of the output by following the symlink in `actuallyCompleteAction` (#14125). However, the re-calculation is redundant because the digest was already computed by Bazel when checking the outputs of the generating action. Bazel should be smart enough to reuse the result. There is a global cache in Bazel for digest computation. Symlink action didn't make use of the cache because it uses the path of symlink as key to look up the cache. This PR changes to use the path of input file (i.e. target path) to query the cache to avoid recalculation. For a large target (700MB), the time for symlink action is reduced from 2000ms to 1ms. Closes #17478. PiperOrigin-RevId: 509524641 Change-Id: Id3c9dc07d68758770c092f6307e2433dad40ba10
keertk
added a commit
that referenced
this pull request
Feb 20, 2023
* Fix symlink file creation overhead The cost of symlink action scales with the size of input because Bazel re-calculates the digest of the output by following the symlink in `actuallyCompleteAction` (#14125). However, the re-calculation is redundant because the digest was already computed by Bazel when checking the outputs of the generating action. Bazel should be smart enough to reuse the result. There is a global cache in Bazel for digest computation. Symlink action didn't make use of the cache because it uses the path of symlink as key to look up the cache. This PR changes to use the path of input file (i.e. target path) to query the cache to avoid recalculation. For a large target (700MB), the time for symlink action is reduced from 2000ms to 1ms. Closes #17478. PiperOrigin-RevId: 509524641 Change-Id: Id3c9dc07d68758770c092f6307e2433dad40ba10 * Update ActionMetadataHandler.java * Create OutputPermissions.java --------- Co-authored-by: Chi Wang <[email protected]> Co-authored-by: keertk <[email protected]>
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.
The cost of symlink action scales with the size of input because Bazel re-calculates the digest of the output by following the symlink in
actuallyCompleteAction(#14125). However, the re-calculation is redundant because the digest was already computed by Bazel when checking the outputs of the generating action. Bazel should be smart enough to reuse the result.There is a global cache in Bazel for digest computation. Symlink action didn't make use of the cache because it uses the path of symlink as key to look up the cache. This PR changes to use the path of input file (i.e. target path) to query the cache to avoid recalculation.