我们知道,java程序一般都很耗资源,动辄都是好几G的内存,甚至几十G。对外,java进程是一个整体,而且java是共享内存模型,所以很难分析单个线程,但是我们还是可以分析java线程消耗cpu的情况。
1 首先得到要分析的java程序进程id,可以通过jps命令:
[make@master ~]$ jps -l
10242 com.intellij.idea.Main
10597 org.jetbrains.idea.maven.server.RemoteMavenServer
15127 sun.tools.jps.Jps
11613 org.jetbrains.jps.cmdline.Launcher
因为intellij不仅耗内存也耗cpu,不如就来分析它吧。
2 用top命令分析进程号为10242内的线程:
top - 21:14:03 up 3:26, 2 users, load average: 0.02, 0.17, 0.26
Threads: 50 total, 0 running, 50 sleeping, 0 stopped, 0 zombie
%Cpu(s): 1.1 us, 0.5 sy, 0.0 ni, 97.9 id, 0.5 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 3792584 total, 226952 free, 2734716 used, 830916 buff/cache
KiB Swap: 4063228 total, 3474208 free, 589020 used. 308556 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
10599 make 20 0 4981576 977.4m 12568 S 0.7 26.4 0:42.29 java
10600 make 20 0 4981576 977.4m 12568 S 0.7 26.4 0:42.24 java
10258 make 20 0 4981576 977.4m 12568 S 0.3 26.4 0:03.62 java
10242 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10243 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:01.33 java
10244 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:01.47 java
10245 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:01.48 java
10246 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:01.51 java
10247 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:01.46 java
10248 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:04.06 java
10249 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:02.62 java
10250 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.05 java
10251 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.04 java
10252 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10253 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10254 make 20 0 4981576 977.4m 12568 S 0.0 26.4 1:00.67 java
10255 make 20 0 4981576 977.4m 12568 S 0.0 26.4 1:06.17 java
10256 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:33.85 java
10257 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10259 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.01 java
10260 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:02.39 java
10262 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 gdbus
10263 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 gmain
10269 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10270 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.31 java
10272 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.06 java
10273 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:04.32 java
10274 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10277 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:56.89 java
10281 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10282 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.04 java
10283 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10288 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10290 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.29 java
10582 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:01.15 java
10598 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10620 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.03 java
10621 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10622 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10624 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.00 java
10625 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.03 java
10629 make 20 0 4981576 977.4m 12568 S 0.0 26.4 0:00.23 java
3 下面来分析进程号为10599 的线程情况(java 线程对外映射称操作系统的进程)
首先得到该java进程的线程堆栈信息:
jstack -l 10242 > stack.log
然后得到线程10599对应的16进制:
[make@master ~]$ python -c "print hex(10599)"
0x2967
然后在stack.log 文件里找到对应线程:
"BaseDataReader: output stream of java" #57 prio=4 os_prio=0 tid=0x00007f88f0914000 nid=0x2967 in Object.wait() [0x00007f88d8df4000]
java.lang.Thread.State: TIMED_WAITING (on object monitor)
at java.lang.Object.wait(Native Method)
at com.intellij.util.io.BaseDataReader.doRun(BaseDataReader.java:173)
- locked <0x00000000e5227758> (a java.lang.Object)
at com.intellij.util.io.BaseDataReader$1.run(BaseDataReader.java:61)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)
Locked ownable synchronizers:
- <0x00000000e20ee510> (a java.util.concurrent.ThreadPoolExecutor$Worker)
下面就可以针对该线程做进一步分析,为啥这么耗cpu