当前位置: 首页 > news >正文

jvm调优工具arthas(阿尔萨斯)安装与使用---实践

jvm调优工具arthas(阿尔萨斯)安装与使用—实践

Arthas 是Alibaba开源的Java诊断工具,深受开发者喜爱。

当你遇到以下类似问题而束手无策时,Arthas可以帮助你解决:

  1. 这个类从哪个 jar 包加载的?为什么会报各种类相关的 Exception?
  2. 我改的代码为什么没有执行到?难道是我没 commit?分支搞错了?
  3. 遇到问题无法在线上 debug,难道只能通过加日志再重新发布吗?
  4. 线上遇到某个用户的数据处理有问题,但线上同样无法 debug,线下无法重现!
  5. 是否有一个全局视角来查看系统的运行状况?
  6. 有什么办法可以监控到JVM的实时运行状态?
  7. 怎么快速定位应用的热点,生成火焰图?
  8. 怎样直接从JVM内查找某个类的实例?

Arthas支持JDK 6+(4.x 版本不再支持 JDK 6 和 JDK 7),支持Linux/Mac/Windows,采用命令行交互模式,同时提供丰富的 Tab 自动补全功能,进一步方便进行问题的定位和诊断。

官方文档

1. 启动 math-game

curl -O https://arthas.aliyun.com/math-game.jar
java -jar math-game.jar

math-game是一个简单的程序,每隔一秒生成一个随机数,再执行质因数分解,并打印出分解结果。

2. 启动 arthas

在命令行下面执行(使用和目标进程一致的用户启动,否则可能 attach 失败):

curl -O https://arthas.aliyun.com/arthas-boot.jar
java -jar arthas-boot.jar
  • 执行该程序的用户需要和目标进程具有相同的权限。比如以admin用户来执行:sudo su admin && java -jar arthas-boot.jarsudo -u admin -EH java -jar arthas-boot.jar
  • 如果 attach 不上目标进程,可以查看~/logs/arthas/ 目录下的日志。
  • 如果下载速度比较慢,可以使用 aliyun 的镜像:java -jar arthas-boot.jar --repo-mirror aliyun --use-http
  • java -jar arthas-boot.jar -h 打印更多参数信息。

选择应用 java 进程:

$ $ java -jar arthas-boot.jar
* [1]: 35542[2]: 71560 math-game.jar

math-game进程是第 2 个,则输入 2,再输入回车/enter。Arthas 会 attach 到目标进程上,并输出日志:

[INFO] Try to attach process 71560
[INFO] Attach process 71560 success.
[INFO] arthas-client connect 127.0.0.1 3658,---.  ,------. ,--------.,--.  ,--.  ,---.   ,---./  O  \ |  .--. ''--.  .--'|  '--'  | /  O  \ '   .-'
|  .-.  ||  '--'.'   |  |   |  .--.  ||  .-.  |`.  `-.
|  | |  ||  |\  \    |  |   |  |  |  ||  | |  |.-'    |
`--' `--'`--' '--'   `--'   `--'  `--'`--' `--'`-----'wiki: https://arthas.aliyun.com/doc
version: 3.0.5.20181127201536
pid: 71560
time: 2018-11-28 19:16:24$

3. 查看 dashboard

输入dashboard,按回车/enter,会展示当前进程的信息,按ctrl+c可以中断执行。

$ dashboard
ID     NAME                   GROUP          PRIORI STATE  %CPU    TIME   INTERRU DAEMON
17     pool-2-thread-1        system         5      WAITIN 67      0:0    false   false
27     Timer-for-arthas-dashb system         10     RUNNAB 32      0:0    false   true
11     AsyncAppender-Worker-a system         9      WAITIN 0       0:0    false   true
9      Attach Listener        system         9      RUNNAB 0       0:0    false   true
3      Finalizer              system         8      WAITIN 0       0:0    false   true
2      Reference Handler      system         10     WAITIN 0       0:0    false   true
4      Signal Dispatcher      system         9      RUNNAB 0       0:0    false   true
26     as-command-execute-dae system         10     TIMED_ 0       0:0    false   true
13     job-timeout            system         9      TIMED_ 0       0:0    false   true
1      main                   main           5      TIMED_ 0       0:0    false   false
14     nioEventLoopGroup-2-1  system         10     RUNNAB 0       0:0    false   false
18     nioEventLoopGroup-2-2  system         10     RUNNAB 0       0:0    false   false
23     nioEventLoopGroup-2-3  system         10     RUNNAB 0       0:0    false   false
15     nioEventLoopGroup-3-1  system         10     RUNNAB 0       0:0    false   false
Memory             used   total max    usage GC
heap               32M    155M  1820M  1.77% gc.ps_scavenge.count  4
ps_eden_space      14M    65M   672M   2.21% gc.ps_scavenge.time(m 166
ps_survivor_space  4M     5M    5M           s)
ps_old_gen         12M    85M   1365M  0.91% gc.ps_marksweep.count 0
nonheap            20M    23M   -1           gc.ps_marksweep.time( 0
code_cache         3M     5M    240M   1.32% ms)
Runtime
os.name                Mac OS X
os.version             10.13.4
java.version           1.8.0_162
java.home              /Library/Java/JavaVirtualMachines/jdk1.8.0_162.jdk/Contents/Home/jre

