jvm高级特性第4章-虚拟机性能监控与故障处理工具

【4.2】jdk命令行工具

1、jps：虚拟机进程状况工具；

C:\Users\pacoson>jps -l // 输出主类名称
12272 sun.tools.jps.Jps
12736 chapter3.Page93
2808C:\Users\pacoson>jps -v // 查看虚拟机进程启动时的参数
12736 Page93 -verbose:gc -Xms20M -Xmx20M -Xmn10M -XX:+PrintGCDetails -XX:SurvivorRatio=8 -Dfile.encoding=UTF-8
2808  -Dosgi.requiredJavaVersion=1.6 -Xms40m -Xmx512m -Dgrails.console.enable.interactive=false -Dgrails.console.enable.terminal=false -Djline.terminal=jline.UnsupportedTerminal -Dgrails.console.class=grails.build.logging.GrailsEclipseConsole
6072 Jps -Dapplication.home=D:\Java\jdk1.8.0_172 -Xms8mC:\Users\pacoson>jps // 列出正在运行的虚拟机进程 
12736 Page93
2808
7992 Jps

补充：

jps -q 仅输出本地虚拟机唯一id， LVMID；

jps -m 输出虚拟机启动时传递给主类的main函数的参数；

jps可以通过rmi 协议查看开启了rmi服务的远程虚拟机进程状态；

2、jstat：虚拟机统计信息监视工具（java statistics monitoring tool）

C:\Users\pacoson>jstat -gcutil 12736 1000 20 // 每隔1000毫秒查看虚拟机12736的内存使用率，共查看20次S0     S1     E      O      M     CCS    YGC     YGCT    FGC    FGCT     GCT0.00   0.00  52.00  64.89  53.87  54.43      6    0.015     8    0.405    0.4210.00   0.00  52.00  64.89  53.87  54.43      6    0.015     8    0.405    0.4210.00   0.00  52.00  64.89  53.87  54.43      6    0.015     8    0.405    0.4210.00   0.00  52.00  64.89  53.87  54.43      6    0.015     8    0.405    0.421

S0=Survivor0， S=Suivivor1， E=Eden区域， O=Old区域， YGC=Young GC=收集年轻代次数=6， YGCT=0.015秒（年轻代收集耗时）， FGC=Full GC=老年代收集次数=8， FGCT=Full GC Time=老年代收集耗时=0.421；

3、jinfo：java配置信息工具：实时查看和调整虚拟机各项参数；

C:\Users\pacoson>jps // 查看本地java虚拟机id
12736 Page93
11496 Jps
2808C:\Users\pacoson>jinfo -flags //查看虚拟机参数 
Usage:jinfo [option] <pid>(to connect to running process)jinfo [option] <executable <core>(to connect to a core file)jinfo [option] [server_id@]<remote server IP or hostname>(to connect to remote debug server)where <option> is one of:-flag <name>         to print the value of the named VM flag-flag [+|-]<name>    to enable or disable the named VM flag // +启用/-停用-flag <name>=<value> to set the named VM flag to the given value // 设置参数值-flags               to print VM flags-sysprops            to print Java system properties<no option>          to print both of the above-h | -help           to print this help messageC:\Users\pacoson>
C:\Users\pacoson>jinfo -flags 12736
Attaching to process ID 12736, please wait...
Debugger attached successfully.
Server compiler detected.
JVM version is 25.172-b11
Non-default VM flags: -XX:CICompilerCount=4 -XX:InitialHeapSize=20971520 -XX:MaxHeapSize=20971520 -XX:MaxNewSize=10485760 -XX:MinHeapDeltaBytes=524288 -XX:NewSize=10485760 -XX:OldSize=10485760 -XX:+PrintGC -XX:+PrintGCDetails -XX:SurvivorRatio=8 -XX:+UseCompressedClassPointers -XX:+UseCompressedOops -XX:+UseFastUnorderedTimeStamps -XX:-UseLargePagesIndividualAllocation -XX:+UseParallelGC
Command line:  -verbose:gc -Xms20M -Xmx20M -Xmn10M -XX:+PrintGCDetails -XX:SurvivorRatio=8 -Dfile.encoding=UTF-8

4、jmap：java内存映像工具，用于生成堆转储快照文件，一般称为heapdump 或 dump文件（堆转储快照文件）；

使用-XX:+HeapDumpOnOutOfMemoryError参数，可以让jvm在oom异常出现之后自动生成dump文件；

jmap在windows平台下的功能是受限的；

jmap和jhat连用以查看虚拟机堆转储快照如下：

5、jhat：虚拟机堆转储快照分析工具；jvm Heap Analysis Tool，与jmap搭配使用；（不推荐使用jhat命令分析dump文件）

推荐的分析堆转储快照工具有：VisualVM， Eclipse Memory Analyzer， IBM HeapAnalyzer等工具；

C:\Users\pacoson>jmap -dump:format=b,file=eclipse.bin 12736  // 生成jvm12736的堆转储快照文件
Dumping heap to C:\Users\pacoson\eclipse.bin ...
File existsC:\Users\pacoson>
C:\Users\pacoson>jhat eclipse.bin // 启用内嵌web服务器分析eclpse.bin 文件（访问localhost:7000）
Reading from eclipse.bin...
Dump file created Tue May 07 23:43:59 CST 2019
Snapshot read, resolving...
Resolving 7066 objects...
Chasing references, expect 1 dots.
Eliminating duplicate references.
Snapshot resolved.
Started HTTP server on port 7000
Server is ready.

