Repository navigation
fix: gate serving readiness and admission on PD producer availability - #247
Closed
gspark-etri wants to merge 1 commit into
Closed
gspark-etri wants to merge 1 commit into
gspark-etri wants to merge 1 commit into
Conversation
This was referenced Oct 23, 2026
gspark-etri
added a commit
that referenced
this pull request
Oct 30, 2026
…#246, #249) (#282) From feat/249-pd-stage-metrics (#239, #247, #250): SubprocessPrefillWorker failure settling, producer readiness gating /ready and admission, producer stage-duration metrics; only the /ready hunk of llmd.py. Fixes on top: - SubprocessPrefillWorker.close() closes stdin and waits for a clean exit before SIGTERM, then SIGKILL (prefill_server exits on stdin EOF; SIGTERM first skipped SDK cleanup that releases the device). - ProcessPrefillWorker.ready: a dead child now fails readiness and admission is rejected with 503 instead of quarantining the batcher. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Collaborator
Author
|
Landed on optimus-rm via #282. Closing; the branch is archived locally (experiment records and evidence included). |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Background
After its required subprocess producer died, batched serving still returned
/ready200/ready=true and admitted completions that returned500/non-retryable. A real tiny Torch subprocess/HTTP reproduction confirms this on ea09aca.Changes
Expose subprocess readiness and propagate it through configured worker pools. Use that availability consistently in batched readiness and request admission: terminally unavailable producers now return retryable503 before scheduling. Explicit healthy replacement restores readiness; a single request error does not poison a live producer. No implicit rerouting, replay or automatic restart.
Verification
Compatibility and limits
Stacked on #243. Non-PD services and legacy/custom workers without the optional readiness property retain prior behavior. Pools conservatively require all configured members because routing still includes them. Already-admitted work retains its existing drain/error handling; the decoder is not quarantined solely because a producer can be replaced.
This is real CPU subprocess/HTTP evidence. Historical hardware fault evidence #244/#245 is pinned to its previous later-request500 behavior; the corrected admission contract is503. The change does not add automatic production restart or establish performance/overlap benefits.
Related Issue
Closes #246; follows #238/#239 and #244/#245.