[smg] Work around sglang's notorious orphan process problem#16756
Merged
[smg] Work around sglang's notorious orphan process problem#16756
Conversation
sglang is notorious for getting stuck during model loading and leaving orphan child processes (TP workers) hanging around holding GPU memory. There's even a dedicated killall_sglang.sh script for this very reason. This commit works around these issues in e2e tests: - Kill entire process group on terminate() instead of just parent process, since sglang doesn't clean up its children when the parent dies - Increase LAUNCH_STAGGER_DELAY from 5s to 10s because sglang can't handle multiple servers loading weights simultaneously without choking
Contributor
|
Warning You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again! |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
sglang is notorious for getting stuck during model loading and leaving orphan child processes (TP workers) hanging around holding GPU memory. There's even a dedicated killall_sglang.sh script for this very reason.
This commit works around these issues in e2e tests:
Checklist
Review Process
/tag-run-ci-label,/rerun-failed-ci,/tag-and-rerun-ci) or contact authorized users to do so.