6、jstack：java堆栈跟踪工具

jstack=stack trace for java，用于生成虚拟机当前时刻的线程快照（一般称为threaddump文件或 javacore文件）；

线程快照就是当前虚拟机内每一条线程正在执行的方法堆栈集合，生成线程快照的主要目的是定位线程出现长时间停顿的原因，如线程死锁，死循环，请求外部资源超时等；

C:\Users\pacoson>jstack -l 12736  // 查看12736进程堆栈信息
2019-05-12 20:55:41
Full thread dump Java HotSpot(TM) 64-Bit Server VM (25.172-b11 mixed mode):"Service Thread" #10 daemon prio=9 os_prio=0 tid=0x000000001557f000 nid=0x794 runnable [0x0000000000000000]java.lang.Thread.State: RUNNABLELocked ownable synchronizers:- None"C1 CompilerThread3" #9 daemon prio=9 os_prio=2 tid=0x0000000015505000 nid=0x36e8 waiting on condition [0x0000000000000000]java.lang.Thread.State: RUNNABLELocked ownable synchronizers:- None"C2 CompilerThread2" #8 daemon prio=9 os_prio=2 tid=0x00000000154fe000 nid=0x3ab4 waiting on condition [0x0000000000000000]java.lang.Thread.State: RUNNABLELocked ownable synchronizers:- None"C2 CompilerThread1" #7 daemon prio=9 os_prio=2 tid=0x00000000154f1000 nid=0x32b4 waiting on condition [0x0000000000000000]java.lang.Thread.State: RUNNABLELocked ownable synchronizers:- None"C2 CompilerThread0" #6 daemon prio=9 os_prio=2 tid=0x0000000015499000 nid=0x16e4 waiting on condition [0x0000000000000000]java.lang.Thread.State: RUNNABLELocked ownable synchronizers:- None"Attach Listener" #5 daemon prio=5 os_prio=2 tid=0x0000000015498000 nid=0x578 waiting on condition [0x0000000000000000]java.lang.Thread.State: RUNNABLELocked ownable synchronizers:- None"Signal Dispatcher" #4 daemon prio=9 os_prio=2 tid=0x00000000154ef800 nid=0x5b0 runnable [0x0000000000000000]java.lang.Thread.State: RUNNABLELocked ownable synchronizers:- None"Finalizer" #3 daemon prio=8 os_prio=1 tid=0x0000000015480800 nid=0x2c3c in Object.wait() [0x000000001595f000]java.lang.Thread.State: WAITING (on object monitor)at java.lang.Object.wait(Native Method)- waiting on <0x00000000fec08a78> (a java.lang.ref.ReferenceQueue$Lock)at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:144)- locked <0x00000000fec08a78> (a java.lang.ref.ReferenceQueue$Lock)at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:165)at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:216)Locked ownable synchronizers:- None"Reference Handler" #2 daemon prio=10 os_prio=2 tid=0x000000001357d000 nid=0x103c in Object.wait() [0x000000001545f000]java.lang.Thread.State: WAITING (on object monitor)at java.lang.Object.wait(Native Method)- waiting on <0x00000000fec01158> (a java.lang.ref.Reference$Lock)at java.lang.Object.wait(Object.java:502)at java.lang.ref.Reference.tryHandlePending(Reference.java:191)- locked <0x00000000fec01158> (a java.lang.ref.Reference$Lock)at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)Locked ownable synchronizers:- None"main" #1 prio=5 os_prio=0 tid=0x0000000002f2d800 nid=0x780 waiting on condition [0x0000000002a8f000]java.lang.Thread.State: TIMED_WAITING (sleeping)at java.lang.Thread.sleep(Native Method)at chapter3.Page93.minorGC(Page93.java:24)at chapter3.Page93.main(Page93.java:10)Locked ownable synchronizers:- None"VM Thread" os_prio=2 tid=0x0000000013578000 nid=0x3350 runnable"GC task thread#0 (ParallelGC)" os_prio=0 tid=0x0000000002df7800 nid=0x3958 runnable"GC task thread#1 (ParallelGC)" os_prio=0 tid=0x0000000002df9000 nid=0x4a7c runnable"GC task thread#2 (ParallelGC)" os_prio=0 tid=0x0000000002dfa800 nid=0xcdc runnable"GC task thread#3 (ParallelGC)" os_prio=0 tid=0x0000000002dfc000 nid=0x3a2c runnable"GC task thread#4 (ParallelGC)" os_prio=0 tid=0x0000000002dfe800 nid=0x3b30 runnable"GC task thread#5 (ParallelGC)" os_prio=0 tid=0x0000000002dff800 nid=0x3a0 runnable"GC task thread#6 (ParallelGC)" os_prio=0 tid=0x0000000002e02800 nid=0x2a0 runnable"GC task thread#7 (ParallelGC)" os_prio=0 tid=0x0000000002e05000 nid=0x1ed4 runnable"VM Periodic Task Thread" os_prio=2 tid=0x00000000155ae000 nid=0x35e8 waiting on conditionJNI global references: 5

