apache/dolphinscheduler · error · BaseException

The TaskExecutor id ${taskInstanceId} is not exist.

Error message

The TaskExecutor id ${taskInstanceId} is not exist.

What it means

The lifecycle event reporter's channel runner (run -> handleTaskExecutionEventChannel) looks up the ITaskExecutor for a pending event's taskInstanceId before reporting it to the master. If the event is still pending/retrying but the executor is no longer in the repository, it throws BaseException "The TaskExecutor id %d is not exist." — the event has no owner to report for.

Source

Thrown at dolphinscheduler-task-executor/src/main/java/org/apache/dolphinscheduler/task/executor/eventbus/TaskExecutorLifecycleEventRemoteReporter.java:206

            log.warn("TaskExecutor is not exists: {}", taskExecutorId);
        }
    }

    private void handleTaskExecutionEventChannel(final ReportableTaskExecutorLifecycleEventChannel reportableTaskExecutorLifecycleEventChannel) {
        if (reportableTaskExecutorLifecycleEventChannel.isEmpty()) {
            return;
        }
        while (!reportableTaskExecutorLifecycleEventChannel.isEmpty()) {
            final IReportableTaskExecutorLifecycleEvent headEvent = reportableTaskExecutorLifecycleEventChannel.peek();
            try (
                    final TaskExecutorMDCUtils.MDCAutoClosable ignore =
                            TaskExecutorMDCUtils.logWithMDC(headEvent.getTaskInstanceId())) {
                try {
                    if (isTaskExecutorEventNeverSent(headEvent) || isRetryIntervalExceeded(headEvent)) {
                        final Optional<ITaskExecutor> taskExecutorOptional =
                                taskExecutorRepository.get(headEvent.getTaskInstanceId());
                        if (!taskExecutorOptional.isPresent()) {
                            throw new BaseException(String.format("The TaskExecutor id %d is not exist.",
                                    headEvent.getTaskInstanceId()));
                        }
                        final String masterAddress =
                                taskExecutorOptional.get().getTaskExecutionContext().getWorkflowInstanceHost();
                        taskExecutorEventRemoteReporterClient.reportTaskExecutionEventToMaster(masterAddress,
                                headEvent);
                        continue;
                    }
                    if (log.isDebugEnabled()) {
                        log.debug(
                                "The ReportableTaskExecutorLifecycleEvent: {} latest send time: {} doesn't exceeded retry interval",
                                headEvent,
                                headEvent.getLatestReportTime());
                    }
                    break;
                } catch (Exception ex) {
                    log.error("Send TaskExecutionEvent: {} to master error will retry after {} mills",
                            headEvent,

View on GitHub (pinned to 02eac45a1b)

Solutions

1. Verify events are drained/reported before removing executors from the repository
2. Investigate why reporting failed repeatedly until the executor was removed (network to master, master down)
3. Reduce event backlog by ensuring the master address is reachable and the reporter client healthy
4. On occurrence, treat the stale event as droppable or purge orphaned events after executor removal

if (taskExecutorRepository.get(event.getTaskInstanceId()).isEmpty()) {
    // event is stale: drop or move to dead-letter instead of reporting
    return;
}

try {
    reporterChannel.run();
} catch (BaseException e) {
    if (e.getMessage().contains("is not exist")) {
        log.warn("Dropping stale lifecycle event; executor removed");
    } else throw e;
}

When it happens

Trigger: A task execution event was queued (or is being retried after failures) and, before it is successfully reported, the corresponding task executor was removed from taskExecutorRepository (task finished/killed/cleaned up), so repository.get(taskInstanceId) returns empty.

Common situations: Master restart or task cleanup racing with a backlog of unreported events; events retained past the task's lifecycle; memory pressure evicting executors while retrying failed reports.

Understand the failure class

Background: "Not found" and "does not exist" errors: why "Task not found", "No such folder", and "Can't find" fire when a lookup comes back empty — this error's family across 14 libraries.

Related errors

• ${this} has already started
• DELETE_QUEUE_BY_ID_ERROR
• "Take over task: " + taskExecution.getName() + " from execut
• "Pause task: " + taskExecution.getName() + " from executor f
• "Kill task: " + taskExecution.getName() + " from executor fa

AI-assisted analysis of apache/dolphinscheduler@02eac45a1b (2026-09-06). Data as JSON: /api/errors/0147f1eae8322be5. Report an issue: GitHub.