首页 > 代码库 > CPU高的排查

CPU高的排查

<style></style>

之前有朋友反馈说发的内容希望有个梯度,逐步加深,前面发了几篇关于jvm源码分析的文章,可能我觉得我已经把内容写得浅显易懂了,但是对于某些没怎么接触的同学来说还是比较难理解,这个我以后慢慢改进吧,今天发篇轻松点的文章,可能大家在工作过程中也会可能碰到类似的问题,或许有经验的同学看到这个题目就知道我要说什么了,也有自己的定位方法。

话不多说了,先来看代码吧

public class Test{

        public static void main(String args[]){

                for(int i=0;i<10;i++){

                        new Thread(){

                                public void run(){

                                        try{

                                                Thread.sleep(100000);

                                        }catch(Exception e){}

                                }

                        }.start();

                }

                Thread t=new Thread(){

                        public void run(){

                                int i=0;

                                while(true){

                                        i=(i++)/100;

                                }

                        }

                };

                t.setName("Busiest Thread");

                t.start();

        }

}

这个例子里新创建了11个线程,其中10个线程没干什么事,主要是sleep,另外有一个线程在循环里一直跑着,可以想象这个线程是这个进程里最耗cpu的线程了,那怎么把这个线程给抓出来呢?

首先我们可以通过top -Hp <pid>来看这个进程里所有线程的cpu消耗情况,得到类似下面的数据

$ top -Hp 18207

top - 19:11:43 up 573 days,  2:43,  2 users,  load average: 3.03, 3.03, 3.02

Tasks:  44 total,   1 running,  43 sleeping,   0 stopped,   0 zombie

Cpu(s): 18.8%us,  0.0%sy,  0.0%ni, 81.1%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st

Mem:  99191752k total, 98683576k used,   508176k free,   128248k buffers

Swap:  1999864k total,   191064k used,  1808800k free, 17413760k cached

 

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND

18250 admin     20   0 26.1g  28m  10m R 99.9  0.0   0:19.50 java Test

18207 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18208 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.09 java Test

18209 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18210 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18211 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18212 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18213 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18214 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18215 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18216 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18217 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18218 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18219 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18220 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18221 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18222 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18223 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18224 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18225 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18226 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

18227 admin     20   0 26.1g  28m  10m S  0.0  0.0   0:00.00 java Test

拿到这个结果之后,我们可以看到cpu最高的线程是pid为18250的线程,占了99.8%:

 PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND

18250 admin     20   0 26.1g  28m  10m R 99.9  0.0   0:19.50 java Test

接着我们可以通过jstack <pid>的输出来看各个线程栈:

$ jstack 18207

2016-03-30 19:12:23

Full thread dump OpenJDK 64-Bit Server VM (25.66-b60 mixed mode):

 

"Attach Listener" #30 daemon prio=9 os_prio=0 tid=0x00007fb90be13000 nid=0x47d7 waiting on condition [0x0000000000000000]

   java.lang.Thread.State: RUNNABLE

 

"DestroyJavaVM" #29 prio=5 os_prio=0 tid=0x00007fb96245b800 nid=0x4720 waiting on condition [0x0000000000000000]

   java.lang.Thread.State: RUNNABLE

 

"Busiest Thread" #28 prio=5 os_prio=0 tid=0x00007fb91498d000 nid=0x474a runnable [0x00007fb9065fe000]

   java.lang.Thread.State: RUNNABLE

    at Test$2.run(Test.java:18)

 

"Thread-9" #27 prio=5 os_prio=0 tid=0x00007fb91498c800 nid=0x4749 waiting on condition [0x00007fb906bfe000]

   java.lang.Thread.State: TIMED_WAITING (sleeping)

    at java.lang.Thread.sleep(Native Method)

    at Test$1.run(Test.java:9)

 

"Thread-8" #26 prio=5 os_prio=0 tid=0x00007fb91498b800 nid=0x4748 waiting on condition [0x00007fb906ffe000]

   java.lang.Thread.State: TIMED_WAITING (sleeping)

    at java.lang.Thread.sleep(Native Method)

    at Test$1.run(Test.java:9)

 

"Thread-7" #25 prio=5 os_prio=0 tid=0x00007fb91498b000 nid=0x4747 waiting on condition [0x00007fb9073fe000]

   java.lang.Thread.State: TIMED_WAITING (sleeping)

    at java.lang.Thread.sleep(Native Method)

    at Test$1.run(Test.java:9)

 

"Thread-6" #24 prio=5 os_prio=0 tid=0x00007fb91498a000 nid=0x4746 waiting on condition [0x00007fb9077fe000]

   java.lang.Thread.State: TIMED_WAITING (sleeping)

    at java.lang.Thread.sleep(Native Method)

    at Test$1.run(Test.java:9)

...

上面的线程栈我们注意到nid的值其实就是线程ID,它是十六进制的,我们将消耗cpu最高的线程18250,转成十六进制0X47A,然后从上面的线程栈里找到nid=0X47A的线程,其栈为:

"Busiest Thread" #28 prio=5 os_prio=0 tid=0x00007fb91498d000 nid=0x474a runnable [0x00007fb9065fe000]

   java.lang.Thread.State: RUNNABLE

    at Test$2.run(Test.java:18)

即将最耗cpu的线程找出来了,是Businest Thread

CPU高的排查