4. 通过 thread 命令来获取到math-game进程的 Main Class

thread 1会打印线程 ID 1 的栈,通常是 main 函数的线程。

$ thread 1 | grep 'main('at demo.MathGame.main(MathGame.java:17)

5. 通过 jad 来反编译 Main Class

$ jad demo.MathGameClassLoader:
+-sun.misc.Launcher$AppClassLoader@3d4eac69+-sun.misc.Launcher$ExtClassLoader@66350f69Location:
/tmp/math-game.jar/** Decompiled with CFR 0_132.*/
package demo;import java.io.PrintStream;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import java.util.Random;
import java.util.concurrent.TimeUnit;public class MathGame {private static Random random = new Random();private int illegalArgumentCount = 0;public static void main(String[] args) throws InterruptedException {MathGame game = new MathGame();do {game.run();TimeUnit.SECONDS.sleep(1L);} while (true);}public void run() throws InterruptedException {try {int number = random.nextInt();List<Integer> primeFactors = this.primeFactors(number);MathGame.print(number, primeFactors);}catch (Exception e) {System.out.println(String.format("illegalArgumentCount:%3d, ", this.illegalArgumentCount) + e.getMessage());}}public static void print(int number, List<Integer> primeFactors) {StringBuffer sb = new StringBuffer("" + number + "=");Iterator<Integer> iterator = primeFactors.iterator();while (iterator.hasNext()) {int factor = iterator.next();sb.append(factor).append('*');}if (sb.charAt(sb.length() - 1) == '*') {sb.deleteCharAt(sb.length() - 1);}System.out.println(sb);}public List<Integer> primeFactors(int number) {if (number < 2) {++this.illegalArgumentCount;throw new IllegalArgumentException("number is: " + number + ", need >= 2");}ArrayList<Integer> result = new ArrayList<Integer>();int i = 2;while (i <= number) {if (number % i == 0) {result.add(i);number /= i;i = 2;continue;}++i;}return result;}
}Affect(row-cnt:1) cost in 970 ms.

6. watch

通过watch命令来查看demo.MathGame#primeFactors函数的返回值:

$ watch demo.MathGame primeFactors returnObj
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 107 ms.
ts=2018-11-28 19:22:30; [cost=1.715367ms] result=null
ts=2018-11-28 19:22:31; [cost=0.185203ms] result=null
ts=2018-11-28 19:22:32; [cost=19.012416ms] result=@ArrayList[@Integer[5],@Integer[47],@Integer[2675531],
]
ts=2018-11-28 19:22:33; [cost=0.311395ms] result=@ArrayList[@Integer[2],@Integer[5],@Integer[317],@Integer[503],@Integer[887],
]
ts=2018-11-28 19:22:34; [cost=10.136007ms] result=@ArrayList[@Integer[2],@Integer[2],@Integer[3],@Integer[3],@Integer[31],@Integer[717593],
]
ts=2018-11-28 19:22:35; [cost=29.969732ms] result=@ArrayList[@Integer[5],@Integer[29],@Integer[7651739],
]

7. 退出 arthas

如果只是退出当前的连接,可以用quit或者exit命令。Attach 到目标进程上的 arthas 还会继续运行,端口会保持开放,下次连接时可以直接连接上。

如果想完全退出 arthas,可以执行stop命令。

参考文档

  • 在线教程(推荐)

  • 用户文档

  • 安装

  • 下载

  • 快速入门

  • 进阶使用

  • 命令列表

  • WebConsole

  • Docker

  • Arthas Spring Boot Starter

  • 用户案例

  • FAQ/常见问题

  • 编译调试/参与贡献

  • Release Notes

案例展示

Dashboard
  • https://arthas.aliyun.com/doc/dashboard
Thread
  • https://arthas.aliyun.com/doc/thread

一目了然的了解系统的状态,哪些线程比较占cpu?他们到底在做什么?

