CPU飙高了，这样玩

服务CPU飙高了

项目运行时，JVM占用的CPU飙高，跟我们预期的不符，CPU在干什么呢？是什么让CPU如此忙碌呢？让我们来一次来追踪一下。

1、查找JVM进程ID：ps aux |grep java

2、根据pid查找占用CPU较高的线程：ps -mp pid -o THREAD,tid,time

  [xxxxxxxxx-22027 ~]$ ps -mp 10605 -o THREAD,tid,time
  USER     %CPU PRI SCNT WCHAN  USER SYSTEM   TID     TIME
  root      697   -    - -         -      -     - 41-23:49:14
  root      0.0  19    - futex_    -      - 10605 00:00:00
  root      0.0  19    - futex_    -      - 10606 00:05:15
  root      0.0  19    - poll_s    -      - 10607 00:00:03
  root     99.6  33    - ?         -      - 10608 5-23:54:02
  root     99.6  33    - ?         -      - 10609 6-00:01:38
  root     99.6  33    - ?         -      - 10610 6-00:00:17
  root     99.6  33    - ?         -      - 10611 6-00:01:01
  root     99.6  33    - ?         -      - 10612 5-23:58:03
  root     99.6  33    - ?         -      - 10613 5-23:54:01
  root     99.6  33    - ?         -      - 10614 5-23:54:51

3、利用tid转换成16进制的数字：printf “%x\n” tid

  [xxxxxxxxx-22027 ~]$ printf "%x\n" 10608
  2970

4、使用jstack命令，查询线程信息，从而定位到具体线程和代码：jstack pid | grep 16进制数字 -A 30

  checking pid(10605)
  ******************************************************************* ThreadId=0 **************************************************************************
  2017-11-08 11:54:54
  Full thread dump Java HotSpot(TM) 64-Bit Server VM (24.55-b03 mixed mode):

  "Attach Listener" daemon prio=10 tid=0x00007ffa98002000 nid=0x28e29 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

  "scheduler-10" prio=10 tid=0x00007ffa7000e000 nid=0x28c99 waiting on condition [0x00007ff9bbfbf000]
   java.lang.Thread.State: WAITING (parking)
  at sun.misc.Unsafe.park(Native Method)
  - parking to wait for  <0x00000000cc91c8c0> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
  at java.util.concurrent.locks.LockSupport.park(LockSupport.java:186)
  at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2043)
  at java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(ScheduledThreadPoolExecutor.java:1085)
  at java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(ScheduledThreadPoolExecutor.java:807)
  at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1068)
  at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1130)
  at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
  at java.lang.Thread.run(Thread.java:745)

可直接查出原因。
觉得以上操作麻烦的话，可直接执行此脚本。

  #!/bin/bash
  #
  # 当JVM占用CPU特别高时，查看CPU正在做什么
  # 可输入两个参数：1、pid Java进程ID，必须参数  2、打印线程ID上下文行数，可选参数，默认打印10行
  #

  pid=$1

  if test -z $pid
  then
  echo "pid can not be null!"
  exit
  else
  echo "checking pid($pid)"
  fi

  if test -z "$(jps -l | cut -d '' -f 1 | grep $pid)"
  then
  echo "process of $pid is not exists"
  exit
  fi

  lineNum=$2
  if test -z $lineNum
  then
    lineNum=10
  fi

  jstack $pid >> "$pid".bak

  ps -mp $pid -o THREAD,tid,time | sort -k2r | awk '{if ($1 !="USER" && $2 != "0.0" && $8 !="-") print $8;}' | xargs printf "%x\n" >> "$pid".tmp

  tidArray="$( cat $pid.tmp)"

  for tid in $tidArray
  do
    echo "******************************************************************* ThreadId=$tid **************************************************************************"
    cat "$pid".bak | grep $tid -A $lineNum
  done

  rm -rf $pid.bak
  rm -rf $pid.tmp

本文转自 简书【http://www.jianshu.com/p/90579ec3113f】