Java Thread Dump_kill -3 $pid和jstack-CSDN博客

本文链接：https://blog.csdn.net/west_609/article/details/79394610

Thread dump是对于程序的诊断是非常有用的，它可以帮助我们分析诸如程序没有响应，响应慢，CPU抖动，死锁，内存问题以及其它的系统问题。今天主要想分享如果打印Thread Dump

假定我们现在怀疑某个java进程出现了问题，比如没有响应，首先可以尝试拿到这个java应用对应的process Id (ps or jps)。然后可以看一下这个应用启动了那一些线程，这个步骤可以top命令：

# top -H -p 3627
top - 07:15:00 up  5:46,  2 users,  load average: 0.22, 0.33, 0.47
Tasks:  12 total,   0 running,  12 sleeping,   0 stopped,   0 zombie
Cpu(s): 23.5%us,  9.8%sy,  2.2%ni, 62.3%id,  1.2%wa,  0.0%hi,  0.9%si,  0.0%st
Mem:   8061348k total,  7981468k used,    79880k free,    73332k buffers
Swap:  8208380k total,        0k used,  8208380k free,  1370492k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 3627 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.00 java
 3628 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.17 java
 3629 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.00 java
 3630 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.00 java
 3631 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.00 java
 3632 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.00 java
 3633 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.00 java
 3634 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.00 java
 3635 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.01 java
 3636 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.01 java
 3637 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.00 java
 3638 cms       20   0 3042m  23m  10m S  0.0  0.3   0:00.01 java

CPU的summary各个field的意义：

us = user mode (application运行时间）
sy = system mode （kernel运行时间）
ni = low priority user mode (nice)
id = idle task
wa = I/O waiting
hi = servicing IRQs
si = servicing soft IRQs
st = steal (time given to other DomU instances)

其中第一列PID就是Thread 的系统Id, 对应于后面thread dump中的nid(native thread id)。这里需要特别注意：这个id是十进制的，而thread dump中的nid是十六进制的，所以需要做一个转换。

从top命令output的summary来看，它有12个线程，而且12个线程处于sleeping状态。

接下来就是如果得到这个进程3627的thread dump。这里主要介绍jstack 以及kill -3，当然还有其它的方式，如jvisualVM，这里不详细介绍了。

jstack

jstack是一个非常有用的工具，用于抓取thread dump，这个工具通常在JDK_HOME/bin目录下。用法也相对简单:

jstack -l <pid>

如果没有响应，可以尝试使用-F选项， 3627进程thread dump：

2018-02-27 07:27:18
Full thread dump Java HotSpot(TM) 64-Bit Server VM (24.141-b31 mixed mode):

"Attach Listener" daemon prio=10 tid=0x00007faba0001000 nid=0x3b52 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

   Locked ownable synchronizers:
        - None

"Service Thread" daemon prio=10 tid=0x00007fabc411a800 nid=0xe35 runnable [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

   Locked ownable synchronizers:
        - None

"C2 CompilerThread1" daemon prio=10 tid=0x00007fabc4118000 nid=0xe34 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

   Locked ownable synchronizers:
        - None

"C2 CompilerThread0" daemon prio=10 tid=0x00007fabc4115000 nid=0xe33 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

   Locked ownable synchronizers:
        - None

"Signal Dispatcher" daemon prio=10 tid=0x00007fabc4113000 nid=0xe32 runnable [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

   Locked ownable synchronizers:
        - None

"Finalizer" daemon prio=10 tid=0x00007fabc40f1800 nid=0xe31 in Object.wait() [0x00007fabc0ad0000]
   java.lang.Thread.State: WAITING (on object monitor)
        at java.lang.Object.wait(Native Method)
        - waiting on <0x00000007d6f04860> (a java.lang.ref.ReferenceQueue$Lock)
        at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:135)
        - locked <0x00000007d6f04860> (a java.lang.ref.ReferenceQueue$Lock)
        at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:151)
        at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:209)

   Locked ownable synchronizers:
        - None

"Reference Handler" daemon prio=10 tid=0x00007fabc40ef800 nid=0xe30 in Object.wait() [0x00007fabafffe000]
   java.lang.Thread.State: WAITING (on object monitor)
        at java.lang.Object.wait(Native Method)
        - waiting on <0x00000007d6f04478> (a java.lang.ref.Reference$Lock)
        at java.lang.Object.wait(Object.java:503)
        at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:133)
        - locked <0x00000007d6f04478> (a java.lang.ref.Reference$Lock)

