转载

Java性能 -- 命令行工具

$ free -m
             total       used       free     shared    buffers     cached
Mem:         15948      15261        687        304         37       6343
-/+ buffers/cache:       8880       7068
Swap:            0          0          0

Mem 是从 操作系统 的角度来看的
- 总共有15948M物理内存，其中15261M被使用了，还有687可用， 15948 = 15261 + 687
- 有若干线程共享了304M物理内存，已经被弃用（值总为0）
- buffer / cached ：为了 提高IO性能 ，由OS管理
  - A buffer is something that has yet to be “written” to disk.
  - A cache is something that has been “read” from the disk and stored for later use.
  - buffer和cached占用的内存可以 被快速回收

-/+ buffers/cache 是从 应用程序 的角度来看的

8880 ≈ (15261 - 37 - 6343 = 8881)
7068 ≈ (687 + 37 + 6343 = 7067)

vmstat

$ vmstat -S M 1 5
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu-----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 2  0      0    295      1   5684    0    0     8    23    1    2  5  1 93  0  0
 0  0      0    295      1   5684    0    0     0     0 11326 25112  6  5 90  0  0
 0  0      0    295      1   5685    0    0     0     0 10855 24065  2  2 96  0  0
 0  0      0    294      1   5686    0    0     0     4 11756 25472  4  2 94  0  0
 0  0      0    294      1   5686    0    0     0     0 10208 22726  4  4 91  0  0

$ free -m
             total       used       free     shared    buffers     cached
Mem:         15948      15655        293       5362          1       5687
-/+ buffers/cache:       9966       5981
Swap:            0          0          0

vmstat是一款指定采样周期和次数的 功能性监控工具 ，可以用来监控进程上下文切换的情况
procs
- r: The number of runnable processes ( running or waiting for run time)
- b: The number of processes in uninterruptible sleep
memory
- swpd: the amount of virtual memory used
- free: the amount of idle memory
- buff: the amount of memory used as buffers
- cache: the amount of memory used as cache
swap
- si: Amount of memory swapped in from disk (/s)
- so: Amount of memory swapped to disk (/s)
io
- bi: Blocks received from a block device (blocks/s)
- bo: Blocks sent to a block device (blocks/s)
system
- in: The number of interrupts per second, including the clock
- cs: The number of context switches per second
cpu
- us: Time spent running non-kernel code
- sy: Time spent running kernel code
- id: Time spent idle, this includes IO-wait time
- wa: Time spent waiting for IO, included in idle
- st: Time stolen from a virtual machine

pidstat

pidstat可以监测到 具体线程的上下文切换 ，pidstat是 sysstat 中的一个组件
常用参数
- -u : Report CPU utilization
- -r : Report page faults and memory utilization
- -d : Report I/O statistics
- -w : Report task switching activity
- -t : Also display statistics for threads associated with selected tasks
- -p { pid [,...] | SELF | ALL } : Select tasks (processes) for which statistics are to be reported

CPU使用率

$ pidstat -u -p 4256 1 3
Linux 2.6.32-696.el6.x86_64 (XXXX) 	2019年XX月XX日 	_x86_64_	(32 CPU)

22时59分32秒       PID    %usr %system  %guest    %CPU   CPU  Command
22时59分33秒      4256   71.00   13.00    0.00   84.00     1  java
22时59分34秒      4256   60.00   10.00    0.00   70.00     1  java
22时59分35秒      4256   76.24   12.87    0.00   89.11     1  java
平均时间:      4256   69.10   11.96    0.00   81.06     -  java

$ pidstat -u -I -p 4256 1 3
Linux 2.6.32-696.el6.x86_64 (XXXX) 	2019年XX月XX日 	_x86_64_	(32 CPU)

23时06分18秒       PID    %usr %system  %guest    %CPU   CPU  Command
23时06分19秒      4256   69.31   12.87    0.00    2.63     1  java
23时06分20秒      4256   69.00   12.00    0.00    2.57     1  java
23时06分21秒      4256   62.00   12.00    0.00    2.34     1  java
平均时间:      4256   66.78   12.29    0.00    2.51     -  java

