单线程Java应用程序启动超过70个线程

时间：2015-04-05 23:59:04

标签： java linux jvm openjdk single-threaded

在我的计算服务器上启动单线程应用程序时，我刚遇到（openJDK）JVM的一个非常奇怪的行为：

JVM启动了很多（非常多）线程，大大增加了所有处理器的CPU负载，我没有合适的解释[见下图]。

我还测试了一些其他内容：

  1. 我也尝试在拥有~200GB RAM的Amazon EC2实例上运行它，并且只用少量线程就可以得到类似的结果。
  2. 我没有使用java -Xms4096M -Xmx100000M -jar ...启动应用，而是尝试使用较小的参数，例如java -Xms1024M -Xmx1024M -jar ...，但没有任何更改。

    3. 还有其他人已经经历过这样的行为吗？

    甚至更好：有人可以向我解释这里发生了什么吗？

    http://i.stack.imgur.com/E9XtQ.png

    修改 我现在已经尝试了一些你建议的东西，并且确实得到了一些有价值且无价值的输出。 jstack PID不起作用，建议我使用jstack -F PID，其主要流程的输出如下。 kill -3 PID不会输出任何内容。

    Attaching to process ID 39041, please wait...
Debugger attached successfully.
Server compiler detected.
JVM version is 25.40-b25
Deadlock Detection:

No deadlocks found.

Thread 39090: (state = BLOCKED)


Thread 39089: (state = BLOCKED)
 - java.lang.Object.wait(long) @bci=0 (Interpreted frame)
 - java.lang.ref.ReferenceQueue.remove(long) @bci=59, line=143 (Interpreted frame)
 - java.lang.ref.ReferenceQueue.remove() @bci=2, line=164 (Interpreted frame)
 - java.lang.ref.Finalizer$FinalizerThread.run() @bci=36, line=209 (Interpreted frame)


Thread 39088: (state = BLOCKED)
 - java.lang.Object.wait(long) @bci=0 (Interpreted frame)
 - java.lang.Object.wait() @bci=2, line=502 (Interpreted frame)
 - java.lang.ref.Reference$ReferenceHandler.run() @bci=36, line=157 (Interpreted frame)


Thread 39043: (state = BLOCKED)
 - java.lang.AbstractStringBuilder.<init>(int) @bci=6, line=68 (Compiled frame)
 - java.lang.StringBuilder.<init>(java.lang.String) @bci=8, line=112 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=526, line=99 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=413, line=86 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=80, line=39 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=136, line=43 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=676, line=112 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=654, line=111 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=654, line=111 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=654, line=111 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=654, line=111 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=654, line=111 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=654, line=111 (Compiled frame)
 - de.unisb.react.stud.algorithm.BMCLoopConverter.convert(gov.nasa.ltl.trans.Formula, int, int, int[]) @bci=654, line=111 (Compiled frame)
 - de.unisb.react.stud.algorithm.PMC.LoopConvert(gov.nasa.ltl.trans.Formula, int, de.unisb.react.stud.algorithm.Tree) @bci=69, line=274 (Compiled frame)
 - de.unisb.react.stud.algorithm.PMC.main(java.lang.String[]) @bci=207, line=97 (Interpreted frame)

    我不确定它是属于kill还是jstack，但是我在tmux中发现了这个：

    2015-04-06 10:11:49
Full thread dump Java HotSpot(TM) 64-Bit Server VM (25.40-b25 mixed mode):

