Java虚拟机 - 实战篇

一、内存调优

1. 什么是内存泄漏

（1）内存溢出和内存泄漏

2. 监控Java内存的常用工具

（1）Top命令

（2）VisualVM

（3）Arthas

（4）Prometheus + Grafana

（5）堆内存状况的对比

3. 内存泄漏的常见场景

（1）equals()和hashCode()导致的内存泄漏

package com.itheima.jvmoptimize.leakdemo.demo2;

import org.apache.commons.lang3.builder.EqualsBuilder;
import org.apache.commons.lang3.builder.HashCodeBuilder;

import java.util.Objects;

public class Student {
    private String name;
    private Integer id;
    private byte[] bytes = new byte[1024 * 1024];

    public String getName() {
        return name;
    }

    public void setName(String name) {
        this.name = name;
    }

    public Integer getId() {
        return id;
    }

    public void setId(Integer id) {
        this.id = id;
    }

    @Override
    public boolean equals(Object o) {
        if (this == o) {
            return true;
        }

        if (o == null || getClass() != o.getClass()) {
            return false;
        }

        Student student = (Student) o;

        return new EqualsBuilder().append(id, student.id).isEquals();
    }

    @Override
    public int hashCode() {
        return new HashCodeBuilder(17, 37).append(id).toHashCode();
    }
}

（2）内部类引用外部类

package com.itheima.jvmoptimize.leakdemo.demo3;

import java.io.IOException;
import java.util.ArrayList;

public class Outer{
    private byte[] bytes = new byte[1024 * 1024]; //外部类持有数据
    private static String name  = "测试";
    class Inner{
        private String name;
        public Inner() {
            this.name = Outer.name;
        }
    }

    public static void main(String[] args) throws IOException, InterruptedException {
//        System.in.read();
        int count = 0;
        ArrayList<Inner> inners = new ArrayList<>();
        while (true){
            if(count++ % 100 == 0){
                Thread.sleep(10);
            }
            inners.add(new Outer().new Inner());
        }
    }
}

package com.itheima.jvmoptimize.leakdemo.demo4;

import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

public class Outer {
    private byte[] bytes = new byte[1024 * 1024 * 10];
    public List<String> newList() {
        List<String> list = new ArrayList<String>() {{
            add("1");
            add("2");
        }};
        return list;
    }

    public static void main(String[] args) throws IOException {
        System.in.read();
        int count = 0;
        ArrayList<Object> objects = new ArrayList<>();
        while (true){
            System.out.println(++count);
            objects.add(new Outer().newList());
        }
    }
}

package com.itheima.jvmoptimize.leakdemo.demo3;

import java.io.IOException;
import java.util.ArrayList;

public class Outer{
    private byte[] bytes = new byte[1024 * 1024]; //外部类持有数据
    private static String name  = "测试";
    static class Inner{
        private String name;
        public Inner() {
            this.name = Outer.name;
        }
    }

    public static void main(String[] args) throws IOException, InterruptedException {
//        System.in.read();
        int count = 0;
        ArrayList<Inner> inners = new ArrayList<>();
        while (true){
            if(count++ % 100 == 0){
                Thread.sleep(10);
            }
            inners.add(new Inner());
        }
    }
}

package com.itheima.jvmoptimize.leakdemo.demo4;

import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

public class Outer {
    private byte[] bytes = new byte[1024 * 1024 * 10];
    public static List<String> newList() {
        List<String> list = new ArrayList<String>() {{
            add("1");
            add("2");
        }};
        return list;
    }

    public static void main(String[] args) throws IOException {
        System.in.read();
        int count = 0;
        ArrayList<Object> objects = new ArrayList<>();
        while (true){
            System.out.println(++count);
            objects.add(newList());
        }
    }
}

（3）ThreadLocal的使用

package com.itheima.jvmoptimize.leakdemo.demo5;

import java.util.concurrent.*;

public class Demo5 {
    public static ThreadLocal<Object> threadLocal = new ThreadLocal<>();

    public static void main(String[] args) throws InterruptedException {
        ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(Integer.MAX_VALUE, Integer.MAX_VALUE,
                0, TimeUnit.DAYS, new SynchronousQueue<>());
        int count = 0;
        while (true) {
            System.out.println(++count);
            threadPoolExecutor.execute(() -> {
                threadLocal.set(new byte[1024 * 1024]);
                // threadLocal.remove();
            });
            Thread.sleep(10);
        }


