Java 线上cpu占用过高分析

转载于：https://blog.csdn.net/ch999999999999999999/article/details/113151519
感谢作者：ch999999999999999999

Java 线上cpu占用过高分析

1、top 命令查看占用过高线程

 9401 root      20   0 3163948  43592  11932 S   1.3  1.2   0:15.81 java                                                                    
 2609 polkitd   20   0 1333984 209256  11044 S   0.3  5.6   1:27.59 mysqld                                 
13301 root      10 -10  145952  20784   5980 S   0.3  0.6  84:17.34 AliYunDun                              
23125 polkitd   20   0 1780292 514228  11512 S   0.3 13.8  19:45.03 mysqld  

  
  

  
  
1
2
3
4

可以看出pid 为9401的线程占用过高

2、查看是哪个线程占用过高， 哪个线程耗费了多长时间

ps -mp 9401 -o THREAD,tid,time参数
// 参数解释
-m 显示所有的线程
-p pid 进程使用cpu的时间
-o 该参数是用户自定义格式

USER     %CPU PRI SCNT WCHAN  USER SYSTEM   TID     TIME
root      1.8   -    - -         -      -     - 00:00:20
root      0.0  19    - futex_    -      -  9401 00:00:00
root      1.7  19    - n_tty_    -      -  9402 00:00:19
root      0.0  19    - futex_    -      -  9403 00:00:00
root      0.0  19    - futex_    -      -  9404 00:00:00
root      0.0  19    - futex_    -      -  9405 00:00:00

  
  

  
  
1
2
3
4
5
6
7

可以看出9402线程 占用过高

3、将线程id转换为16进制格式（英文字母小写格式） -> printf “%x\n” 有问题的线程id

printf "%x\n" 9402

  
  

  
  
1

结果： 24ba

4、jstack 进程ID|grep tid（16进制线程小写英文） -A60

jstack 9402 | grep 24ba -A60

  
  

  
  
1

结果：

"main" #1 prio=5 os_prio=0 tid=0x00007fc380009800 nid=0x2c84 runnable [0x00007fc387caa000]
   java.lang.Thread.State: RUNNABLE
	at java.io.FileOutputStream.writeBytes(Native Method)
	at java.io.FileOutputStream.write(FileOutputStream.java:326)
	at java.io.BufferedOutputStream.flushBuffer(BufferedOutputStream.java:82)
	at java.io.BufferedOutputStream.flush(BufferedOutputStream.java:140)
	- locked <0x00000000c701b490> (a java.io.BufferedOutputStream)
	at java.io.PrintStream.write(PrintStream.java:482)
	- locked <0x00000000c7004fb0> (a java.io.PrintStream)
	at sun.nio.cs.StreamEncoder.writeBytes(StreamEncoder.java:221)
	at sun.nio.cs.StreamEncoder.implFlushBuffer(StreamEncoder.java:291)
	at sun.nio.cs.StreamEncoder.flushBuffer(StreamEncoder.java:104)
	- locked <0x00000000c7004f70> (a java.io.OutputStreamWriter)
	at java.io.OutputStreamWriter.flushBuffer(OutputStreamWriter.java:185)
	at java.io.PrintStream.newLine(PrintStream.java:546)
	- eliminated <0x00000000c7004fb0> (a java.io.PrintStream)
	at java.io.PrintStream.println(PrintStream.java:737)
	- locked <0x00000000c7004fb0> (a java.io.PrintStream)
	at TestJavaLinux.main(TestJavaLinux.java:11)

“VM Thread” os_prio=0 tid=0x00007fc380073000 nid=0x2c87 runnable

“GC task thread#0 (ParallelGC)” os_prio=0 tid=0x00007fc38001e800 nid=0x2c85 runnable

“GC task thread#1 (ParallelGC)” os_prio=0 tid=0x00007fc380020800 nid=0x2c86 runnable

“VM Periodic Task Thread” os_prio=0 tid=0x00007fc3800bf000 nid=0x2c8e waiting on condition

JNI global references: 5

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29

由此可见是 at TestJavaLinux.main(TestJavaLinux.java:11) 我们的11行代码产出的原因

我们在代码里面写了死循环

5、注意：

注意：这里要使用程序启动的用户执行此命令

使用root会提示：

Unable to open socket file: target process not responding or HotSpot VM not loaded

The -F option can be used when the target process is not responding