microsoft / microsoft/durabletask-java

Functions - Terminating a running orchestrator throws IllegalStateException

未关闭
#237 1 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

还没有人认领这个 Issue。

P2
主要语言
Java
星标
29
派生
18
平均合并
1 天 10 小时
30 天内合并 PR
2

描述

When using this repo with Azure Functions with the MSSQL backend, and using the client.terminate() API to terminate an orchestraton that is still running, the orchestration will throw an IllegalStateException:

[2025-08-22T16:32:56.240Z] Executed 'Functions.LongRunningOrchestrator' (Failed, Id=c6d1d1b9-32ee-4849-a841-5902bbbff2c6, Duration=51ms)
[2025-08-22T16:32:56.241Z] System.Private.CoreLib: Exception while executing function: Functions.LongRunningOrchestrator. System.Private.CoreLib: Result: Failure
Exception: IllegalStateException: The orchestrator has already completed
Stack: java.lang.IllegalStateException: The orchestrator has already completed
[2025-08-22T16:32:56.242Z]      at com.microsoft.durabletask.Helpers.throwIfOrchestratorComplete(Helpers.java:31)
[2025-08-22T16:32:56.243Z]      at com.microsoft.durabletask.TaskOrchestrationExecutor$ContextImplTask.completeInternal(TaskOrchestrationExecutor.java:799)
[2025-08-22T16:32:56.244Z]      at com.microsoft.durabletask.TaskOrchestrationExecutor$ContextImplTask.fail(TaskOrchestrationExecutor.java:787)
[2025-08-22T16:32:56.244Z]      at com.microsoft.durabletask.TaskOrchestrationExecutor.execute(TaskOrchestrationExecutor.java:71)
[2025-08-22T16:32:56.245Z]      at com.microsoft.durabletask.OrchestrationRunner.loadAndRun(OrchestrationRunner.java:136)
[2025-08-22T16:32:56.247Z]      at com.microsoft.durabletask.OrchestrationRunner.loadAndRun(OrchestrationRunner.java:69)
[2025-08-22T16:32:56.249Z]      at com.microsoft.durabletask.OrchestrationRunner.loadAndRun(OrchestrationRunner.java:42)
[2025-08-22T16:32:56.249Z]      at com.microsoft.durabletask.azurefunctions.internal.middleware.OrchestrationMiddleware.invoke(OrchestrationMiddleware.java:37)
[2025-08-22T16:32:56.250Z]      at com.microsoft.azure.functions.worker.chain.InvocationChain.doNext(InvocationChain.java:21)
[2025-08-22T16:32:56.251Z]      at com.microsoft.azure.functions.worker.broker.JavaFunctionBroker.invokeMethod(JavaFunctionBroker.java:125)
[2025-08-22T16:32:56.252Z]      at com.microsoft.azure.functions.worker.handler.InvocationRequestHandler.execute(InvocationRequestHandler.java:34)
[2025-08-22T16:32:56.253Z]      at com.microsoft.azure.functions.worker.handler.InvocationRequestHandler.execute(InvocationRequestHandler.java:10)
[2025-08-22T16:32:56.254Z]      at com.microsoft.azure.functions.worker.handler.MessageHandler.handle(MessageHandler.java:44)
[2025-08-22T16:32:56.255Z]      at com.microsoft.azure.functions.worker.JavaWorkerClient$StreamingMessagePeer.lambda$onNext$0(JavaWorkerClient.java:94)
[2025-08-22T16:32:56.256Z]      at java.base/java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:539)
[2025-08-22T16:32:56.257Z]      at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
[2025-08-22T16:32:56.258Z]      at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)
[2025-08-22T16:32:56.258Z]      at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)
[2025-08-22T16:32:56.259Z]      at java.base/java.lang.Thread.run(Thread.java:842)
[2025-08-22T16:32:56.260Z] .
[2025-08-22T16:32:56.266Z] d88614d4-5cea-453a-9bbd-0a423a9059da: Function 'LongRunningOrchestrator (Orchestrator)' failed with an error. Reason: Microsoft.Azure.WebJobs.Host.FunctionInvocationException

Code:

/**
 * Orchestration and activity functions for simulating long-running orchestration and termination.
 */
public class TerminateOrchestration {

    /**
     * This is the orchestrator function.
     */
    @FunctionName("LongRunningOrchestrator")
    public List<String> longRunningOrchestrator(
            @DurableOrchestrationTrigger(name = "ctx") TaskOrchestrationContext ctx,
            final ExecutionContext context) {
        Logger logger = context.getLogger();
        logger.info("Starting long-running orchestration.");
        List<String> outputs = new ArrayList<>();

        // Call our fake activity 100,000 times to simulate an orchestration that might run for >= 10,000s (2.7 hours)
        for (int i = 0; i < 100000; i++) {
            outputs.add(ctx.callActivity("SimulatedWorkActivity", 100, String.class).await());
        }

        return outputs;
    }

    /**
     * This is the activity function.
     */
    @FunctionName("SimulatedWorkActivity")
    public String simulatedWorkActivity(
            @DurableActivityTrigger(name = "sleepMs") int sleepMs,
            final ExecutionContext context) {
        context.getLogger().info("Sleeping for " + sleepMs + "ms.");
        try {
            Thread.sleep(sleepMs);
        } catch (InterruptedException e) {
            Thread.currentThread().interrupt();
            return "Sleep interrupted.";
        }
        return "Slept for " + sleepMs + "ms.";
    }

    /**
     * HTTP-triggered function to terminate an orchestration instance.
     */
    @FunctionName("TerminateInstance")
    public HttpResponseMessage run(
            @HttpTrigger(name = "req", methods = {HttpMethod.GET, HttpMethod.POST}, authLevel = AuthorizationLevel.ANONYMOUS) HttpRequestMessage<Optional<String>> request,
            @DurableClientInput(name = "durableContext") DurableClientContext durableContext,
            @BindingName("instanceId") String instanceId,
            final ExecutionContext context) {
        DurableTaskClient client = durableContext.getClient();
        String reason = "Long-running orchestration was terminated early.";
        try {
            client.terminate(instanceId, reason);
            return request.createResponseBuilder(HttpStatus.OK).build();
        } catch (Exception ex) {
            context.getLogger().severe(ex.getMessage());
            return request.createResponseBuilder(HttpStatus.BAD_REQUEST)
                    .header("Content-Type", "text/plain")
                    .body(ex.getMessage())
                    .build();
        }
    }
}

贡献指南

打开贡献指南

从这里开始

  1. 先读完整个 Issue,再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库,在一个分支上完成修改。
  4. 提交 Pull Request,并在描述里引用这个 Issue 编号。

调研方向

首先按照堆栈跟踪所示,从 DurableTaskClient.terminate(instanceId, reason) 跟踪到 TaskOrchestrationExecutor.execute 的终止路径,尤其关注 ContextImplTask.fail 和 Helpers.throwIfOrchestratorComplete。使用提供的 TerminateOrchestration 函数重现该问题,并验证终止正在运行的 orchestration 不再导致 IllegalStateException。

由索引模型根据 Issue 内容生成。

评估

技术栈
azure, java
领域
backend, distributed-systems
Issue 类型
缺陷
难度
3/5
预计耗时
1-2 天
活跃度
停滞
描述清晰度
基本清楚
新手友好度
35/100

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。