A frozen node is twice as expensive as an absent one, and marks its peer's completed work as failed
メンテナーはふだん 5 日以内に返信
@Noah-Tervalon-Nvidia がすでに取り組んでいます。
2026年9月11日 から。
評価
この issue はまだ評価されていません。
説明
Two observations from the same failure mode, measured on a two-node cluster (Ubuntu + RTX 3060, macOS, v0.1.1-463). Both matter for a cluster that includes a laptop, since a laptop that sleeps mid-request is exactly this shape: the socket is accepted, nothing answers.
Setup
24 concurrent chat/completions, max_tokens: 200, google/gemma-4-12b resident on both nodes.
| Condition, same load | Wall clock |
|---|---|
| Both nodes healthy | 37.8 s |
| Single node, bypassing the router entirely | 72.8 s |
Peer frozen (SIGSTOP on its nvpair processes) |
155.6 s |
Every request returned 200 in all three. Reliability is not the problem — the cost is.
A frozen peer is 2.1× worse than not having it at all, and 4.1× worse than a healthy cluster. When I instead killed the peer's processes outright, the same load finished in 28.3 s with no failures: a closed socket gives an immediate RST and the proxy moves on in milliseconds. It is specifically the accepted-but-silent socket that hurts.
For context on the eviction path: after the peer was killed, discovery had dropped it from the node list within 75 s — I sampled too coarsely to pin it more precisely. During that window each request pays its own connection timeout.
The second half: false failures
When the frozen peer resumed, its job history marked several of the other node's workloads as
Failed — node serveri stopped reporting this workload
Those workloads had all completed successfully; the requesting node returned 200 for all 24. The peer's view had simply gone stale while it was stopped, and on resume it reaped them as failed. The timestamps match the resume to the second.
That leaves the job history — the only place a user sees what the cluster did — showing failures that did not happen, precisely in the situation where something unusual was going on and they would be looking at it.
Not filed as a defect
Neither is a crash and both may be deliberate. But if a laptop is a first-class node, asleep with the socket accepted and nothing answering is its most likely state, not an edge case. Two directions that would help, if either is welcome:
- a connection-level timeout distinctly shorter than the request timeout, so an unresponsive candidate is abandoned quickly while a slow-but-live generation is not
- reaping a peer's workloads as
unknownrather thanfailedwhen the gap was on our side
Happy to open a PR for either if you have a preferred shape.
- 主要言語
- Go
- スター
- 1.6k
- フォーク
- 266
- 平均マージ
- 3日 6時間
- マージ済み PR（30日）
- 22
環境構築
- Dockerfile・Docker Compose ファイルなし
- プルリクエストのテンプレートあり
- コントリビューションガイドを読む
はじめの一歩
- issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
- 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
- リポジトリをフォークし、ブランチを切って変更します。
- issue 番号を参照したプルリクエストを送ります。
NVIDIA/Personal-AI-Router のほかの issue
-
難易度 5/5 1週間以上 初心者へのやさしさ 20/100
NVIDIA/Personal-AI-Router#165 ·
メンテナーはふだん 5 日以内に返信
-
bug
難易度 4/5 3〜5日 初心者へのやさしさ 46/100
NVIDIA/Personal-AI-Router#164 ·
メンテナーはふだん 5 日以内に返信
-
enhancement
難易度 4/5 1週間以上 初心者へのやさしさ 12/100
NVIDIA/Personal-AI-Router#162 ·
メンテナーはふだん 5 日以内に返信
-
enhancement
難易度 4/5 3〜5日 初心者へのやさしさ 25/100
NVIDIA/Personal-AI-Router#154 · コメント 1 件 ·
メンテナーはふだん 5 日以内に返信
-
[Feature]: [Ollama] Alert the user about model pull failures対応中かも @ckelseynv が 3 日前に担当しました。 オープンenhancement
難易度 3/5 1〜2日 初心者へのやさしさ 56/100
NVIDIA/Personal-AI-Router#152 · コメント 1 件 · 担当者 1 名 ·
メンテナーはふだん 5 日以内に返信
NVIDIA/Personal-AI-Router の issue をすべて見る
似ている issue
-
難易度 2/5 1〜3時間 初心者へのやさしさ 70/100
メンテナーはふだん 1 日以内に返信
-
難易度 2/5 1〜3時間 初心者へのやさしさ 73/100
メンテナーはふだん 1 日以内に返信
-
難易度 2/5 1〜3時間 初心者へのやさしさ 63/100
メンテナーはふだん 1 日以内に返信
-
product/backlog
難易度 2/5 1〜3時間 初心者へのやさしさ 75/100
TencentBlueKing/bk-nodemgr#3645 ·
メンテナーはふだん 1 日以内に返信
-
難易度 2/5 1〜3時間 初心者へのやさしさ 74/100
modelcontextprotocol/go-sdk#1387 ·
メンテナーはふだん 1 日以内に返信