Fix /tmp file leak when API server streams large task logs#66450
Merged
Conversation
dheerajturaga
force-pushed
the
bugfix/tmp-file-leak
branch
from
574209e to
fe57e31
Compare
LogStreamAccumulator spills the in-flight log stream to a NamedTemporaryFile(delete=False) in /tmp once a task's output exceeds 5000 lines. Cleanup was hooked only into the stream generator's finally block, so the file leaked whenever the caller did not iterate the generator to completion: client disconnect on the NDJSON streaming endpoint, exception during list() materialisation on the JSON path, or the caller using only total_lines and never accessing .stream. A single production host accumulated ~39 GB of orphaned files from this code path, with individual files up to 2.5 GB on tasks emitting ~10M log lines.
dheerajturaga
force-pushed
the
bugfix/tmp-file-leak
branch
from
fe57e31 to
f3c7317
Compare
potiuk
approved these changes
May 10, 2026
potiuk
reviewed
May 10, 2026
1 task
potiuk
added a commit
that referenced
this pull request
May 10, 2026
…66667) (cherry picked from commit 31d4837) Co-authored-by: Dheeraj Turaga <[email protected]>
github-actions Bot
pushed a commit
to aws-mwaa/upstream-to-airflow
that referenced
this pull request
May 10, 2026
…s warnings (apache#66663) Failure to remove the spill file is the exact failure mode that caused the prior /tmp leak (~39 GB on a single host), so operators need it visible in their logs rather than buried at debug level. Follow-up to apache#66450. (cherry picked from commit 1abd959) Co-authored-by: Dheeraj Turaga <[email protected]>
potiuk
pushed a commit
that referenced
this pull request
May 10, 2026
…s warnings (#66663) (#66669) Failure to remove the spill file is the exact failure mode that caused the prior /tmp leak (~39 GB on a single host), so operators need it visible in their logs rather than buried at debug level. Follow-up to #66450. (cherry picked from commit 1abd959) Co-authored-by: Dheeraj Turaga <[email protected]>
potiuk
approved these changes
May 10, 2026
potiuk
left a comment
potiuk
left a comment
Member
There was a problem hiding this comment.
LGTM. The three-layer cleanup is overkill in the best sense — guarantees cleanup regardless of caller behaviour, with each layer testable in isolation. The module-level _safe_remove to break the strong-ref cycle and the idempotent _cleanup are nicely subtle. One non-blocking suggestion inline on the OSError log level.
Drafted-by: Claude Opus 4.7 (1M context); reviewed by @potiuk before posting
jason810496
pushed a commit
to jason810496/airflow
that referenced
this pull request
May 11, 2026
vatsrahul1001
pushed a commit
that referenced
this pull request
May 20, 2026
…66667) (cherry picked from commit 31d4837) Co-authored-by: Dheeraj Turaga <[email protected]>
vatsrahul1001
pushed a commit
that referenced
this pull request
May 20, 2026
…s warnings (#66663) (#66669) Failure to remove the spill file is the exact failure mode that caused the prior /tmp leak (~39 GB on a single host), so operators need it visible in their logs rather than buried at debug level. Follow-up to #66450. (cherry picked from commit 1abd959) Co-authored-by: Dheeraj Turaga <[email protected]>
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
2 participants
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.
LogStreamAccumulator spills the in-flight log stream to a
NamedTemporaryFile(delete=False) in /tmp once a task's output exceeds
5000 lines. Cleanup was hooked only into the stream generator's
finally block, so the file leaked whenever the caller did not iterate
the generator to completion: client disconnect on the NDJSON streaming
endpoint, exception during list() materialisation on the JSON path, or
the caller using only total_lines and never accessing .stream.
A single production host accumulated ~39 GB of orphaned files from this
code path, with individual files up to 2.5 GB on tasks emitting
~10M log lines.
Was generative AI tooling used to co-author this PR?
ClaudeCode Opus 4.7