匿名野线程池导致oom

1 背景

1.1 生产环境几次所有操作出现

[java.lang.RuntimeException: java.lang.OutOfMemoryError: Failed creating thread: pthread_create() failed, maxproc limit reached] with root cause

        java.lang.OutOfMemoryError: Failed creating thread: pthread_create() failed, maxproc limit reached

 

线程监控显示,线程数10w资源耗尽,在一个时间点,线程以每秒10个开始激增,直到搞垮系统

 

1.2 threaddump显示为匿名线程池和匿名线程,无法直接定位到线程池/线程定义的代码位置

$ jstack 18652
2020-12-16 12:01:32
Full thread dump Java HotSpot(TM) 64-Bit Server VM (25.221-b27 mixed mode):

"DestroyJavaVM" #12 prio=5 os_prio=0 tid=0x00000000031e2000 nid=0x47cc waiting on condition [0x0000000000000000]
java.lang.Thread.State: RUNNABLE

"pool-1-thread-1" #11 prio=5 os_prio=0 tid=0x000000001b11e800 nid=0x44f8 waiting on condition [0x000000001bbfe000]
java.lang.Thread.State: WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x000000078aeaed10> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442) thread一直在跑,只不过被linkedBlockingqueue.take 阻塞了,所以没回收thread,进而没回收threadpool

at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1074)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1134)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)

 

1.3 threadump为服务器垮掉以后的dump,意味着thread已跑完,没有堆栈

 

2 排查

2.1 初步

推算有人在for与while中新建线程池

 

2.2 排查

2.2.1 使用脚本监控进程号线程激增时自动进行threaddump,谋求其堆栈定位,只要线程在跑,没理由我们抓不到它。

2.2.2 写一个程序遍历所有代码,在for及while后的200个字符内,寻找线程池关键词,但如果在循环里面的是函数,则抓不到,而且是函数的可能性很大,该方案作为次要方案

 

3 结果

jstack定位到堆栈,麻痹还真有人在for里面新建线程池

 

4 扩展

4.1 局部线程池不会回收?

根据

线程池造成服务器内存泄漏

线程池的原理

这两个文章,我们认为,即使是局部线程池也不会回收,因为线程一直在跑,见 线程池的原理 中的while

我们用以下代码(不进入git)调查研究:

import java.util.concurrent.Executor;
import java.util.concurrent.Executors;

public class TestMain {
    public static void main(String []f) {
        Executor executor = Executors.newFixedThreadPool((4));
        InnerThread innerThread = new InnerThread();
        executor.execute(innerThread);
    }

    private static class InnerThread implements Runnable {

        @Override
        public void run() {

        }
    }
}

  

main没有退出,与守护非守护无关,只是因为线程一直在跑好不好,只是被阻塞队列wait了,jstack:

$ jstack 18652
2020-12-16 12:01:32
Full thread dump Java HotSpot(TM) 64-Bit Server VM (25.221-b27 mixed mode):

"DestroyJavaVM" #12 prio=5 os_prio=0 tid=0x00000000031e2000 nid=0x47cc waiting on condition [0x0000000000000000]
java.lang.Thread.State: RUNNABLE

"pool-1-thread-1" #11 prio=5 os_prio=0 tid=0x000000001b11e800 nid=0x44f8 waiting on condition [0x000000001bbfe000]
java.lang.Thread.State: WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x000000078aeaed10> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442) ***********************************************************该dump日志与  线程池的原理 中自己写的线程池一致

不一致,见 线程池的原理 2021.3.9  4.4

at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1074)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1134)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)

 

 

4.2 线程/线程池应务必自定义名称

 

4.3 线程池的shutdown

如果局部方法中,线程池shutdown了,不会出现泄露导致oom的结局

import java.util.concurrent.Executor;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

public class TestMain {
    public static void main(String []f) {
        ExecutorService executor = Executors.newFixedThreadPool((4));
        InnerThread innerThread = new InnerThread();
        executor.execute(innerThread);
        executor.shutdown();
    }

    private static class InnerThread implements Runnable {

        @Override
        public void run() {

        }
    }
}

  

这里就牵扯出,如何shutdown一个线程池,具体可见  线程池的原理终止while循环的thread,让thread跑完回收,进而回收局部的线程池

 

 

4.4 线程池没有任务时的阻塞方式,有2种

见上方4.4标明处

posted on 2020-12-16 11:28  silyvin  阅读(571)  评论(0编辑  收藏  举报