可以通过调用 Thread.getAllStackTraces() 方法完成 jstack的大部分功能；

【荔枝】

/*** Thread.getAllStackTraces获取虚拟机中所有线程的StackTraceElement对象* , 模拟jstack的大部分功能   */
public class Page111 {public static void main(String[] args) {for (int i = 0; i < 5; i++) {new Thread(new Runnable() {@Overridepublic void run() {System.out.println(new Random(17).nextInt(10)); } }).start(); // // 启动线程 }for (int i = 0; i < 3600; i++) {System.out.println("================");try {Thread.sleep(1000);/* Thread.getAllStackTraces获取虚拟机中所有线程的StackTraceElement对象  */Map<Thread, StackTraceElement[]> stackTracks = Thread.getAllStackTraces();for (Map.Entry<Thread, StackTraceElement[]> entry : stackTracks.entrySet()) {System.out.println("key = " + entry.getKey().getName() + ", value = " + entry.getValue());StackTraceElement[] elements = entry.getValue();int index = 1;for (StackTraceElement e : elements) {System.out.println("value[" + index++ +"] = " + e);}}} catch (InterruptedException e) {e.printStackTrace();}}}
}

// Thread.getAllStackTraces获取虚拟机中所有线程的StackTraceElement对象，线程方法堆栈信息
key = Finalizer, value = [Ljava.lang.StackTraceElement;@5c647e05
value[1] = java.lang.Object.wait(Native Method)
value[2] = java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:144)
value[3] = java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:165)
value[4] = java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:216)
key = Attach Listener, value = [Ljava.lang.StackTraceElement;@33909752
key = Signal Dispatcher, value = [Ljava.lang.StackTraceElement;@55f96302
key = Reference Handler, value = [Ljava.lang.StackTraceElement;@3d4eac69
value[1] = java.lang.Object.wait(Native Method)
value[2] = java.lang.Object.wait(Object.java:502)
value[3] = java.lang.ref.Reference.tryHandlePending(Reference.java:191)
value[4] = java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)
key = main, value = [Ljava.lang.StackTraceElement;@42a57993
value[1] = java.lang.Thread.dumpThreads(Native Method)
value[2] = java.lang.Thread.getAllStackTraces(Thread.java:1610)
value[3] = chapter4.Page111.main(Page111.java:25)

7、HSDIS：JIT生成代码反汇编

HSDIS是一个sun官方推荐的hotspot虚拟机 JIT编译代码的反汇编插件，包含在hotspot虚拟机的源码之中，但没有提供编译后的程序。

作用是让hotspot的-XX:+PrintAssembly 指令调用它来吧动态生成的本地代码还原为汇编代码输出，同时还生成了大量有价值的注释；

【4.3】JDK的可视化工具

JConsole和VisualVM；JConsole是虚拟机监控工具，而VisualVM是多合一故障处理工具；

1、JConsole：java监视与管理控制台；基于JMX的可视化监视，管理工具；

1.1、启动JConsole；

【JConsole监控荔枝】

/*** JConsole 监控jvm内存  * vm params: -verbose:gc -Xms20M -Xmx20M -Xmn10M -XX:+PrintGCDetails -XX:SurvivorRatio=8*/
public class Page114JConsoleTest {private static final int _1MB = 1024 * 1024;public static void main(String[] args) {for (int i = 0; i < 3600; i++) {new Thread(new Runnable() {@Overridepublic void run() {byte[] allocation1 = new byte[1 * _1MB]; byte[] allocation2 = new byte[2 * _1MB];byte[] allocation4 = new byte[4 * _1MB];} }).start(); // // 启动线程try {Thread.sleep(1000);} catch (InterruptedException e) {e.printStackTrace(); }}}
}

