13.4.虚拟化工具--jmap详解

Posted to.to

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了13.4.虚拟化工具--jmap详解相关的知识,希望对你有一定的参考价值。

13.4.虚拟化工具–jmap详解
13.4.1.jmap命令基本概述
13.4.2.命令格式
13.4.3.参数说明

13.4.虚拟化工具–jmap详解

13.4.1.jmap命令基本概述

jmap命令是一个可输出所有内存中对象的工具,甚至可以将VM中的heap,以二进制输出成文本。

64位机上使用需要使用如下方式:

jmap -J-d64 -heap pid

13.4.2.命令格式

jmap [option] <pid>
        (to connect to running process) 连接到正在运行的进程

    jmap [option] <executable <core>
        (to connect to a core file)     连接到核心文件

    jmap [option] [server_id@]<remote server IP or hostname>
        (to connect to remote debug server) 连接到远程调试服务

13.4.3.参数说明

1)options

pid:         目标进程的PID,进程编号,可以采用ps -ef | grep java 查看java进程的PID;
executable:     产生core dump的java可执行程序;
core:          将被打印信息的core dump文件;
remote-hostname-or-IP:     远程debug服务的主机名或ip;
server-id:      唯一id,假如一台主机上多个远程debug服务;

2)基本参数:
[就是替换[option]位置的参数]

1> -dump:[live,]format=b,file= 使用hprof二进制形式,输出jvm的heap内容到文件=. live子选项是可选的,假如指定live选项,那么只输出活的对象到文件.

命令:

jmap -dump:live,format=b,file=myjmapfile.txt 19570

再如:

[admin@hadoop3 ~]$ jmap -dump:live,format=b,file=aaa.hprof 449415

结果:

即可在/root目录打开myjmapfile.txt文件。
当然,file=后面也可以指定文件存放的目录,就可以在指定目录查看文件了。

2> -finalizerinfo 打印正等候回收的对象的信息
命令:

jmap -finalizerinfo 3772


3> -heap打印heap的概要信息,GC使用的算法,heap(堆)的配置及JVM堆内存的使用情况.

命令:

[admin@hadoop3 ~]$ jmap -finalizerinfo 449415

结果:

Attaching to process ID 449415, please wait...
Debugger attached successfully.
Server compiler detected.
JVM version is 25.60-b23
Number of objects pending for finalization: 0
[admin@hadoop3 ~]$ jmap -heap 449415
Attaching to process ID 449415, please wait...
Debugger attached successfully.
Server compiler detected.
JVM version is 25.60-b23

using parallel threads in the new generation.
using thread-local object allocation.
Concurrent Mark-Sweep GC

Heap Configuration:
   MinHeapFreeRatio         = 40
   MaxHeapFreeRatio         = 70
   MaxHeapSize              = 53687091200 (51200.0MB)
   NewSize                  = 3315138560 (3161.5625MB)
   MaxNewSize               = 3315138560 (3161.5625MB)
   OldSize                  = 50371952640 (48038.4375MB)
   NewRatio                 = 2
   SurvivorRatio            = 8
   MetaspaceSize            = 21807104 (20.796875MB)
   CompressedClassSpaceSize = 1073741824 (1024.0MB)
   MaxMetaspaceSize         = 17592186044415 MB
   G1HeapRegionSize         = 0 (0.0MB)

Heap Usage:
New Generation (Eden + 1 Survivor Space):
   capacity = 2983657472 (2845.4375MB)
   used     = 1304315248 (1243.8919525146484MB)
   free     = 1679342224 (1601.5455474853516MB)
   43.715314517175244% used
Eden Space:
   capacity = 2652176384 (2529.3125MB)
   used     = 1237992080 (1180.6412506103516MB)
   free     = 1414184304 (1348.6712493896484MB)
   46.67834641272486% used
From Space:
   capacity = 331481088 (316.125MB)
   used     = 66323168 (63.250701904296875MB)
   free     = 265157920 (252.87429809570312MB)
   20.008130297919138% used
To Space:
   capacity = 331481088 (316.125MB)
   used     = 0 (0.0MB)
   free     = 331481088 (316.125MB)
   0.0% used
concurrent mark-sweep generation:
   capacity = 50371952640 (48038.4375MB)
   used     = 6296298032 (6004.617721557617MB)
   free     = 44075654608 (42033.81977844238MB)
   12.499610799284671% used

32892 interned Strings occupying 4809320 bytes.

