java进程CPU高分析

JVM导致系统CPU高的常见场景：

内存不足，JVM gc频繁，一般会伴随OOM
JVM某个线程死循环或者递归调用

定位和解决
1.内存不足，gc频繁可参考我的这遍文章解决。https://blog.csdn.net/moranzi1/article/details/88670204
2.JVM某个线程死循环或者递归调用。这种情况关键是找到导致CPU高的线程。然后根据具体线程具体分析为什么该线程会导致CPU高。需要线程的步骤如下。

top——命令查看cpu高的进程

[root@iZwz9gs2zseivevv1k5vnkZ syp]# top

top - :: up  days, :,  users, load average: 4.59, 4.16, 3.26

Tasks:  total,  running,  sleeping,  stopped,  zombie

%Cpu(s): 84.1 us, 4.5 sy, 0.0 ni, 10.8 id, 0.1 wa, 0.0 hi, 0.6 si, 0.0 st

KiB Mem :  total,  free,  used,  buff/cache

KiB Swap:  total,  free,  used.  avail Mem

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

 root      S 293.0 6.0 :23.05 java

 root      S 15.9 3.2 : java

 root      S 12.0 5.0 : java

 root      S 2.3 5.3 :16.27 java

 root      S 2.0 4.8 :26.20 java

 root  -    S 2.0 0.1 :57.96 AliYunDun

ps aux | grep PID——命令查看具体进程信息

ps -mp pid -o THREAD,tid,time——命令查看进程线程信息，找到导致CPU高的线程TID

[root@iZwz9gs2zseivevv1k5vnkZ syp]# ps -mp  -o THREAD,tid,time

USER %CPU PRI SCNT WCHAN USER SYSTEM TID TIME

root 22.7 - - - - - - ::

root 0.0  - futex_ - -  ::

root 0.1  - futex_ - -  ::

root 4.6  - - - -  ::

root 4.6  - - - -  ::

root 4.6  - - - -  ::

root 4.6  - - - -  ::

root 1.2  - futex_ - -  ::

printf "%x\n" tid——命令转换线程为16进制格式

[root@iZwz9gs2zseivevv1k5vnkZ syp]# printf “%x\n”

“2f4dn”

[root@iZwz9gs2zseivevv1k5vnkZ syp]# printf “%x\n”

“2f4en”

[root@iZwz9gs2zseivevv1k5vnkZ syp]# printf “%x\n”

“2f4fn”

[root@iZwz9gs2zseivevv1k5vnkZ syp]# printf “%x\n”

“2f50n”

jstack pid |grep tid -A 30——命令查看线程信息，定位到具体线程

[root@iZwz9gs2zseivevv1k5vnkZ syp]# jstack  | grep 2f4d -A

"GC task thread#0 (ParallelGC)" os_prio= tid=0x00007f223405e000 nid=0x2f4d runnable

"GC task thread#1 (ParallelGC)" os_prio= tid=0x00007f2234060000 nid=0x2f4e runnable

"GC task thread#2 (ParallelGC)" os_prio= tid=0x00007f2234061800 nid=0x2f4f runnable

"GC task thread#3 (ParallelGC)" os_prio= tid=0x00007f2234063800 nid=0x2f50 runnable

"VM Periodic Task Thread" os_prio= tid=0x00007f22341ed800 nid=0x2f59 waiting on condition

JNI global references: 

[root@iZwz9gs2zseivevv1k5vnkZ syp]# jstack  | grep 2f4e -A

"GC task thread#1 (ParallelGC)" os_prio= tid=0x00007f2234060000 nid=0x2f4e runnable

"GC task thread#2 (ParallelGC)" os_prio= tid=0x00007f2234061800 nid=0x2f4f runnable

"GC task thread#3 (ParallelGC)" os_prio= tid=0x00007f2234063800 nid=0x2f50 runnable

"VM Periodic Task Thread" os_prio= tid=0x00007f22341ed800 nid=0x2f59 waiting on condition

JNI global references: 

"GC task thread#2 (ParallelGC)" os_prio= tid=0x00007f2234061800 nid=0x2f4f runnable

"GC task thread#3 (ParallelGC)" os_prio= tid=0x00007f2234063800 nid=0x2f50 runnable

"VM Periodic Task Thread" os_prio= tid=0x00007f22341ed800 nid=0x2f59 waiting on condition

JNI global references: 

[root@iZwz9gs2zseivevv1k5vnkZ syp]# jstack  | grep 2f50 -A

"GC task thread#3 (ParallelGC)" os_prio= tid=0x00007f2234063800 nid=0x2f50 runnable

"VM Periodic Task Thread" os_prio= tid=0x00007f22341ed800 nid=0x2f59 waiting on condition

JNI global references:

参考：

https://blog.csdn.net/moranzi1/article/details/89341480

巴特西

java进程CPU高分析

最新文章

热门文章