1.2、内存监控

/*** JConsole监视代码 * jvm params: -Xms100m -Xmx100m -XX:+UseSerialGC* @author tang rong* 2019/05/12 */
public class Page117 {public static void main(String[] args) throws Exception {fillHeap(1000);}public static void fillHeap(int num) throws InterruptedException {List<OOMObject> list = new ArrayList<>(); for (int i=0; i<num; i++) {System.out.println(i);/* 稍作延时，令监视曲线的变化更加明显 */Thread.sleep(50);list.add(new OOMObject());}System.gc();  }static class OOMObject {byte[] placeholder = new byte[64 * 1024];} 
}

1.3、线程监控

线程长时间停顿的主要原因有：等待外部资源（数据库连接，网络资源，设备资源等），死循环，锁等待（活锁和死锁）；

【荔枝-线程等待演示代码】

/*** 线程等待演示代码* @author tr* 2019/05/12 */
public class Page119 {public static void createBusyThread() {new Thread(new Runnable() {@Overridepublic void run() {while (true) // 死循环;}}, "testBusyThread").start();}public static void createLockThread(final Object lock) {new Thread(new Runnable() {@Overridepublic void run() {synchronized (lock) {try {lock.wait(); // 线程等待 } catch (InterruptedException e) {e.printStackTrace();}}}}, "testLockThread").start(); }public static void main(String[] args) throws Exception {BufferedReader br = new BufferedReader(new InputStreamReader(System.in));br.readLine();createBusyThread(); // 死循环br.readLine();Object obj = new Object(); createLockThread(obj); // 线程等待}
}

【荔枝-线程死锁等待演示代码】

/*** 线程死锁等待演示 * @author tr* 2019/05/12 */
public class SynAddRunnable implements Runnable {int a, b; public SynAddRunnable(int a, int b) {this.a = a;this.b = b; }@Override public void run() {synchronized (Integer.valueOf(a)) {synchronized (Integer.valueOf(b)) {System.out.println(a+b);}}}public static void main(String[] args) {for (int i=0; i<100; i++) {System.out.println("i = " + i);new Thread(new SynAddRunnable(1, 2)).start(); // 死锁， 线程1拿到了a锁，准备拿b锁，但线程2拿到了b锁且没有释放；new Thread(new SynAddRunnable(2, 1)).start(); // 死锁，线程2拿到了b锁，准备拿a锁，但线程1拿到了a锁且没有释放；} }
}

出现死锁后，查看线程列表中被阻塞的线程；

【4.3.2】VisualVM：多合一故障处理工具（jvisualvm.exe）

1、VisualVM可以做到：

显示jvm进程配置信息，环境信息（jps， jinfo）；

监视cpu， gc，堆，方法区以及线程信息（jstat, jstack）；

dump以及分析堆转储快照（jmap，jhat）；

方法级的程序运行性能分析，找出被调用最多，运行时间最长的方法；

离线程序快照：收集程序的运行时配置，线程dump，内存dump等信息建立一个快照，可以将快照发送给开发者；

其他plugin的无限可能性；

【补充】VisualVM的兼容性

如果不给VisualVM安装插件，就放弃了其最精华的功能；

2、生成、浏览堆转储快照-dump文件

点击堆dump；

3、分析程序性能：profiler页签中分析；

注意： profiling堆程序运行性能有很大影响，一般不在生产环境做；

4、BTrace动态日志跟踪

4.1、BTrace的作用是：在不停止目标程序运行的前提下，通过hotspot虚拟机的hotswap技术动态加入原本并不存在的调试代码（如日志信息）；

补充：hotswap技术，即代码热替换技术， hotspot虚拟机允许在不停止运行的情况下，更新已经加载的类的代码；

/*** BTrace 跟踪演示 * @date 2019/05/13*/
public class BTraceTest {public int add(int a, int b) {return a + b; }public static void main(String[] args) throws IOException {BTraceTest test = new BTraceTest();BufferedReader reader = new BufferedReader(new InputStreamReader(System.in));for (int i = 0; i < 10; i++) {                                                                                                                                                                                                                                                                                                                reader.readLine();int a = (int) Math.round(Math.random() * 1000);int b = (int) Math.round(Math.random() * 1000);System.out.println("test.add(a, b) = " + test.add(a, b));}}
}

trace script

/* BTrace Script Template */
import com.sun.btrace.annotations.*;
import static com.sun.btrace.BTraceUtils.*;@BTrace
public class TracingScript {/* put your code here */@OnMethod(clazz = "chapter4.BTraceTest",method = "add",location = @Location(Kind.RETURN))public static void func(@Self chapter4.BTraceTest instance,int a,int b,@Return int result){println("调用堆栈:");jstack();println(strcat("方法参数A:",str(a)));println(strcat("方法参数B:",str(b)));println(strcat("方法结果：",str(result)));    }
}

BTrace的用法包括：打印调用堆栈，参数，返回值只是最基本的应用，还可以进行性能监视，定位连接泄漏和内存泄漏，解决多线程竞争问题等；