    }
}

（4）String的intern方法

package com.itheima.jvmoptimize.leakdemo.demo6;

import org.apache.commons.lang3.RandomStringUtils;

import java.util.ArrayList;
import java.util.List;

public class Demo6 {
    public static void main(String[] args) {
        while (true){
            List<String> list = new ArrayList<String>();
            int i = 0;
            while (true) {
                //String.valueOf(i++).intern(); //JDK1.6 perm gen 不会溢出
                list.add(String.valueOf(i++).intern()); //溢出
            }
        }
    }
}

（5）通过静态字段保存对象

单例模式 -> 懒加载

package com.itheima.jvmoptimize.leakdemo.demo7;

import org.springframework.context.annotation.Lazy;
import org.springframework.stereotype.Component;

@Lazy //懒加载
@Component
public class TestLazy {
    private byte[] bytes = new byte[1024 * 1024 * 1024];
}

Spring的Bean中不要长期存放大对象，如果是缓存用于提升性能，尽量设置过期时间

package com.itheima.jvmoptimize.leakdemo.demo7;

import com.github.benmanes.caffeine.cache.Cache;
import com.github.benmanes.caffeine.cache.Caffeine;

import java.time.Duration;

public class CaffineDemo {
    public static void main(String[] args) throws InterruptedException {
        Cache<Object, Object> build = Caffeine.newBuilder()
        //设置100ms之后就过期
                 .expireAfterWrite(Duration.ofMillis(100))
                .build();
        int count = 0;
        while (true){
            build.put(count++,new byte[1024 * 1024 * 10]);
            Thread.sleep(100L);
        }
    }
}

（6）资源没有正常关闭

package com.itheima.jvmoptimize.leakdemo.demo1;

import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.net.URL;
import java.net.URLConnection;
import java.sql.*;

//-Xmx50m -Xms50m
public class Demo1 {

    // JDBC driver name and database URL
    static final String JDBC_DRIVER = "com.mysql.cj.jdbc.Driver";
    static final String DB_URL = "jdbc:mysql:///bank1";

    //  Database credentials
    static final String USER = "root";
    static final String PASS = "123456";

    public static void leak() throws SQLException {
        //Connection conn = null;
        Statement stmt = null;
        Connection conn = DriverManager.getConnection(DB_URL, USER, PASS);

        // executes a valid query
        stmt = conn.createStatement();
        String sql;
        sql = "SELECT id, account_name FROM account_info";
        ResultSet rs = stmt.executeQuery(sql);

        //STEP 4: Extract data from result set
        while (rs.next()) {
            //Retrieve by column name
            int id = rs.getInt("id");
            String name = rs.getString("account_name");

            //Display values
            System.out.print("ID: " + id);
            System.out.print(", Name: " + name + "\n");
        }

    }

    public static void main(String[] args) throws InterruptedException, SQLException {
        while (true) {
            leak();
        }
    }
}

4. 内存泄漏的解决方案

（1）内存快照

package com.itheima.jvmoptimize.matdemo;

import org.openjdk.jol.info.ClassLayout;

import java.util.ArrayList;
import java.util.List;

//-XX:+HeapDumpBeforeFullGC -XX:HeapDumpPath=D:/jvm/dump/mattest.hprof
public class HeapDemo {
    public static void main(String[] args) {
        TestClass a1 = new TestClass();
        TestClass a2 = new TestClass();
        TestClass a3 = new TestClass();
        String s1 = "itheima1";
        String s2 = "itheima2";
        String s3 = "itheima3";

        a1.list.add(s1);

        a2.list.add(s1);
        a2.list.add(s2);

        a3.list.add(s3);

