下面的java代码通过提供一个远程进程的PID来获取java进程的堆转储。该程序使用远程JMX连接转储堆到一个文件。它可能对某人有帮助。不需要jmap。

import java.lang.management.ManagementFactory;
import javax.management.MBeanServerConnection;
import javax.management.remote.JMXConnector;
import javax.management.remote.JMXConnectorFactory;
import javax.management.remote.JMXServiceURL;
import java.lang.reflect.Method;

public class HeapDumper {

public static final String HOST = "192.168.11.177";
public static final String PORT = "1600";
public static final String FILE_NAME = "heapDump.hprof";
public static final String FOLDER_PATH = "C:/";
private static final String HOTSPOT_BEAN_NAME ="com.sun.management:type=HotSpotDiagnostic";

public static void main(String[] args) {
    if(args.length == 0) {
        System.out.println("Enter PID of the Java Process !!!");
        return;
    }
    
    String pidString = args[0];
    int pid = -1;
    if(pidString!=null && pidString.length() > 0) {
        try {
            pid = Integer.parseInt(pidString);
        }
        catch(Exception e) {
            System.out.println("PID is not Valid !!!");
            return;
        }
    }
    boolean isHeapDumpSuccess = false;
    boolean live = true;
    if(pid > 0) {
        MBeanServerConnection beanServerConn = getJMXConnection();
        
        if(beanServerConn!=null) {
            Class clazz = null;
            String dumpFile = FOLDER_PATH+"/"+FILE_NAME;
            try{
                clazz = Class.forName("com.sun.management.HotSpotDiagnosticMXBean");
                Object hotspotMBean = ManagementFactory.newPlatformMXBeanProxy(beanServerConn, HOTSPOT_BEAN_NAME, clazz);
                Method method = clazz.getMethod("dumpHeap", new Class[]{String.class , boolean.class});
                method.setAccessible(true);
                method.invoke(hotspotMBean , new Object[] {dumpFile, new Boolean(live)});
                isHeapDumpSuccess = true;
            }
            catch(Exception e){
                e.printStackTrace();
                isHeapDumpSuccess = false;
            }
            finally{
                clazz = null;
            }
        }
    }
    
    if(isHeapDumpSuccess){
        System.out.println("HeapDump is Success !!!");
    }
    else{
        System.out.println("HeapDump is not Success !!!");
    }
}

private static MBeanServerConnection getJMXConnection() {
    MBeanServerConnection mbeanServerConnection = null;
    String urlString = "service:jmx:rmi:///jndi/rmi://" + HOST + ":" + PORT + "/jmxrmi";
    try {
        JMXServiceURL url = new JMXServiceURL(urlString);
        JMXConnector jmxConnector = JMXConnectorFactory.connect(url);
        mbeanServerConnection = jmxConnector.getMBeanServerConnection();
        System.out.println("JMX Connection is Success for the URL :"+urlString);
    }
    catch(Exception e) {
        System.out.println("JMX Connection Failed !!!");
    }
    return mbeanServerConnection;
}

}

如果由于某些原因不能(或不想)使用控制台/终端，还有另一种解决方案。您可以让Java应用程序为您打印线程转储。收集堆栈跟踪的代码相当简单，可以附加到按钮或web界面。

private static String getThreadDump() {
    Map<Thread, StackTraceElement[]> allStackTraces = Thread.getAllStackTraces();

    StringBuilder out = new StringBuilder();
    for (Map.Entry<Thread, StackTraceElement[]> entry : allStackTraces.entrySet()) {
        Thread thread = entry.getKey();
        StackTraceElement[] elements = entry.getValue();
        out.append(String.format("%s | prio=%d | %s", thread.getName(), thread.getPriority(), thread.getState()));
        out.append('\n');

        for (StackTraceElement element : elements) {
            out.append(element.toString()).append('\n');
        }
        out.append('\n');
    }
    return out.toString();
}

