Repository navigation
CI flake: Smoketests (Windows 4/4) #5778
Description
Activity
Diagnosis for the exact linked job/attempt: run
32394780835, job96514081285,Smoketests (Windows 4/4), attempt 1.GitHub's job API reports the linked job as
run_attempt: 1, matching the ticket's failedattempt 1 -> 2rerun. The workflow was publicCIfor PR #5571 (Add cause to init failures), head SHAb697c382e1546d6afab06ed29acd99c9c68eea9f.The job failed before Windows setup completed and before any smoketest partition ran. Job metadata shows:
runner: ghr-c95a7188 started: 2026-08-20T17:13:16Z completed: 2026-08-20T17:23:19Z conclusion: failure last completed setup step: Checkout sources last started step: Run dsherret/rust-toolchain-file@v1 Set default rust toolchain and later setup/test steps: pending Run smoketest partition (Windows): pendingThe job-specific log endpoint for
96514081285is still unavailable. A manual re-check returned AzureBlobNotFound:HTTP 404 The specified blob does not exist. X-Ms-Error-Code: BlobNotFoundRoot cause visible from the exact attempt: Windows CI/runner infrastructure failed during very early setup, while the Rust toolchain action was in progress. The job never reached Node, pnpm, emscripten, artifact download, cargo-nextest, or the Windows smoketest step, and GitHub also has no retained job log blob. This is not a SpacetimeDB smoketest assertion failure.
Suggested updated title:
CI infra flake: Windows Smoketests lost logs during rust-toolchain setupDuplicate analysis:
- Similar Windows Smoketests missing-log / runner-loss tickets exist: CI flake: lost connection with windows runner #5346, CI flake: Smoketests (Windows) #5545, CI flake: Smoketests (Windows) #5595, and CI flake: Smoketests (Windows) #5673. CI flake: Smoketests (Windows) #5595 is the closest by last-started step (
Run dsherret/rust-toolchain-file@v1), but it was closed not-planned after a human noted the run was canceled. - I also searched broad fallback terms including
BlobNotFound,The specified blob does not exist,Run dsherret/rust-toolchain-file@v1,Run ./.github/actions/setup-pnpm,Install emscripten (Windows), andWindows Smoketests setup. - Top duplicate candidate: none close enough to mark as duplicate. The related tickets are closed/not-planned runner-loss cases, and I did not find an active canonical Test Flakes issue for this exact missing-log setup-failure mechanism.
Leaving this open because there is no close duplicate candidate and no automatic duplicate-close allowlist target applies.
- Similar Windows Smoketests missing-log / runner-loss tickets exist: CI flake: lost connection with windows runner #5346, CI flake: Smoketests (Windows) #5545, CI flake: Smoketests (Windows) #5595, and CI flake: Smoketests (Windows) #5673. CI flake: Smoketests (Windows) #5595 is the closest by last-started step (
Closing this one out because the useful failure signal was lost while GitHub was having log/API issues. If this happens again when GitHub can serve the job logs, we want a fresh ticket with the exact failed job logs available so the failure can be diagnosed instead of guessed.
Detected by CI Gremlin after a failed GitHub Actions attempt succeeded on rerun.
Tests:
Smoketests (Windows 4/4)Error:
Could not fetch job logs: Command '['gh', 'api', '-X', 'GET', 'repos/clockworklabs/SpacetimeDB/actions/jobs/96514081285/logs']' returned non-zero exit status 1.
Failed job runs:
Timeout detection:
Context: