From 9fc24238ea5c15c910cf59033ca1d4ca0f65cf55 Mon Sep 17 00:00:00 2001 From: Luke Curley Date: Fri, 9 Oct 2026 18:18:14 -0700 Subject: [PATCH] quest: plan the cluster burst overrun flake Co-Authored-By: Claude Opus 5.5 --- quest/m1/test-flakes-2/README.md | 1 + .../m1/test-flakes-2/cluster-burst-overrun.md | 20 +++++++++++++++++++ 2 files changed, 21 insertions(+) create mode 100644 quest/m1/test-flakes-2/cluster-burst-overrun.md diff --git a/quest/m1/test-flakes-2/README.md b/quest/m1/test-flakes-2/README.md index b0fc70f158..7018882a64 100644 --- a/quest/m1/test-flakes-2/README.md +++ b/quest/m1/test-flakes-2/README.md @@ -36,4 +36,5 @@ Public API: none. Wire: none. - [Import catalog finish](/quest/m1/test-flakes-2/import-catalog-finish.md) - `moq-cli`'s subprocess EOF catalog-finish test holds up under load with event-based fixture coordination - [Relay restart rebind](/quest/m1/test-flakes-2/relay-restart-rebind.md) - the crash drill restarts on its original UDP address under concurrent load - [Impaired handshake](/quest/m1/test-flakes-2/impaired-handshake.md) - the impaired cluster drills' clients never time out while connecting +- [Cluster burst overrun](/quest/m1/test-flakes-2/cluster-burst-overrun.md) - the impaired cluster burst drill always overruns its bottleneck, so it always exercises it - [Media audio tone](/quest/m1/media-audio-tone.md) - the `just test media` audio-tone check passes under load, fixed at its cause diff --git a/quest/m1/test-flakes-2/cluster-burst-overrun.md b/quest/m1/test-flakes-2/cluster-burst-overrun.md new file mode 100644 index 0000000000..20ad283657 --- /dev/null +++ b/quest/m1/test-flakes-2/cluster-burst-overrun.md @@ -0,0 +1,20 @@ +# [S] Cluster burst drill always overruns its bottleneck + +## Goal + +`bursts_cross_a_cluster::impaired` (`rs/moq-relay/tests/drills.rs`) always +exercises its bottleneck: every run has at least one burst that overruns +the 100 kbit/s link, so its "no burst overran a bottleneck, so none was +exercised" precondition never fails. + +## Plan + +Found in #5160's loaded runs (2026-10-09): 5 of about 450 runs of the drills +binary at load average 40-60 failed that precondition. Decided 2026-10-09: +keep the precondition, since a run that never overruns tests nothing. Find +why a burst sometimes fits (for example, pacing that lets the shaper drain +between frames under load), then size or schedule the bursts from the +shaper's capacity so the overrun is guaranteed. No retry, and no lowering +of what counts as an overrun. + +Public API: none. Wire: none.