Symptom
On the EXEC3_PARALLEL test matrix (#21017), the QA - Stage Exec Smoke Test workflow's stage-exec-test jobs fail in both modes:
stage-exec-test (from-0, serial) ❌ and (from-0, parallel) ❌stage-exec-test (chaintip, serial) ❌ and (chaintip, parallel) ❌stage-exec-test (resume-nonchaintip, serial) ❌ and (resume-nonchaintip, parallel) ❌
Because these fail on the serial path too, the root cause is not an exec3-parallel race — it's a separate concern (re-exec-from-0 / chaintip sync behavior). It was explicitly scoped out of #21088 (which is the parallel-commitment correctness fix).
What to do
- Triage the
stage-exec-test failures starting from the serial-mode logs (since serial failing rules out the parallel-specific paths).
- Identify whether this is a pre-existing failure on
main or a regression introduced by a recent change.
- Fix or, if it's a known-flaky/infra issue, document and stabilize.
Related: #21017, #21088
Symptom
On the
EXEC3_PARALLELtest matrix (#21017), theQA - Stage Exec Smoke Testworkflow'sstage-exec-testjobs fail in both modes:stage-exec-test (from-0, serial)❌ and(from-0, parallel)❌stage-exec-test (chaintip, serial)❌ and(chaintip, parallel)❌stage-exec-test (resume-nonchaintip, serial)❌ and(resume-nonchaintip, parallel)❌Because these fail on the serial path too, the root cause is not an exec3-parallel race — it's a separate concern (re-exec-from-0 / chaintip sync behavior). It was explicitly scoped out of #21088 (which is the parallel-commitment correctness fix).
What to do
stage-exec-testfailures starting from the serial-mode logs (since serial failing rules out the parallel-specific paths).mainor a regression introduced by a recent change.Related: #21017, #21088