$ thread -n 3
"as-command-execute-daemon" Id=29 cpuUsage=75% RUNNABLEat sun.management.ThreadImpl.dumpThreads0(Native Method)at sun.management.ThreadImpl.getThreadInfo(ThreadImpl.java:440)at com.taobao.arthas.core.command.monitor200.ThreadCommand$1.action(ThreadCommand.java:58)at com.taobao.arthas.core.command.handler.AbstractCommandHandler.execute(AbstractCommandHandler.java:238)at com.taobao.arthas.core.command.handler.DefaultCommandHandler.handleCommand(DefaultCommandHandler.java:67)at com.taobao.arthas.core.server.ArthasServer$4.run(ArthasServer.java:276)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)at java.lang.Thread.run(Thread.java:745)Number of locked synchronizers = 1- java.util.concurrent.ThreadPoolExecutor$Worker@6cd0b6f8"as-session-expire-daemon" Id=25 cpuUsage=24% TIMED_WAITINGat java.lang.Thread.sleep(Native Method)at com.taobao.arthas.core.server.DefaultSessionManager$2.run(DefaultSessionManager.java:85)"Reference Handler" Id=2 cpuUsage=0% WAITING on java.lang.ref.Reference$Lock@69ba0f27at java.lang.Object.wait(Native Method)-  waiting on java.lang.ref.Reference$Lock@69ba0f27at java.lang.Object.wait(Object.java:503)at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:133)
jad
  • https://arthas.aliyun.com/doc/jad

对类进行反编译:

$ jad javax.servlet.ServletClassLoader:
+-java.net.URLClassLoader@6108b2d7+-sun.misc.Launcher$AppClassLoader@18b4aac2+-sun.misc.Launcher$ExtClassLoader@1ddf84b8Location:
/Users/xxx/work/test/lib/servlet-api.jar/** Decompiled with CFR 0_122.*/
package javax.servlet;import java.io.IOException;
import javax.servlet.ServletConfig;
import javax.servlet.ServletException;
import javax.servlet.ServletRequest;
import javax.servlet.ServletResponse;public interface Servlet {public void init(ServletConfig var1) throws ServletException;public ServletConfig getServletConfig();public void service(ServletRequest var1, ServletResponse var2) throws ServletException, IOException;public String getServletInfo();public void destroy();
}
mc
  • https://arthas.aliyun.com/doc/mc

Memory Compiler/内存编译器,编译.java文件生成.class

mc /tmp/Test.java
retransform
  • https://arthas.aliyun.com/doc/retransform

加载外部的.class文件,retransform 热更新jvm已加载的类。

retransform /tmp/Test.class
retransform -c 327a647b /tmp/Test.class /tmp/Test\$Inner.class
sc
  • https://arthas.aliyun.com/doc/sc

查找JVM中已经加载的类

$ sc -d org.springframework.web.context.support.XmlWebApplicationContextclass-info        org.springframework.web.context.support.XmlWebApplicationContextcode-source       /Users/xxx/work/test/WEB-INF/lib/spring-web-3.2.11.RELEASE.jarname              org.springframework.web.context.support.XmlWebApplicationContextisInterface       falseisAnnotation      falseisEnum            falseisAnonymousClass  falseisArray           falseisLocalClass      falseisMemberClass     falseisPrimitive       falseisSynthetic       falsesimple-name       XmlWebApplicationContextmodifier          publicannotationinterfacessuper-class       +-org.springframework.web.context.support.AbstractRefreshableWebApplicationContext+-org.springframework.context.support.AbstractRefreshableConfigApplicationContext+-org.springframework.context.support.AbstractRefreshableApplicationContext+-org.springframework.context.support.AbstractApplicationContext+-org.springframework.core.io.DefaultResourceLoader+-java.lang.Objectclass-loader      +-org.apache.catalina.loader.ParallelWebappClassLoader+-java.net.URLClassLoader@6108b2d7+-sun.misc.Launcher$AppClassLoader@18b4aac2+-sun.misc.Launcher$ExtClassLoader@1ddf84b8classLoaderHash   25131501
vmtool
  • https://arthas.aliyun.com/doc/vmtool

