一、top命令定位cpu占用率高的进程号（pid），找到pid后，top -p pid -H,记录出cpu占用率高的线程号（tid），用printf "%x\n" tid将其转换为16进制
二、使用jstack打印进程信息，为准确定位可以多来几次，jstack pid > pid_1.log
三、打开日志文件，找到十六进制的线程信息，可定位到具体类的某一行。
示例
1、
1 Tasks: 546 total, 1 running, 544 sleeping, 0 stopped, 1 zombie 2 Cpu(s): 15.3%us, 0.1%sy, 0.0%ni, 84.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st 3 Mem: 24633804k total, 23710152k used, 923652k free, 784472k buffers 4 Swap: 32764556k total, 127120k used, 32637436k free, 7000668k cached 5 6 PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7 25930 weblogic 18 0 5311m 4.5g 29m S 488.1 19.3 35663:17 java 8 16830 weblogic 15 0 13168 1472 824 R 0.7 0.0 0:00.04 top 9 13834 weblogic 16 0 13028 1424 812 S 0.3 0.0 5900:50 top 10 17263 weblogic 23 0 2410m 1.5g 28m S 0.3 6.2 390:24.31 java 11 17438 weblogic 25 0 3429m 1.7g 25m S 0.3 7.2 217:35.39 java 12 26316 weblogic 23 0 4695m 2.5g 29m S 0.3 10.8 31:47.38 java 13 1 root 15 0 10372 600 568 S 0.0 0.0 2:12.76 init 14 2 root RT -5 0 0 0 S 0.0 0.0 22:18.89 migration/0 15 3 root 34 19 0 0 0 S 0.0 0.0 2:38.19 ksoftirqd/0 16 4 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/0 17 5 root RT -5 0 0 0 S 0.0 0.0 19:16.83 migration/1 18 6 root 34 19 0 0 0 S 0.0 0.0 6:37.27 ksoftirqd/1 19 7 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/1 20 8 root RT -5 0 0 0 S 0.0 0.0 11:11.42 migration/2 21 9 root 34 19 0 0 0 S 0.0 0.0 6:59.29 ksoftirqd/2 22 10 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/2 23 11 root RT -5 0 0 0 S 0.0 0.0 5:36.90 migration/3
进程号 为25930的java进程cpu使用率很高，top -p 25930 -H
1 Tasks: 220 total, 1 running, 219 sleeping, 0 stopped, 0 zombie 2 Cpu(s): 8.7%us, 0.1%sy, 0.0%ni, 91.2%id, 0.0%wa, 0.0%hi, 0.1%si, 0.0%st 3 Mem: 24633804k total, 23712492k used, 921312k free, 784472k buffers 4 Swap: 32764556k total, 127120k used, 32637436k free, 7000868k cached 5 6 PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 7 25960 weblogic 15 0 5311m 4.5g 29m S 51.9 19.3 5615:15 java 8 25941 weblogic 16 0 5311m 4.5g 29m S 15.0 19.3 798:46.36 java 9 25939 weblogic 15 0 5311m 4.5g 29m S 11.3 19.3 798:26.28 java 10 25951 weblogic 15 0 5311m 4.5g 29m S 11.3 19.3 799:05.00 java 11 25936 weblogic 15 0 5311m 4.5g 29m S 11.0 19.3 799:14.41 java 12 25937 weblogic 15 0 5311m 4.5g 29m S 11.0 19.3 798:37.65 java 13 25943 weblogic 15 0 5311m 4.5g 29m S 11.0 19.3 799:30.19 java 14 25944 weblogic 15 0 5311m 4.5g 29m S 11.0 19.3 799:58.25 java 15 25945 weblogic 15 0 5311m 4.5g 29m S 11.0 19.3 798:56.29 java
tid为25960的线程cpu使用率很高。
1 [weblogic@shtpeb2b2eapp01 adminserver]$ printf "%x\n" 25960 2 6568
2、使用jstack打印进程信息并导入日志文件
jstack 25930 >25930_1.log
jstack 25930 >25930_2.log
jstack 25930 >25930_3.log
3、查看相关日志信息
1 [weblogic@shtpeb2b2eapp01 bin]$ grep -A 10 0x6568 25930_1.log 2 "Concurrent Mark-Sweep GC Thread" prio=10 tid=0x0000000048b59000 nid=0x6568 runnable 3 "Gang worker#0 (Parallel CMS Threads)" prio=10 tid=0x0000000048b4f000 nid=0x6563 runnable 4 5 "Gang worker#1 (Parallel CMS Threads)" prio=10 tid=0x0000000048b51000 nid=0x6564 runnable 6 7 "Gang worker#2 (Parallel CMS Threads)" prio=10 tid=0x0000000048b53000 nid=0x6565 runnable 8 9 "Gang worker#3 (Parallel CMS Threads)" prio=10 tid=0x0000000048b54800 nid=0x6566 runnable 10 11 "Gang worker#4 (Parallel CMS Threads)" prio=10 tid=0x0000000048b56800 nid=0x6567 runnable 12 13 [weblogic@shtpeb2b2eapp01 bin]$ grep -A 10 0x6568 25930_2.log 14 "Concurrent Mark-Sweep GC Thread" prio=10 tid=0x0000000048b59000 nid=0x6568 runnable 15 "Gang worker#0 (Parallel CMS Threads)" prio=10 tid=0x0000000048b4f000 nid=0x6563 runnable 16 17 "Gang worker#1 (Parallel CMS Threads)" prio=10 tid=0x0000000048b51000 nid=0x6564 runnable 18 19 "Gang worker#2 (Parallel CMS Threads)" prio=10 tid=0x0000000048b53000 nid=0x6565 runnable 20 21 "Gang worker#3 (Parallel CMS Threads)" prio=10 tid=0x0000000048b54800 nid=0x6566 runnable 22 23 "Gang worker#4 (Parallel CMS Threads)" prio=10 tid=0x0000000048b56800 nid=0x6567 runnable
从输出信息来看，此线程是JVM的GC线程，此时可以基本确定是内存不足或内存泄露导致gc线程持续运行，导致CPU占用过高。
所以接下来我们要找的内存方面的问题。
1 [weblogic@shtpeb2b2eapp01 bin]$ jstat -gcutil 25930 2000 10 2 S0 S1 E O P YGC YGCT FGC FGCT GCT 3 0.00 0.00 94.78 100.00 87.22 1035 92.722 251436 158050.083 158142.805 4 0.00 0.00 94.78 100.00 87.22 1035 92.722 251437 158050.083 158142.805 5 0.00 0.00 94.78 100.00 87.22 1035 92.722 251438 158051.345 158144.067 6 0.00 0.00 94.79 100.00 87.22 1035 92.722 251438 158052.279 158145.001 7 0.00 0.00 94.79 100.00 87.22 1035 92.722 251439 158052.279 158145.001 8 0.00 0.00 94.79 100.00 87.22 1035 92.722 251440 158053.553 158146.275 9 0.00 0.00 94.79 100.00 87.22 1035 92.722 251440 158054.475 158147.197 10 0.00 0.00 94.79 100.00 87.22 1035 92.722 251441 158054.475 158147.197 11 0.00 0.00 94.79 100.00 87.22 1035 92.722 251441 158055.753 158148.475
从输出信息可以看出，Eden区内存占用95%，Old区内存占用100%，根据这些信息，基本可以确定是程序代码上出现了问题，可能存在不合理创建对象的地方。
使用文档编辑器打开日志
发现大量的线程等待以及被锁住的资源，然后丢给相关开发同事
来源：https://www.cnblogs.com/breeze-24/p/9098978.html