        //System.out.print(ClassLayout.parseClass(TestClass.class).toPrintable());
        s1 = null;
        s2 = null;
        s3 = null;
        System.gc();
    }
}

class TestClass {
    public List<String> list = new ArrayList<>(10);
}

或

（2）修复问题

总结：

二、GC调优

1. GC调优的核心指标

（1）垃圾回收吞吐量

（2）延迟

（3）内存使用量

2. GC调优的方法

（1）发现问题

（2）常见的GC模式

（3）解决GC问题的手段

（3-1）优化基础JVM参数

（3-2）减少对象产生

（3-3）更换垃圾回收器

（3-4）优化垃圾回收器参数

三、性能调优

1. 性能调优解决的问题

2. 性能调优的方法

（1）线程转储的查看方式

（2）请求单个服务处理时间特别长

（3）程序启动之后运行正常，但是在运行一段时间之后无法处理任何的请求（内存和GC正常）

3. JMH基准测试框架

测试代码：

package org.sample;

import org.openjdk.jmh.annotations.*;
import org.openjdk.jmh.results.format.ResultFormatType;
import org.openjdk.jmh.runner.Runner;
import org.openjdk.jmh.runner.RunnerException;
import org.openjdk.jmh.runner.options.Options;
import org.openjdk.jmh.runner.options.OptionsBuilder;

import java.text.SimpleDateFormat;
import java.time.LocalDateTime;
import java.time.format.DateTimeFormatter;
import java.util.Date;
import java.util.concurrent.TimeUnit;

//执行5轮预热，每次持续1秒
@Warmup(iterations = 5, time = 1, timeUnit = TimeUnit.SECONDS)
//执行一次测试
@Fork(value = 1, jvmArgsAppend = {"-Xms1g", "-Xmx1g"})
//显示平均时间，单位纳秒
@BenchmarkMode(Mode.AverageTime)
// 指定显示结果单位
@OutputTimeUnit(TimeUnit.NANOSECONDS)
// 变量共享范围
@State(Scope.Benchmark)
public class HelloWorldBench {

    @Benchmark
    public int test1() {
        int i = 0;
        i++;
        return i;
    }

    public static void main(String[] args) throws RunnerException {
        Options opt = new OptionsBuilder()
                .include(HelloWorldBench.class.getSimpleName())
                .resultFormat(ResultFormatType.JSON)
                .forks(1)
                .build();

        new Runner(opt).run();
    }
}

如果不将i返回，JIT会直接将这段代码去掉，因为它认为你不会使用i那么我们对i进行的任何处理都是没有意义的，这种代码无法执行的现象称之为死代码

我们可以将i返回，或者添加黑洞来消费这些变量，让JIT无法消除这些代码:

package org.sample;

import org.openjdk.jmh.annotations.*;
import org.openjdk.jmh.results.format.ResultFormatType;
import org.openjdk.jmh.runner.Runner;
import org.openjdk.jmh.runner.RunnerException;
import org.openjdk.jmh.runner.options.Options;
import org.openjdk.jmh.runner.options.OptionsBuilder;

import java.text.SimpleDateFormat;
import java.time.LocalDateTime;
import java.time.format.DateTimeFormatter;
import java.util.Date;
import java.util.concurrent.TimeUnit;

//执行5轮预热，每次持续1秒
@Warmup(iterations = 5, time = 1, timeUnit = TimeUnit.SECONDS)
//执行一次测试
@Fork(value = 1, jvmArgsAppend = {"-Xms1g", "-Xmx1g"})
//显示平均时间，单位纳秒
@BenchmarkMode(Mode.AverageTime)
@OutputTimeUnit(TimeUnit.NANOSECONDS)
@State(Scope.Thread)
public class DateBench {


    private static String sDateFormatString = "yyyy-MM-dd HH:mm:ss";
    private Date date = new Date();
    private LocalDateTime localDateTime = LocalDateTime.now();
    private static ThreadLocal<SimpleDateFormat> simpleDateFormatThreadLocal = new ThreadLocal();
    private static final DateTimeFormatter formatter = DateTimeFormatter.ofPattern("yyyy-MM-dd HH:mm:ss");

    @Setup
    public void setUp() {

        SimpleDateFormat sdf = new SimpleDateFormat(sDateFormatString);
        simpleDateFormatThreadLocal.set(sdf);

    }

    @Benchmark
    public String date() {
        SimpleDateFormat simpleDateFormat = new SimpleDateFormat(sDateFormatString);
        return simpleDateFormat.format(date);
    }

    @Benchmark
    public String localDateTime() {
        return localDateTime.format(formatter);
    }
    @Benchmark
    public String localDateTimeNotSave() {
        return localDateTime.format(DateTimeFormatter.ofPattern("yyyy-MM-dd HH:mm:ss"));
    }

