Java 中怎么获取一份线程 dump 文件
当服务器挂起,崩溃或者性能底下时,就需要抓取服务器的线程堆栈(Thread Dump)用于后续的分析.
Thread dump提供了当前活动的线程的快照. 它提供了JVM中所有Java线程的栈跟踪信息
有很多方式可用于获取Thread Dump, 一些是操作系统特定的命令.
操作系统命令获取ThreadDump:
Windows:
1. 转向服务器的标准输出窗口并按下Control + Break组合键, 之后需要将线程堆栈复制到文件中
UNIX/ Linux
首先查找到服务器的进程号(process id), 然后获取堆栈.
1. ps –ef | grep java
2. kill -3 pid
注意一定要谨慎, 一步不慎就可能让服务器进程被杀死!
JVM 自带的工具获取线程堆栈:
JDK自带命令行工具获取PID并做ThreadDump:
1. jps
2.jstack pid
使用JVisualVM:
Threads 标签页 →ThreadDump按钮
WebLogic 自带的获取 thread dump的工具:
1. webLogic.Admin 工具
a. 打开命令提示符, 通过运行DOMAIN_HOME/bin/setDomain.env设置相关类路径
b. 执行下面的命令
java weblogic.Admin -url t3://www.easyaq.com:7001 -username weblogic -password weblogic1 THREAD_DUMP
注意: Thread Dump 会打印到标准输出, 如nohup日志或者进程窗口.
2. 使用 Admin Console
a. 登录 Admin Console , 点击对应的服务器
b. 点击Server à Monitoring àThreads
c. 点击: Dump Thread Stack 按钮
3. 使用WLST (WebLogic Scripting Tool)
connect(‘weblogic’,’weblogic1’,’t3://www.easyaq.com:7001’)
cd(‘Servers’)
cd(‘AdminServer’)
threadDump()
disconnect()
exit()
注意: 线程堆栈将会保存在运行wlst的当前目录下.
4. 使用utils.ThreadDumper
用法:
C:\bea\wlserver_10.3\server\libjava -cp weblogic.jar utils.ThreadDumper
Broadcast Thread dumps disabled: must specify weblogic.debug.dumpThreadAddr and
weblogic.debug.dumpThreadPort
Exception in thread “main” java.lang.IllegalArgumentException: Port out of range
:-1
at java.net.DatagramPacket.setPort(Unknown Source)
at java.net.DatagramPacket.init(Unknown Source)
at java.net.DatagramPacket.init(Unknown Source)
at utils.ThreadDumper.sendDumpMsg(ThreadDumper.java:124)
at utils.ThreadDumper.main(ThreadDumper.java:145)
5. 如果服务器是作为Windows服务的方式运行, 请运行下列命令:
WL_HOME\bin\beasvc -dump -svcname:service-name
其它一些获取Thread Dump的工具有jrcmd, jrmc(JRockit VM自带) ,Samurai, JProfiler等, 还可通过JMX编程的方式获取, 如JDK自带示例代码:
$JAVA_HOME\demo\management\FullThreadDump
jstack命令详解是什么?
jstack用于打印出给定的java进程ID或corefile或远程调试服务的Java堆栈信息,如果是在64位机器上,需要指定选项”-J-d64″,Windows的jstack使用方式只支持以下的这种方式:
命令格式
jstack[option]pid
jstack[option]executablecore
jstack[option][server-id@]remote-hostname-or-IP
用法结构:
[root@wwwwangxiaoxiao]#jstack
Usage:
jstack[-l]pid
jstack-F[-m][-l]pid
Options:
-F强制dump线程堆栈信息.用于进程hung住,jstackpid命令没有响应的情况
-m同时打印java和本地(native)线程栈信息,m是mixedmode的简写
-l打印锁的额外信息
怎样分析 JAVA 的 Thread Dumps
当看到线程栈的log的时候,第一反应是以下几个thread state
线程状态为 “waiting for monitor entry”
意味着它 在等待进入一个临界区 ,所以它在”Entry Set“队列中等待。
此时线程状态一般都是 Blocked:
java.lang.Thread.State: BLOCKED (on object monitor)
线程状态为“waiting on condition”
说明它在等待另一个条件的发生,来把自己唤醒,或者干脆它是调用了 sleep(N)。
此时线程状态大致为以下几种:
java.lang.Thread.State: WAITING (parking):一直等那个条件发生;
java.lang.Thread.State: TIMED_WAITING (parking或sleeping):定时的,那个条件不到来,也将定时唤醒自己。
如果大量线程在“waiting for monitor entry”
可能是一个全局锁阻塞住了大量线程。
如果短时间内打印的 thread dump 文件反映,随着时间流逝,waiting for monitor entry 的线程越来越多,没有减少的趋势,可能意味着某些线程在临界区里呆的时间太长了,以至于越来越多新线程迟迟无法进入临界区。
如果大量线程在“waiting on condition”
可能是它们又跑去获取第三方资源,尤其是第三方网络资源,迟迟获取不到Response,导致大量线程进入等待状态。
所以如果你发现有大量的线程都处在 Wait on condition,从线程堆栈看,正等待网络读写,这可能是一个网络瓶颈的征兆,因为网络阻塞导致线程无法执行。
线程状态为“in Object.wait()”
说明它获得了监视器之后,又调用了 java.lang.Object.wait() 方法。
每个 Monitor在某个时刻,只能被一个线程拥有,该线程就是 “Active Thread”,而其它线程都是 “Waiting Thread”,分别在两个队列 “ Entry Set”和 “Wait Set”里面等候。在 “Entry Set”中等待的线程状态是 “Waiting for monitor entry”,而在 “Wait Set”中等待的线程状态是 “in Object.wait()”。