   Locked ownable synchronizers:
        - None

"main" prio=10 tid=0x00007fabc4008800 nid=0xe2c runnable [0x00007fabc9ed6000]
   java.lang.Thread.State: RUNNABLE
        at java.io.FileInputStream.readBytes(Native Method)
        at java.io.FileInputStream.read(FileInputStream.java:272)
        at java.io.BufferedInputStream.read1(BufferedInputStream.java:273)
        at java.io.BufferedInputStream.read(BufferedInputStream.java:334)
        - locked <0x00000007d6f19af0> (a java.io.BufferedInputStream)
        at sun.nio.cs.StreamDecoder.readBytes(StreamDecoder.java:283)
        at sun.nio.cs.StreamDecoder.implRead(StreamDecoder.java:325)
        at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:177)
        - locked <0x00000007d721e808> (a java.io.InputStreamReader)
        at java.io.InputStreamReader.read(InputStreamReader.java:184)
        at java.io.BufferedReader.fill(BufferedReader.java:154)
        at java.io.BufferedReader.readLine(BufferedReader.java:317)
        - locked <0x00000007d721e808> (a java.io.InputStreamReader)
        at java.io.BufferedReader.readLine(BufferedReader.java:382)
        at com.tools.Utilities.readInput(LogUtilities.java:175)
        at com.tools.Utilities.main(LogUtilities.java:184)

   Locked ownable synchronizers:
        - None

"VM Thread" prio=10 tid=0x00007fabc40eb000 nid=0xe2f runnable

"GC task thread#0 (ParallelGC)" prio=10 tid=0x00007fabc401e000 nid=0xe2d runnable

"GC task thread#1 (ParallelGC)" prio=10 tid=0x00007fabc4020000 nid=0xe2e runnable

"VM Periodic Task Thread" prio=10 tid=0x00007fabc4125800 nid=0xe36 waiting on condition

JNI global references: 129

可以看到上面有12个线程，跟top命令看到是一样的。每个线程那是以下面的格式开始：

"main" prio=10 tid=0x00007fabc4008800 nid=0xe2c runnable [0x00007fabc9ed6000]
   java.lang.Thread.State: RUNNABLE

其中：

main: thread name，如果没有指定线程名字，由系统自动生成，比如”Thread-1”
tid： jvm的线程id，由jvm自已维护其唯一性
nid： thread的系统id，这个通常跟操作系统相关，所以称为本地线程Id, 在Linux下可以在/proc/processId/task/查看这个所以的线程Id
runnable: 线程的状态

kill -3

如果生产环境没有jdk，只有jre，那就没有办法使用jstack。作为备用，我们可以使用kill -3或者kill -QUIT来抓取thread dump

kill -3 <pid>

这里需要注意的是，thread dump可能不会在你当前的console上打印

如果你的应用是布置在tomcat，那么你可以从catalina.out查看
如果你是用nohup启动的后台进程，那么应该在nohup.out查看

kill -3的thread dump跟jstack很类似，有一点不一样，kill -3会在最后打印堆相关的信息：

Heap
 PSYoungGen      total 36864K, used 3193K [0x00000007d6f00000, 0x00000007d9800000, 0x0000000800000000)
  eden space 31744K, 10% used [0x00000007d6f00000,0x00000007d721e670,0x00000007d8e00000)
  from space 5120K, 0% used [0x00000007d9300000,0x00000007d9300000,0x00000007d9800000)
  to   space 5120K, 0% used [0x00000007d8e00000,0x00000007d8e00000,0x00000007d9300000)
 ParOldGen       total 83968K, used 0K [0x0000000784e00000, 0x000000078a000000, 0x00000007d6f00000)
  object space 83968K, 0% used [0x0000000784e00000,0x0000000784e00000,0x000000078a000000)
 PSPermGen       total 21504K, used 4553K [0x000000077fc00000, 0x0000000781100000, 0x0000000784e00000)
  object space 21504K, 21% used [0x000000077fc00000,0x0000000780072558,0x0000000781100000)