    @Benchmark
    public String dateThreadLocal() {
        return simpleDateFormatThreadLocal.get().format(date);
    }


    public static void main(String[] args) throws RunnerException {
        Options opt = new OptionsBuilder()
                .include(DateBench.class.getSimpleName())
                .resultFormat(ResultFormatType.JSON)
                .forks(1)
                .build();

        new Runner(opt).run();
    }
}

package com.itheima.jvmoptimize.performance.practice.controller;

import com.itheima.jvmoptimize.performance.practice.entity.User;
import com.itheima.jvmoptimize.performance.practice.entity.UserDetails;
import com.itheima.jvmoptimize.performance.practice.service.UserService;
import com.itheima.jvmoptimize.performance.practice.vo.UserVO;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RestController;

import java.text.SimpleDateFormat;
import java.time.format.DateTimeFormatter;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.stream.Collectors;

@RestController
@RequestMapping("/puser")
public class UserController {

    @Autowired
    private UserService userService;

    private final DateTimeFormatter formatter = DateTimeFormatter.ofPattern("yyyy-MM-dd HH:mm:ss");

    //初始代码
    public List<UserVO> user1(){
        //1.从数据库获取前端需要的详情数据
        List<UserDetails> userDetails = userService.getUserDetails();

        //2.获取缓存中的用户数据
        List<User> users = userService.getUsers();

        SimpleDateFormat simpleDateFormat = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
        //3.遍历详情集合，从缓存中获取用户名，生成VO进行填充
        ArrayList<UserVO> userVOS = new ArrayList<>();
        for (UserDetails userDetail : userDetails) {
            UserVO userVO = new UserVO();
            //可以使用BeanUtils对象拷贝
            userVO.setId(userDetail.getId());
            userVO.setRegister(simpleDateFormat.format(userDetail.getRegister2()));
            //填充name
            for (User user : users) {
                if(user.getId().equals(userDetail.getId())){
                    userVO.setName(user.getName());
                }
            }
            //加入集合
            userVOS.add(userVO);
        }

        return userVOS;

    }


    //使用HasmMap存放用户名字
    public List<UserVO> user2(){
        //1.从数据库获取前端需要的详情数据
        List<UserDetails> userDetails = userService.getUserDetails();

        //2.获取缓存中的用户数据
        List<User> users = userService.getUsers();
        //将list转换成hashmap
        HashMap<Long, User> map = new HashMap<>();
        for (User user : users) {
            map.put(user.getId(),user);
        }

        SimpleDateFormat simpleDateFormat = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
        //3.遍历详情集合，从缓存中获取用户名，生成VO进行填充
        ArrayList<UserVO> userVOS = new ArrayList<>();
        for (UserDetails userDetail : userDetails) {
            UserVO userVO = new UserVO();
            //可以使用BeanUtils对象拷贝
            userVO.setId(userDetail.getId());
            userVO.setRegister(simpleDateFormat.format(userDetail.getRegister2()));
            //填充name
            userVO.setName(map.get(userDetail.getId()).getName());
            //加入集合
            userVOS.add(userVO);
        }

        return userVOS;

    }


    //优化日期格式化
    public List<UserVO> user3(){
        //1.从数据库获取前端需要的详情数据
        List<UserDetails> userDetails = userService.getUserDetails();

        //2.获取缓存中的用户数据
        List<User> users = userService.getUsers();
        //将list转换成hashmap
        HashMap<Long, User> map = new HashMap<>();
        for (User user : users) {
            map.put(user.getId(),user);
        }

        //3.遍历详情集合，从缓存中获取用户名，生成VO进行填充
        ArrayList<UserVO> userVOS = new ArrayList<>();
        for (UserDetails userDetail : userDetails) {
            UserVO userVO = new UserVO();
            //可以使用BeanUtils对象拷贝
            userVO.setId(userDetail.getId());
            userVO.setRegister(userDetail.getRegister().format(formatter));
            //填充name
            userVO.setName(map.get(userDetail.getId()).getName());
            //加入集合
            userVOS.add(userVO);
        }

        return userVOS;

    }

    @GetMapping
    //使用stream流改写for循环
    public List<UserVO> user4(){
        //1.从数据库获取前端需要的详情数据
        List<UserDetails> userDetails = userService.getUserDetails();

        //2.获取缓存中的用户数据
        List<User> users = userService.getUsers();
        //将list转换成hashmap
        Map<Long, User> map = users.stream().collect(Collectors.toMap(User::getId, o -> o));