从JVM heap中获取指定类的实例。

$ vmtool --action getInstances --className java.lang.String --limit 10
@String[][@String[com/taobao/arthas/core/shell/session/Session],@String[com.taobao.arthas.core.shell.session.Session],@String[com/taobao/arthas/core/shell/session/Session],@String[com/taobao/arthas/core/shell/session/Session],@String[com/taobao/arthas/core/shell/session/Session.class],@String[com/taobao/arthas/core/shell/session/Session.class],@String[com/taobao/arthas/core/shell/session/Session.class],@String[com/],@String[java/util/concurrent/ConcurrentHashMap$ValueIterator],@String[java/util/concurrent/locks/LockSupport],
]
stack
  • https://arthas.aliyun.com/doc/stack

查看方法 test.arthas.TestStack#doGet 的调用堆栈:

$ stack test.arthas.TestStack doGet
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 286 ms.
ts=2018-09-18 10:11:45;thread_name=http-bio-8080-exec-10;id=d9;is_daemon=true;priority=5;TCCL=org.apache.catalina.loader.ParallelWebappClassLoader@25131501@test.arthas.TestStack.doGet()at javax.servlet.http.HttpServlet.service(HttpServlet.java:624)at javax.servlet.http.HttpServlet.service(HttpServlet.java:731)at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(ApplicationFilterChain.java:303)at org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationFilterChain.java:208)at org.apache.tomcat.websocket.server.WsFilter.doFilter(WsFilter.java:52)at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(ApplicationFilterChain.java:241)at org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationFilterChain.java:208)at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(ApplicationFilterChain.java:241)at org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationFilterChain.java:208)at org.apache.catalina.core.StandardWrapperValve.invoke(StandardWrapperValve.java:220)at org.apache.catalina.core.StandardContextValve.invoke(StandardContextValve.java:110)...at org.apache.catalina.core.StandardHostValve.invoke(StandardHostValve.java:169)at org.apache.catalina.valves.ErrorReportValve.invoke(ErrorReportValve.java:103)at org.apache.catalina.core.StandardEngineValve.invoke(StandardEngineValve.java:116)at org.apache.catalina.connector.CoyoteAdapter.service(CoyoteAdapter.java:451)at org.apache.coyote.http11.AbstractHttp11Processor.process(AbstractHttp11Processor.java:1121)at org.apache.coyote.AbstractProtocol$AbstractConnectionHandler.process(AbstractProtocol.java:637)at org.apache.tomcat.util.net.JIoEndpoint$SocketProcessor.run(JIoEndpoint.java:316)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)at org.apache.tomcat.util.threads.TaskThread$WrappingRunnable.run(TaskThread.java:61)at java.lang.Thread.run(Thread.java:745)
Trace
  • https://arthas.aliyun.com/doc/trace

观察方法执行的时候哪个子调用比较慢:

Watch
  • https://arthas.aliyun.com/doc/watch

观察方法 test.arthas.TestWatch#doGet 执行的入参,仅当方法抛出异常时才输出。

$ watch test.arthas.TestWatch doGet {params[0], throwExp} -e
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 65 ms.
ts=2018-09-18 10:26:28;result=@ArrayList[@RequestFacade[org.apache.catalina.connector.RequestFacade@79f922b2],@NullPointerException[java.lang.NullPointerException],
]
Monitor
  • https://arthas.aliyun.com/doc/monitor

监控某个特殊方法的调用统计数据,包括总调用次数,平均rt,成功率等信息,每隔5秒输出一次。

$ monitor -c 5 org.apache.dubbo.demo.provider.DemoServiceImpl sayHello
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 109 ms.timestamp            class                                           method    total  success  fail  avg-rt(ms)  fail-rate
----------------------------------------------------------------------------------------------------------------------------2018-09-20 09:45:32  org.apache.dubbo.demo.provider.DemoServiceImpl  sayHello  5      5        0     0.67        0.00%timestamp            class                                           method    total  success  fail  avg-rt(ms)  fail-rate
----------------------------------------------------------------------------------------------------------------------------2018-09-20 09:45:37  org.apache.dubbo.demo.provider.DemoServiceImpl  sayHello  5      5        0     1.00        0.00%timestamp            class                                           method    total  success  fail  avg-rt(ms)  fail-rate
----------------------------------------------------------------------------------------------------------------------------2018-09-20 09:45:42  org.apache.dubbo.demo.provider.DemoServiceImpl  sayHello  5      5        0     0.43        0.00%
Time Tunnel(tt)
  • https://arthas.aliyun.com/doc/tt