"Service Thread" #23 daemon prio=9 os_prio=0 tid=0x00007f0e54040000 nid=0x98c5 runnable [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C1 CompilerThread17" #22 daemon prio=9 os_prio=0 tid=0x00007f0e54039800 nid=0x98c4 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C1 CompilerThread16" #21 daemon prio=9 os_prio=0 tid=0x00007f0e54037800 nid=0x98c3 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C1 CompilerThread15" #20 daemon prio=9 os_prio=0 tid=0x00007f0e54035000 nid=0x98c2 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C1 CompilerThread14" #19 daemon prio=9 os_prio=0 tid=0x00007f0e54033000 nid=0x98c1 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C1 CompilerThread13" #18 daemon prio=9 os_prio=0 tid=0x00007f0e54031000 nid=0x98c0 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C1 CompilerThread12" #17 daemon prio=9 os_prio=0 tid=0x00007f0e5402f000 nid=0x98bf waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread11" #16 daemon prio=9 os_prio=0 tid=0x00007f0e5402d000 nid=0x98be waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread10" #15 daemon prio=9 os_prio=0 tid=0x00007f0e5402b000 nid=0x98bd waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread9" #14 daemon prio=9 os_prio=0 tid=0x0000000001bd9000 nid=0x98bc runnable [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread8" #13 daemon prio=9 os_prio=0 tid=0x0000000001bd7000 nid=0x98bb waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread7" #12 daemon prio=9 os_prio=0 tid=0x0000000001bd5000 nid=0x98ba waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread6" #11 daemon prio=9 os_prio=0 tid=0x0000000001bd3000 nid=0x98b9 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread5" #10 daemon prio=9 os_prio=0 tid=0x0000000001bd1000 nid=0x98b8 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread4" #9 daemon prio=9 os_prio=0 tid=0x0000000001bc6800 nid=0x98b7 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread3" #8 daemon prio=9 os_prio=0 tid=0x0000000001bc4800 nid=0x98b6 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread2" #7 daemon prio=9 os_prio=0 tid=0x0000000001bc0000 nid=0x98b5 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread1" #6 daemon prio=9 os_prio=0 tid=0x0000000001bbe000 nid=0x98b4 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C2 CompilerThread0" #5 daemon prio=9 os_prio=0 tid=0x0000000001bbb800 nid=0x98b3 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"Signal Dispatcher" #4 daemon prio=9 os_prio=0 tid=0x0000000001bb9000 nid=0x98b2 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"Finalizer" #3 daemon prio=8 os_prio=0 tid=0x0000000001b89000 nid=0x98b1 in Object.wait() [0x00007f0e5b5ba000]
   java.lang.Thread.State: WAITING (on object monitor)
        at java.lang.Object.wait(Native Method)
        - waiting on <0x00007f0f73829c78> (a java.lang.ref.ReferenceQueue$Lock)
        at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:143)
        - locked <0x00007f0f73829c78> (a java.lang.ref.ReferenceQueue$Lock)
        at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:164)
        at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:209)

"Reference Handler" #2 daemon prio=10 os_prio=0 tid=0x0000000001b87000 nid=0x98b0 in Object.wait() [0x00007f0e5b6bb000]
   java.lang.Thread.State: WAITING (on object monitor)
        at java.lang.Object.wait(Native Method)
        - waiting on <0x00007f0f73829cc0> (a java.lang.ref.Reference$Lock)
        at java.lang.Object.wait(Object.java:502)
        at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:157)
        - locked <0x00007f0f73829cc0> (a java.lang.ref.Reference$Lock)

"main" #1 prio=5 os_prio=0 tid=0x000000000151d800 nid=0x9883 runnable [0x00007f27cbc8c000]
   java.lang.Thread.State: RUNNABLE
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convertLoop(BMCLoopConverter.java:156)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:118)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.BMCLoopConverter.convert(BMCLoopConverter.java:111)
        at de.unisb.react.stud.algorithm.PMC.LoopConvert(PMC.java:274)
        at de.unisb.react.stud.algorithm.PMC.main(PMC.java:97)

"VM Thread" os_prio=0 tid=0x0000000001b82000 nid=0x98af runnable

"GC task thread#0 (ParallelGC)" os_prio=0 tid=0x0000000001533000 nid=0x9884 runnable

"GC task thread#1 (ParallelGC)" os_prio=0 tid=0x0000000001535000 nid=0x9885 runnable

"GC task thread#2 (ParallelGC)" os_prio=0 tid=0x0000000001536800 nid=0x9886 runnable

"GC task thread#3 (ParallelGC)" os_prio=0 tid=0x0000000001538800 nid=0x9887 runnable