%usr
- Percentage of CPU used by the task while executing at the user level ( application )
- Note that this field does NOT include time spent running a virtual processor
%system
- Percentage of CPU used by the task while executing at the system level ( kernel )
%guest
- Percentage of CPU spent by the task in virtual machine (running a virtual processor )
%CPU
- Total percentage of CPU time used by the task
- In an SMP environment, the task’s CPU usage will be divided by the total number of CPU’s if option -I has been entered on the command line
CPU
- Processor number to which the task is attached

页错误 + 内存使用率

$ pidstat -r -p 4256 1 3
23时11分26秒       PID  minflt/s  majflt/s     VSZ    RSS   %MEM  Command
23时11分27秒      4256   2132.00      0.00 22427892 5172756   7.84  java
23时11分28秒      4256   1623.00      0.00 22427892 5172728   7.84  java
23时11分29秒      4256   1942.00      0.00 22427892 5172728   7.84  java
平均时间:      4256   1899.00      0.00 22427892 5172737   7.84  java

minflt/s
- Total number of minor faults the task has made per second, those which have not required loading a memory page from disk
majflt/s
- Total number of major faults the task has made per second, those which have required loading a memory page from disk
VSZ
- Virtual Size: The virtual memory usage of entire task in kilobytes
RSS
- Resident Set Size: The non-swapped physical memory used by the task in kilobytes

IO统计

$ pidstat -d -p 4256 1 3
23时24分59秒       PID   kB_rd/s   kB_wr/s kB_ccwr/s  Command
23时25分00秒      4256      0.00    408.00      0.00  java
23时25分01秒      4256      0.00    356.44      0.00  java
23时25分02秒      4256      0.00    375.76      0.00  java
平均时间:      4256      0.00    380.00      0.00  java

kB_rd/s
- Number of kilobytes the task has caused to be read from disk per second
kB_wr/s
- Number of kilobytes the task has caused, or shall cause to be written to disk per second
kB_ccwr/s
- Number of kilobytes whose writing to disk has been cancelled by the task
- This may occur when the task truncates some dirty pagecache
  - In this case, some IO which another task has been accounted for will not be happening

进程上下文切换

$ pidstat -w -p 4256 1 3
22时54分33秒       PID   cswch/s nvcswch/s  Command
22时54分34秒      4256      0.00      0.00  java
22时54分35秒      4256      0.00      0.00  java
22时54分36秒      4256      0.00      0.00  java
平均时间:      4256      0.00      0.00  java

cswch/s
- Total number of voluntary context switches the task made per second
- A voluntary context switch occurs when a task blocks because it requires a resource that is unavailable
nvcswch/s
- Total number of non voluntary context switches the task made per second
- A involuntary context switch takes place when a task executes for the duration of its time slice and then is forced to relinquish the processor

线程上下文切换

$ pidstat -w -t -p 4256 | head -n 10

23时32分03秒      TGID       TID   cswch/s nvcswch/s  Command
23时32分03秒      4256         -      0.00      0.00  java
23时32分03秒         -      4256      0.00      0.00  |__java
23时32分03秒         -      4259      0.00      0.00  |__java
23时32分03秒         -      4306      0.04      0.01  |__java
23时32分03秒         -      4307      0.04      0.01  |__java
23时32分03秒         -      4308      0.04      0.01  |__java
23时32分03秒         -      4315      0.04      0.01  |__java

jstack

jstack可以查看 线程堆栈 的运行情况，可以结合 pidstat -t -p PID 来查看具体线程的状态

$ sudo /usr/local/java/bin/jstack -F 5850
Thread 5853: (state = BLOCKED)
 - java.lang.Object.wait(long) @bci=0 (Interpreted frame)
 - java.lang.Object.wait() @bci=2, line=503 (Interpreted frame)
 - xx.xx.xx.service.client.proxy.Provider.start() @bci=889, line=208 (Interpreted frame)
 - xx.xx.xx.xx.xx.main(java.lang.String[]) @bci=368, line=90 (Interpreted frame)

原文 http://zhongmingmao.me/2019/09/04/java-performance-command-line-tool/

正文到此结束