记录方法调用信息,支持事后查看方法调用的参数,返回值,抛出的异常等信息,仿佛穿越时空隧道回到调用现场一般。

$ tt -t org.apache.dubbo.demo.provider.DemoServiceImpl sayHello
Press Ctrl+C to abort.
Affect(class-cnt:1 , method-cnt:1) cost in 75 ms.INDEX   TIMESTAMP            COST(ms)  IS-RET  IS-EXP   OBJECT         CLASS                          METHOD
-------------------------------------------------------------------------------------------------------------------------------------1000    2018-09-20 09:54:10  1.971195  true    false    0x55965cca     DemoServiceImpl                sayHello1001    2018-09-20 09:54:11  0.215685  true    false    0x55965cca     DemoServiceImpl                sayHello1002    2018-09-20 09:54:12  0.236303  true    false    0x55965cca     DemoServiceImpl                sayHello1003    2018-09-20 09:54:13  0.159598  true    false    0x55965cca     DemoServiceImpl                sayHello1004    2018-09-20 09:54:14  0.201982  true    false    0x55965cca     DemoServiceImpl                sayHello1005    2018-09-20 09:54:15  0.214205  true    false    0x55965cca     DemoServiceImpl                sayHello1006    2018-09-20 09:54:16  0.241863  true    false    0x55965cca     DemoServiceImpl                sayHello1007    2018-09-20 09:54:17  0.305747  true    false    0x55965cca     DemoServiceImpl                sayHello1008    2018-09-20 09:54:18  0.18468   true    false    0x55965cca     DemoServiceImpl                sayHello
Classloader
  • https://arthas.aliyun.com/doc/classloader

了解当前系统中有多少类加载器,以及每个加载器加载的类数量,帮助您判断是否有类加载器泄露。

$ classloadername                                                  numberOfInstances  loadedCountTotalBootstrapClassLoader                                  1                  3346com.taobao.arthas.agent.ArthasClassloader             1                  1262java.net.URLClassLoader                               2                  1033org.apache.catalina.loader.ParallelWebappClassLoader  1                  628sun.reflect.DelegatingClassLoader                     166                166sun.misc.Launcher$AppClassLoader                      1                  31com.alibaba.fastjson.util.ASMClassLoader              6                  15sun.misc.Launcher$ExtClassLoader                      1                  7org.jvnet.hk2.internal.DelegatingClassLoader          2                  2sun.reflect.misc.MethodUtil                           1                  1
Web Console
  • https://arthas.aliyun.com/doc/web-console

外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传

Profiler/FlameGraph/火焰图
  • https://arthas.aliyun.com/doc/profiler
$ profiler start
Started [cpu] profiling
$ profiler stop
profiler output file: /tmp/demo/arthas-output/20211207-111550.html
OK

通过浏览器查看profiler结果:

外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传

Arthas Spring Boot Starter
  • Arthas Spring Boot Starter

相关文章:

  • 宫庭海:跨界融合与家国情怀的影视开拓者
  • 机器学习 Day11 决策树
  • Ubuntu 20.04.6编译安装COMFAST CF-AX90无线网卡驱动
  • OpenHarmony Camera开发指导(四):相机会话管理(ArkTS)
  • js day2
  • 十五种光电器件综合对比——《器件手册--光电器件》
  • FastAPI与SQLAlchemy数据库集成
  • 2025.04.17【Dendrogram】生信数据可视化:Dendrogram图表详解
  • 无人机姿态稳定与动态控制模块概述!
  • 第六章:6.3求一个3*3的整型矩阵对角线元素之和
  • 基于FreeRTOS和STM32的微波炉
  • Google C++ Style Guide
  • Linux 文件传输:系统数据交互的动脉
  • 极狐GitLab 外部流水线验证功能解读
  • SpringAI+DeepSeek大模型应用开发——4 对话机器人
  • 多模态大模型的算力需求预测:从理论FLOPs到实际集群配置(搭建算力成本评估模型的方法论)
  • Linux,redis群集模式,主从复制,读写分离
  • 山东大学软件学院创新项目实训开发日志(15)之中医知识问答历史对话查看bug处理后端信息响应成功但前端未获取到
  • 在Pycharm配置stable diffusion环境(使用conda虚拟环境)
  • Python(19)Python并发编程:深入解析多线程与多进程的差异及锁机制实战
  • “80后”德州市接待事务中心副主任刘巍“拟进一步使用”
  • 市场监管总局等五部门约谈外卖平台企业
  • 中国创面修复学科发起者之一陆树良教授病逝,享年64岁
  • 长三角议事厅·周报|从模速空间看上海街区化AI孵化模式
  • 云南一男子持刀致邻居3死1重伤案二审开庭,未当庭宣判
  • 上海国际电影节特设“今日亚洲”单元