"GC task thread#4 (ParallelGC)" os_prio=0 tid=0x000000000153a000 nid=0x9888 runnable

"GC task thread#5 (ParallelGC)" os_prio=0 tid=0x000000000153c000 nid=0x9889 runnable

"GC task thread#6 (ParallelGC)" os_prio=0 tid=0x000000000153d800 nid=0x988a runnable

"GC task thread#7 (ParallelGC)" os_prio=0 tid=0x000000000153f800 nid=0x988b runnable

"GC task thread#8 (ParallelGC)" os_prio=0 tid=0x0000000001541000 nid=0x988c runnable

"GC task thread#9 (ParallelGC)" os_prio=0 tid=0x0000000001543000 nid=0x988d runnable

"GC task thread#10 (ParallelGC)" os_prio=0 tid=0x0000000001544800 nid=0x988e runnable

"GC task thread#11 (ParallelGC)" os_prio=0 tid=0x0000000001546800 nid=0x988f runnable

"GC task thread#12 (ParallelGC)" os_prio=0 tid=0x0000000001548000 nid=0x9890 runnable

"GC task thread#12 (ParallelGC)" os_prio=0 tid=0x0000000001548000 nid=0x9890 runnable                                                                                       [31/1020]

"GC task thread#13 (ParallelGC)" os_prio=0 tid=0x000000000154a000 nid=0x9891 runnable

"GC task thread#14 (ParallelGC)" os_prio=0 tid=0x000000000154b800 nid=0x9892 runnable

"GC task thread#15 (ParallelGC)" os_prio=0 tid=0x000000000154d800 nid=0x9893 runnable

"GC task thread#16 (ParallelGC)" os_prio=0 tid=0x000000000154f000 nid=0x9894 runnable

"GC task thread#17 (ParallelGC)" os_prio=0 tid=0x0000000001551000 nid=0x9895 runnable

"GC task thread#18 (ParallelGC)" os_prio=0 tid=0x0000000001552800 nid=0x9896 runnable

"GC task thread#19 (ParallelGC)" os_prio=0 tid=0x0000000001554800 nid=0x9897 runnable

"GC task thread#20 (ParallelGC)" os_prio=0 tid=0x0000000001556000 nid=0x9898 runnable

"GC task thread#21 (ParallelGC)" os_prio=0 tid=0x0000000001558000 nid=0x9899 runnable

"GC task thread#22 (ParallelGC)" os_prio=0 tid=0x0000000001559800 nid=0x989a runnable

"GC task thread#23 (ParallelGC)" os_prio=0 tid=0x000000000155b800 nid=0x989b runnable

"GC task thread#24 (ParallelGC)" os_prio=0 tid=0x000000000155d000 nid=0x989c runnable

"GC task thread#25 (ParallelGC)" os_prio=0 tid=0x000000000155f000 nid=0x989d runnable

"GC task thread#26 (ParallelGC)" os_prio=0 tid=0x0000000001560800 nid=0x989e runnable

"GC task thread#27 (ParallelGC)" os_prio=0 tid=0x0000000001562800 nid=0x989f runnable

"GC task thread#28 (ParallelGC)" os_prio=0 tid=0x0000000001564000 nid=0x98a0 runnable

"GC task thread#29 (ParallelGC)" os_prio=0 tid=0x0000000001566000 nid=0x98a1 runnable

"GC task thread#30 (ParallelGC)" os_prio=0 tid=0x0000000001567800 nid=0x98a2 runnable

"GC task thread#31 (ParallelGC)" os_prio=0 tid=0x0000000001569800 nid=0x98a3 runnable

"GC task thread#32 (ParallelGC)" os_prio=0 tid=0x000000000156b000 nid=0x98a4 runnable

"GC task thread#33 (ParallelGC)" os_prio=0 tid=0x000000000156d000 nid=0x98a5 runnable

"GC task thread#34 (ParallelGC)" os_prio=0 tid=0x000000000156e800 nid=0x98a6 runnable

"GC task thread#35 (ParallelGC)" os_prio=0 tid=0x0000000001570800 nid=0x98a7 runnable