获取转储快照信息等
解析:

using parallel threads in the new generation.  ##新生代采用的是并行线程处理方式
using thread-local object allocation.   
Concurrent Mark-Sweep GC   ##同步并行垃圾回收

Heap Configuration:  ##堆配置情况,也就是JVM参数配置的结果[平常说的tomcat配置JVM参数,就是在配置这些]
   MinHeapFreeRatio = 40 ##最小堆使用比例
   MaxHeapFreeRatio = 70 ##最大堆可用比例
   MaxHeapSize      = 2147483648 (2048.0MB) ##最大堆空间大小
   NewSize          = 268435456 (256.0MB) ##新生代分配大小
   MaxNewSize       = 268435456 (256.0MB) ##最大可新生代分配大小
   OldSize           = 5439488 (5.1875MB) ##老年代大小
   NewRatio         = 2  ##新生代比例
   SurvivorRatio    = 8 ##新生代与suvivor的比例
   PermSize         = 134217728 (128.0MB) ##perm区 永久代大小
   MaxPermSize      = 134217728 (128.0MB) ##最大可分配perm区 也就是永久代大小
 
Heap Usage: ##堆使用情况【堆内存实际的使用情况】

New Generation (Eden + 1 Survivor Space):  ##新生代(伊甸区Eden区 + 幸存区survior(1+2)空间)
   capacity = 241631232 (230.4375MB)  ##伊甸区容量
   used     = 77776272 (74.17323303222656MB) ##已经使用大小
   free     = 163854960 (156.26426696777344MB) ##剩余容量
   32.188004570534986% used ##使用比例

Eden Space:  ##伊甸区
   capacity = 214827008 (204.875MB) ##伊甸区容量
   used     = 74442288 (70.99369812011719MB) ##伊甸区使用
   free     = 140384720 (133.8813018798828MB) ##伊甸区当前剩余容量
   34.65220164496263% used ##伊甸区使用情况

From Space: ##survior1区
   capacity = 26804224 (25.5625MB) ##survior1区容量
   used     = 3333984 (3.179534912109375MB) ##surviror1区已使用情况
   free     = 23470240 (22.382965087890625MB) ##surviror1区剩余容量
   12.43827838477995% used ##survior1区使用比例

To Space: ##survior2 区
   capacity = 26804224 (25.5625MB) ##survior2区容量
   used     = 0 (0.0MB) ##survior2区已使用情况
   free     = 26804224 (25.5625MB) ##survior2区剩余容量
   0.0% used ## survior2区使用比例

PS Old  Generation: ##老年代使用情况
   capacity = 1879048192 (1792.0MB) ##老年代容量
   used     = 30847928 (29.41887664794922MB) ##老年代已使用容量
   free     = 1848200264 (1762.5811233520508MB) ##老年代剩余容量
   1.6416783843721663% used ##老年代使用比例

Perm Generation: ##永久代使用情况
   capacity = 134217728 (128.0MB) ##perm区容量
   used     = 47303016 (45.111671447753906MB) ##perm区已使用容量
   free     = 86914712 (82.8883285522461MB) ##perm区剩余容量
   35.24349331855774% used ##perm区使用比例

4> -histo[:live] 打印每个class的实例数目,内存占用,类全名信息. VM的内部类名字开头会加上前缀”*”. 如果live子参数加上后,只统计活的对象数量.

命令:

jmap -histo:live 19570

结果:

num     #instances(实例)   #bytes(字节大小)  class name(类名)

----------------------------------------------

   1:         65220        9755240      <constMethodKlass>
   2:         65220        8880384      <methodKlass>
   3:         11721        8252112      [B
   4:          6300        6784040      <constantPoolKlass>
   5:         75224        6218208      [C
   6:         93969        5163280      <symbolKlass>
   7:          6300        4854440      <instanceKlassKlass>
   8:          5482        4203152      <constantPoolCacheKlass>
   9:         72097        2307104      java.lang.String
  10:         15102        2289912      [I
  11:          4089        2227728      <methodDataKlass>
  12:         28887        1386576      org.apache.velocity.runtime.parser.Token
  13:          6792         706368          java.lang.Class
  14:          7445         638312          [Ljava.util.HashMap$Entry;

