The GitHub Actions job "Backport Auto Label" on texera.git/portcompleted-reply-first has failed. Run started by GitHub user aglinxinyuan (triggered by aglinxinyuan).
Head commit for run: 905816c91cfa9920e8199086d3ca684b470fa140 / Xinyuan Lin <[email protected]> fix(amber): advance region executions in a later coordinator round, not inside portCompleted On every normal region termination the coordinator advanced region executions inline in the portCompleted handler's continuation. That advance terminates the completed region, which sends EndWorker to each of its workers - so EndWorker went out before the reply to the very portCompleted being handled, and both travel the same FIFO control channel to that worker. The worker then processed EndWorker while the reply was still queued behind it, its unprocessed-messages check rejected the termination, and the first attempt failed with 2 ERRORs + 2 WARNs + 2 stack traces before the 200 ms retry succeeded (the LoopIntegrationSpec flake). The advance now goes through the coordinator's own control channel: WorkflowExecutionManager.requestAdvanceRegionExecutions() sends CoordinatorInitiateAdvanceRegionExecutions to the coordinator itself, and its handler runs advanceRegionExecutions. A self-directed control message is transmitted and then received before it is handled, so it lands a whole message behind every reply the requesting round still owed - EndWorker can no longer overtake a reply. Everything else is left alone: portCompleted still replies at the end of its statistics/bookkeeping continuation, so a failure there is still returned to the sending worker as a ControlError, and an advance failure still reaches the client as a FatalError. The worker-side unprocessed-messages check is untouched. Report URL: https://github.com/apache/texera/actions/runs/30349646184 With regards, GitHub Actions via GitBox