"GC task thread#36 (ParallelGC)" os_prio=0 tid=0x0000000001572000 nid=0x98a8 runnable

"GC task thread#37 (ParallelGC)" os_prio=0 tid=0x0000000001574000 nid=0x98a9 runnable

"GC task thread#38 (ParallelGC)" os_prio=0 tid=0x0000000001575800 nid=0x98aa runnable

"GC task thread#39 (ParallelGC)" os_prio=0 tid=0x0000000001577800 nid=0x98ab runnable

"GC task thread#40 (ParallelGC)" os_prio=0 tid=0x0000000001579000 nid=0x98ac runnable

"GC task thread#41 (ParallelGC)" os_prio=0 tid=0x000000000157b000 nid=0x98ad runnable

"GC task thread#42 (ParallelGC)" os_prio=0 tid=0x000000000157c800 nid=0x98ae runnable

"VM Periodic Task Thread" os_prio=0 tid=0x00007f0e54042800 nid=0x98c6 waiting on condition

JNI global references: 13

Heap                                                                                                                                                                       [163/1020]
 PSYoungGen      total 22755328K, used 10392421K [0x00007f1f96300000, 0x00007f27b9800000, 0x00007f27b9800000)
  eden space 11377664K, 91% used [0x00007f1f96300000,0x00007f22107d97a8,0x00007f224ca00000)
  from space 11377664K, 0% used [0x00007f224ca00000,0x00007f224ca00000,0x00007f2503100000)
  to   space 11377664K, 0% used [0x00007f2503100000,0x00007f2503100000,0x00007f27b9800000)
 ParOldGen       total 68267008K, used 68266986K [0x00007f0f4f800000, 0x00007f1f96300000, 0x00007f1f96300000)
  object space 68267008K, 99% used [0x00007f0f4f800000,0x00007f1f962fa968,0x00007f1f96300000)
 Metaspace       used 3156K, capacity 4240K, committed 4352K, reserved 8192K

    我检查了生成高CPU负载的正在运行的进程的PID - ＆gt;它们属于GC线程。 我还不确定单个线程程序如何产生如此多的垃圾，以至于它让所有GC线程都忙碌......但至少这是一个起点。 完全解释和调试代码（由于它是一个研究项目，这有点复杂）显然超出了范围。

    可能是关于该主题的最后一个问题：是否可以调整JVM产生的GC线程数？ 我现在将开始调试垃圾收集，例如使用-verbose:gc -XX:+PrintGCTimeStamps -XX:+PrintGCDetails，如果您知道有用的工具或有任何提示，请告诉我。 ;）

1 个答案:

答案 0 :(得分：1)

我能给你的唯一答案是：

  1. 尽管其他答案可能意味着什么，但您确实有证据表明出现了问题。在查看操作系统级线程时，Java线程与操作系统线程之间存在一对一的对应关系......至少在Linux上，与当前的JVM一致。

  2. 这不是JVM的错。 Java不会自发地创建大量线程;例如你看到~70岁。

    （根据Java版本，JVM参数等，您可能会看到十个左右的内部线程。但是大部分时间它们都将处于空闲状态...不会像监控似乎建议的那样一直运行。）

  3. 摆弄Java堆和堆栈大小参数不会影响启动的线程数...以一种可能解决问题的方式。

  4. 在不同的平台（EC2）上运行不太可能有所作为。

    5. 您需要查看应用程序正在执行的操作以及它正在使用的第三方库。除非您提供更多详细信息，否则我们无法帮助您。

    但是，获取线程转储（使用jstack或kill -3 PID）应提供一些线索，了解这些线程是什么以及它们是如何创建的。

    话虽如此，你似乎确实有一个可疑的大堆，并且（可能）有大量核心。显然，吞吐量收集器的默认策略是为每个物理核心创建一个GC线程。但是，很难想象一个用例，单线程应用程序可以创建足够的垃圾来保持大量GC线程长时间忙碌......除非它已经用可填充的对象填充堆。

相关问题
最新问题