4380:             1             16          com.sun.proxy.$Proxy208
4381:             1             16          sun.reflect.GeneratedMethodAccessor198
4382:             1             16          com.sun.proxy.$Proxy46
4383:             1             16          org.apache.ibatis.ognl.SetPropertyAccessor
4384:             1             16          oracle.jdbc.driver.OracleDriver
4385:             1             16          com.sun.proxy.$Proxy181
4386:             1             16          com.sun.proxy.$Proxy82
4387:             1             16          java.util.jar.JavaUtilJarAccessImpl
4388:             1             16          com.sun.proxy.$Proxy171
4389:             1             16          sun.reflect.GeneratedMethodAccessor136
4390:             1             16            sun.reflect.GeneratedConstructorAccessor22
4391:             1             16          org.elasticsearch.action.search.SearchAction
4392:             1             16          org.springframework.core.annotation.AnnotationAwareOrderComparator
Total       1756265      162523736

采用jmap -histo pid>a.log日志将其保存,在一段时间后,使用文本对比工具,可以对比出GC回收了哪些对象。

jmap -dump:format=b,file=outfile 3024可以将3024进程的内存heap输出出来到outfile文件里,再配合MAT(内存分析工具)。

5> -permstat 打印classload和jvm heap长久层的信息. 包含每个classloader的名字,活泼性,地址,父classloader和加载的class数量. 另外,内部String的数量和占用内存数也会打印出来.

命令:

jmap -permstat 19570

结果:

解析:

Attaching to process ID 19570, please wait...
Debugger attached successfully.
Server compiler detected.
JVM version is 24.80-b11
finding class loader instances ..done.
computing per loader stat ..done.
please wait.. computing liveness.liveness analysis may be inaccurate ...
class_loader    classes    bytes    parent_loader    alive?    type

<bootstrap>    2538    14654264      null      live    <internal>
0x000000070af968c8    63    399160    0x0000000707db1788    dead    org/apache/catalina/loader/WebappClassLoader@0x000000070367d2a8
0x000000070cba7b08    1    3064    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba6a38    28    221344    0x0000000707e709a8    dead    org/apache/jasperrvlet/JasperLoader@0x0000000705b11878
0x000000070baed8b8    32    297296    0x0000000707e709a8    dead    org/apache/jasperrvlet/JasperLoader@0x0000000705b11878
0x000000070919a610    1    3056    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070bdd1e18    1    3064    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x0000000707f1d480    1    3072      null      dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba7f48    1    1912    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba8508    1    3064    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba6c40    1    3064    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070bd4c6a0    1    3056      null      dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba62b0    28    235472    0x0000000707e709a8    dead    org/apache/jasperrvlet/JasperLoader@0x0000000705b11878
0x000000070cba77c8    1    3064    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba7388    1    3064    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba8148    1    3064    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070afd8b60    1    6704    0x0000000707db1788    dead    org/apache/catalina/loader/WebappClassLoader@0x000000070367d2a8
0x000000070919a410    1    1888      null      dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070bdd05b0    1    1912      null      dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070bc848b8    1    3088    0x0000000707db1788    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba64e8    1    1888    0x0000000707e709a8    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x0000000707f1d2c0    1    3064    0x0000000707db1788    dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070be82e38    1    1912      null      dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98
0x000000070cba7908    1    3208      null      dead    sun/reflect/DelegatingClassLoader@0x0000000702a50b98.........
total = 273    12995    87547304        N/A        alive=1, dead=272        N/A

6> -F 强迫.在pid没有相应的时候使用-dump或者-histo参数. 在这个模式下,live子参数无效.
7> -h | -help 打印辅助信息
8> -J 传递参数给jmap启动的jvm.

除了通过jmap,还可以通过配置以下方式获取到:

以上是关于13.4.虚拟化工具--jmap详解的主要内容,如果未能解决你的问题,请参考以下文章

深入理解Java虚拟机——Java内存映像工具(Jmap)

深入理解Java虚拟机——Java内存映像工具(Jmap)

jmap命令详解

jmap命令详解

jmap命令详解

jmap命令详解