这个方法将返回一个像这样的字符串:

main | prio=5 | RUNNABLE
java.lang.Thread.dumpThreads(Native Method)
java.lang.Thread.getAllStackTraces(Thread.java:1607)
Main.getThreadDump(Main.java:8)
Main.main(Main.java:36)

Monitor Ctrl-Break | prio=5 | RUNNABLE
java.net.PlainSocketImpl.initProto(Native Method)
java.net.PlainSocketImpl.<clinit>(PlainSocketImpl.java:45)
java.net.Socket.setImpl(Socket.java:503)
java.net.Socket.<init>(Socket.java:424)
java.net.Socket.<init>(Socket.java:211)
com.intellij.rt.execution.application.AppMainV2$1.run(AppMainV2.java:59)

Finalizer | prio=8 | WAITING
java.lang.Object.wait(Native Method)
java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:143)
java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:164)
java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:209)

Reference Handler | prio=10 | WAITING
java.lang.Object.wait(Native Method)
java.lang.Object.wait(Object.java:502)
java.lang.ref.Reference.tryHandlePending(Reference.java:191)
java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)

对于那些对带有流的Java 8版本感兴趣的人来说，代码甚至更加紧凑:

private static String getThreadDump() {
    Map<Thread, StackTraceElement[]> allStackTraces = Thread.getAllStackTraces();
    StringBuilder out = new StringBuilder();
    allStackTraces.forEach((thread, elements) -> {
        out.append(String.format("%s | prio=%d | %s", thread.getName(), thread.getPriority(), thread.getState()));
        out.append('\n');

        Arrays.stream(elements).forEach(element -> out.append(element.toString()).append('\n'));
        out.append('\n');
    });
    return out.toString();
}

你可以很容易地测试这段代码:

System.out.print(getThreadDump());

我认为在Linux进程中创建.hprof文件的最好方法是使用jmap命令。例如:jmap -dump:format=b,file=filename。hprof {PID}

您混淆了两个不同的java转储。Kill -3生成线程转储，而不是堆转储。

线程转储= JVM输出到stdout的每个线程的堆栈跟踪作为文本。 堆转储= JVM进程输出到二进制文件的内存内容。

要在Windows上进行线程转储，如果您的JVM是前台进程，CTRL+BREAK是最简单的方法。如果你在Windows上有一个类似Unix的shell，比如Cygwin或MobaXterm，你可以像在Unix中一样使用kill -3 {pid}。

要在Unix中进行线程转储，如果您的JVM是前台进程，则可以使用CTRL+C或kill -3 {pid}，只要您为JVM获得正确的pid即可。

对于这两种平台，Java都提供了一些有用的实用程序。对于线程转储，jstack {pid}是最好的选择。http://docs.oracle.com/javase/1.5.0/docs/tooldocs/share/jstack.html

Just to finish the dump question out: Heap dumps are not commonly used because they are difficult to interpret. But, they have a lot of useful information in them if you know where/how to look at them. The most common usage is to locate memory leaks. It is a good practice to set the -D on the java command-line so that the heap dump is generated automatically upon an OutOfMemoryError, -XX:+HeapDumpOnOutOfMemoryError But, you can manually trigger a heap dump, also. The most common way is to use the java utility jmap.

注意:此实用程序并非在所有平台上都可用。从JDK 1.6开始，jmap可以在Windows上使用。

命令行示例如下所示

jmap -dump:file=myheap.bin {pid of the JVM}

输出“myheap.bin”不是人类可读的(对我们大多数人来说)，您需要一个工具来分析它。我的首选是MAT. http://www.eclipse.org/mat/

你可以从Cygwin发送kill -3 <pid>。你必须使用Cygwin ps选项来查找windows进程，然后将信号发送到该进程。

如果你在服务器-jre 8及以上，你可以使用这个:

jcmd PID GC.heap_dump /tmp/dump