Question

我在Wildfly 8.0.0.Final上运行以下代码：

@Stateful
public class MyJPABean<T> {

    @PersistenceContext(unitName = "myUnitName")
    private EntityManager em;

    @TransactionAttribute(TransactionAttributeType.REQUIRED)
    public void write(T entity) {
        em.persist(entity);
        em.flush();
    }

    @TransactionAttribute(TransactionAttributeType.REQUIRED)
    public void update(T entity) {
        em.merge(entity);
    }
}

写入和更新方法都是并行调用的，因为它们是直接从处理HTTP请求的另一个类调用的：

@Path("api")
public class MyApiController {

    private MyJPABean<MyJpaModel> jpaBean;

    @POST
    @Consumes(MediaType.APPLICATION_JSON)
    @Produces(MediaType.APPLICATION_JSON)
    public void handlePost(MyJpaModel myJpaModel) {
       jpaBean.write(myJpaModel);
    }
}

某些“write”方法调用是成功的，但有时，当多个线程并行调用“write”方法时，它会开始挂起在em.persist（实体）上。

在PostgreSQL服务器端，我可以看到处于状态的事务：“在事务中空闲”，带有查询：“select nextval（'hibernate_sequence'）”或插入从我刚试图保留的实体创建的查询。

发生此问题时，我无法再保留实体。一切都挂在em.persist（实体）上。要么我无法杀死野蝇进程，我必须用-9信号杀死它。

我有什么问题吗？

一段时间后（超时发生时）我可以在应用程序日志中看到以下警告：

16:44:30,590 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012117: TransactionReaper::check timeout for TX 0:ffff7f000101:-231d9b5d:55030494:2d in state  RUN
16:44:30,591 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper Worker 0) ARJUNA012095: Abort of action id 0:ffff7f000101:-231d9b5d:55030494:2d invoked while multiple threads active within it.
16:44:30,591 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper Worker 0) ARJUNA012108: CheckedAction::check - atomic action 0:ffff7f000101:-231d9b5d:55030494:2d aborting with 1 threads active!
16:44:30,598 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012117: TransactionReaper::check timeout for TX 0:ffff7f000101:-231d9b5d:55030494:35 in state  RUN
16:44:31,090 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012117: TransactionReaper::check timeout for TX 0:ffff7f000101:-231d9b5d:55030494:2d in state  CANCEL
16:44:31,092 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012378: ReaperElement appears to be wedged: com.arjuna.ats.arjuna.coordinator.TwoPhaseCoordinator.afterCompletion(TwoPhaseCoordinator.java:487)
com.arjuna.ats.arjuna.coordinator.TwoPhaseCoordinator.afterCompletion(TwoPhaseCoordinator.java:463)
com.arjuna.ats.arjuna.coordinator.TwoPhaseCoordinator.cancel(TwoPhaseCoordinator.java:118)
com.arjuna.ats.arjuna.AtomicAction.cancel(AtomicAction.java:215)
com.arjuna.ats.arjuna.coordinator.TransactionReaper.doCancellations(TransactionReaper.java:377)
com.arjuna.ats.internal.arjuna.coordinator.ReaperWorkerThread.run(ReaperWorkerThread.java:78)

16:44:31,100 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012117: TransactionReaper::check timeout for TX 0:ffff7f000101:-231d9b5d:55030494:35 in state  SCHEDULE_CANCEL
16:44:31,594 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012117: TransactionReaper::check timeout for TX 0:ffff7f000101:-231d9b5d:55030494:2d in state  CANCEL_INTERRUPTED
16:44:31,595 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012120: TransactionReaper::check worker Thread[Transaction Reaper Worker 0,5,main] not responding to interrupt when cancelling TX 0:ffff7f000101:-231d9b5d:55030494:2d -- worker marked as zombie and TX scheduled for mark-as-rollback
16:44:31,596 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper Worker 1) ARJUNA012095: Abort of action id 0:ffff7f000101:-231d9b5d:55030494:35 invoked while multiple threads active within it.
16:44:31,597 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012110: TransactionReaper::check successfuly marked TX 0:ffff7f000101:-231d9b5d:55030494:2d as rollback only
16:44:31,597 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper Worker 1) ARJUNA012108: CheckedAction::check - atomic action 0:ffff7f000101:-231d9b5d:55030494:35 aborting with 1 threads active!
16:44:31,600 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012117: TransactionReaper::check timeout for TX 0:ffff7f000101:-231d9b5d:55030494:35 in state  CANCEL
16:44:31,602 WARN  [com.arjuna.ats.arjuna] (Transaction Reaper) ARJUNA012378: ReaperElement appears to be wedged: org.apache.log4j.AppenderSkeleton.doAppend(AppenderSkeleton.java:231)

作为jdbc驱动程序，使用“postgresql-9.2-1004.jdbc41”。 PostgreSQL服务器版本为“9.3”。我在类路径上也有c3p0，但即使我删除了这个依赖项，问题仍然存在。

Answer 1

我已经使用JAX-RS，EJB和JPA准备了一个示例应用程序 - https://github.com/aparnachaudhary/prototypes/tree/master/jaxrs-db-access。我使用JMeter脚本来访问具有多个线程的REST端点。但我无法重现PostgreSQL 9.3，postgresql-9.3-1103.jdbc41.jar和WildFly 8.1.0.Final的问题。

Wildfly 8.0.0.Final + EntityManager + PostgreSQL - 在事务中空闲

1 个答案: