[GH-ISSUE #4237] Nightly build/test failure - 2026-09-19 #1061

New issue

Open

opened 2026-10-29 21:22:27 +00:00 by bsakaguchi · 1 comment

bsakaguchi commented

2026-10-29 21:22:27 +00:00

Owner

Originally created by @github-actions[bot] on GitHub (Sep 19, 2026).
Original GitHub issue: https://github.com/kvcache-ai/Mooncake/issues/4237

Nightly Failure Report

Run: https://github.com/kvcache-ai/Mooncake/actions/runs/35366974546
Branch: refs/heads/main
Timestamp: 2026-09-19T16:10:48.363Z

Please investigate the failed jobs in the workflow run linked above.

Originally created by @github-actions[bot] on GitHub (Sep 19, 2026). Original GitHub issue: https://github.com/kvcache-ai/Mooncake/issues/4237 ## Nightly Failure Report **Run**: https://github.com/kvcache-ai/Mooncake/actions/runs/35366974546 **Branch**: refs/heads/main **Timestamp**: 2026-09-19T16:10:48.363Z Please investigate the failed jobs in the workflow run linked above.

bsakaguchi added the

nightly-failure

label

2026-10-29 21:22:27 +00:00

bsakaguchi commented

2026-10-29 21:22:27 +00:00

Author

Owner

@he-yufeng commented on GitHub (Sep 20, 2026):

Tonight's reds are runner-side, not code: both failing lanes die before the build at Failed to update package lists / Failed to install system packages (16:14Z, both within ~30s of the apt step), and the tone-sglang lane only gets as far as retrying artifact downloads. That is apt/network flake on the runners, different from #4227's three lanes (FilereadWorkerPoolTest flake tracked in #4213 with fix in #4210, and the 8080 HTTP metadata bind in the Go sanitizer job).

Nothing in the diff range is implicated — a rerun should go green, and the two real issues from #4227 still stand on their own (the Fileread flake waits on #4210; the 8080 bind is worth a per-run unique port if it recurs).

@he-yufeng commented on GitHub (Sep 20, 2026): Tonight's reds are runner-side, not code: both failing lanes die before the build at `Failed to update package lists` / `Failed to install system packages` (16:14Z, both within ~30s of the apt step), and the tone-sglang lane only gets as far as retrying artifact downloads. That is apt/network flake on the runners, different from #4227's three lanes (FilereadWorkerPoolTest flake tracked in #4213 with fix in #4210, and the 8080 HTTP metadata bind in the Go sanitizer job). Nothing in the diff range is implicated — a rerun should go green, and the two real issues from #4227 still stand on their own (the Fileread flake waits on #4210; the 8080 bind is worth a per-run unique port if it recurs).

bsakaguchi referenced this issue

2026-10-29 21:31:54 +00:00

[PR #1061] [MERGED] [Store] add version checking between client and server #1985

bsakaguchi referenced this issue

2026-10-29 21:32:27 +00:00

[PR #1180] [CLOSED] [CI] add sglang e2e integration test #2057