        //3.遍历详情集合，从缓存中获取用户名，生成VO进行填充
        return userDetails.stream().map(userDetail -> {
            UserVO userVO = new UserVO();
            //可以使用BeanUtils对象拷贝
            userVO.setId(userDetail.getId());
            userVO.setRegister(userDetail.getRegister().format(formatter));
            //填充name
            userVO.setName(map.get(userDetail.getId()).getName());
            return userVO;
        }).collect(Collectors.toList());

    }

    //使用并行流优化性能
    public List<UserVO> user5(){
        //1.从数据库获取前端需要的详情数据
        List<UserDetails> userDetails = userService.getUserDetails();

        //2.获取缓存中的用户数据
        List<User> users = userService.getUsers();
        //将list转换成hashmap
        Map<Long, User> map = users.parallelStream().collect(Collectors.toMap(User::getId, o -> o));

        //3.遍历详情集合，从缓存中获取用户名，生成VO进行填充
        return userDetails.parallelStream().map(userDetail -> {
            UserVO userVO = new UserVO();
            //可以使用BeanUtils对象拷贝
            userVO.setId(userDetail.getId());
            userVO.setRegister(userDetail.getRegister().format(formatter));
            //填充name
            userVO.setName(map.get(userDetail.getId()).getName());
            return userVO;
        }).collect(Collectors.toList());

    }
}

在SpringBoot项目中整合JMH:

1、pom文件中添加依赖:

<dependency>
    <groupId>org.openjdk.jmh</groupId>
    <artifactId>jmh-core</artifactId>
    <version>${jmh.version}</version>
    <scope>test</scope>
</dependency>
<dependency>
    <groupId>org.openjdk.jmh</groupId>
    <artifactId>jmh-generator-annprocess</artifactId>
    <version>${jmh.version}</version>
    <scope>test</scope>
</dependency>

<properties>
    <java.version>8</java.version>
    <jmh.version>1.37</jmh.version>
</properties>

2、测试类中编写:

package com.itheima.jvmoptimize;

import com.itheima.jvmoptimize.performance.practice.controller.UserController;
import org.junit.jupiter.api.Test;
import org.openjdk.jmh.annotations.*;
import org.openjdk.jmh.infra.Blackhole;
import org.openjdk.jmh.results.format.ResultFormatType;
import org.openjdk.jmh.runner.Runner;
import org.openjdk.jmh.runner.RunnerException;
import org.openjdk.jmh.runner.options.OptionsBuilder;
import org.springframework.boot.SpringApplication;
import org.springframework.context.ApplicationContext;

import java.io.IOException;
import java.util.concurrent.TimeUnit;

//执行5轮预热，每次持续1秒
@Warmup(iterations = 5, time = 1, timeUnit = TimeUnit.SECONDS)
//执行一次测试
@Fork(value = 1, jvmArgsAppend = {"-Xms1g", "-Xmx1g"})
//显示平均时间，单位纳秒
@BenchmarkMode(Mode.AverageTime)
@OutputTimeUnit(TimeUnit.MILLISECONDS)
@State(Scope.Benchmark)
public class PracticeBenchmarkTest {

    private UserController userController;
    private ApplicationContext context;

    //初始化将springboot容器启动 端口号随机
    @Setup
    public void setup() {
        this.context = new SpringApplication(JvmOptimizeApplication.class).run();
        userController = this.context.getBean(UserController.class);
    }

    //启动这个测试用例进行测试
    @Test
    public void executeJmhRunner() throws RunnerException, IOException {

        new Runner(new OptionsBuilder()
                .shouldDoGC(true)
                .forks(0)
                .resultFormat(ResultFormatType.JSON)
                .shouldFailOnError(true)
                .build()).run();
    }

    //用黑洞消费数据，避免JIT消除代码
    @Benchmark
    public void test1(final Blackhole bh) {

        bh.consume(userController.user1());
    }

    @Benchmark
    public void test2(final Blackhole bh) {

        bh.consume(userController.user2());
    }

    @Benchmark
    public void test3(final Blackhole bh) {

        bh.consume(userController.user3());
    }

    @Benchmark
    public void test4(final Blackhole bh) {

        bh.consume(userController.user4());
    }

    @Benchmark
    public void test5(final Blackhole bh) {

        bh.consume(userController.user5());
    }
}