当线程获得了 Monitor,如果发现线程继续运行的条件没有满足,它则调用对象(一般就是被 synchronized 的对象)的 wait() 方法,放弃了 Monitor,进入 “Wait Set”队列。
此时线程状态大致为以下几种:
java.lang.Thread.State: TIMED_WAITING (on object monitor);
java.lang.Thread.State: WAITING (on object monitor);
一般都是RMI相关线程(RMI RenewClean、 GC Daemon、RMI Reaper),GC线程(Finalizer),引用对象垃圾回收线程(Reference Handler)等系统线程处于这种状态。
范例
示范一
下面这个线程在等待这个锁 0x00000000fe7e3b50,等待进入临界区:
“RMI TCP Connection(64896)-172.16.52.118” daemon prio=10 tid=0x00000000405a6000 nid=0x68fe waiting for monitor entry [0x00007f2be65a3000]
java.lang.Thread.State: BLOCKED (on object monitor)
at com.xyz.goods.service.impl.GoodsServiceImpl.findChanellGoodsCountWithCache(GoodsServiceImpl.java:1734)
– waiting to lock 0x00000000fe7e3b50 (a java.lang.String)
那么谁持有这个锁呢?
是另一个先调用了 findChanellGoodsCountWithCache 函数的线程:
“RMI TCP Connection(64878)-172.16.52.117” daemon prio=10 tid=0x0000000040822000 nid=0x6841 runnable [0x00007f2be76b3000]
java.lang.Thread.State: RUNNABLE
at java.net.SocketInputStream.socketRead0(Native Method)
at java.net.SocketInputStream.read(SocketInputStream.java:129)
at java.io.BufferedInputStream.fill(BufferedInputStream.java:218)
at java.io.BufferedInputStream.read1(BufferedInputStream.java:258)
at java.io.BufferedInputStream.read(BufferedInputStream.java:317)
– locked 0x00000000af4ed638 (a java.io.BufferedInputStream)
at org.bson.io.Bits.readFully(Bits.java:35)
at org.bson.io.Bits.readFully(Bits.java:28)
at com.mongodb.Response.init(Response.java:35)
at com.mongodb.DBPort.go(DBPort.java:110)
– locked 0x00000000af442d48 (a com.mongodb.DBPort)
at com.mongodb.DBPort.go(DBPort.java:75)
– locked 0x00000000af442d48 (a com.mongodb.DBPort)
at com.mongodb.DBPort.call(DBPort.java:65)
at com.mongodb.DBTCPConnector.call(DBTCPConnector.java:202)
at com.mongodb.DBApiLayer$MyCollection.__find(DBApiLayer.java:296)
at com.mongodb.DB.command(DB.java:152)
at com.mongodb.DBCollection.getCount(DBCollection.java:760)
at com.mongodb.DBCollection.getCount(DBCollection.java:731)
at com.mongodb.DBCollection.count(DBCollection.java:697)
at com.xyz.goods.manager.MongodbManager.count(MongodbManager.java:202)
at com.xyz.goods.service.impl.GoodsServiceImpl.findChanellGoodsCount(GoodsServiceImpl.java:1787)
at com.xyz.goods.service.impl.GoodsServiceImpl.findChanellGoodsCountWithCache(GoodsServiceImpl.java:1739)
– locked 0x00000000fe7e3b50 (a java.lang.String)
示范二
等待另一个条件发生来将自己唤醒:
“RMI TCP Connection(idle)” daemon prio=10 tid=0x00007fd50834e800 nid=0x56b2 waiting on condition [0x00007fd4f1a59000]
java.lang.Thread.State: TIMED_WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
– parking to wait for 0x00000000acd84de8 (a java.util.concurrent.SynchronousQueue$TransferStack)
at java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:198)
at java.util.concurrent.SynchronousQueue$TransferStack.awaitFulfill(SynchronousQueue.java:424)
at java.util.concurrent.SynchronousQueue$TransferStack.transfer(SynchronousQueue.java:323)
at java.util.concurrent.SynchronousQueue.poll(SynchronousQueue.java:874)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:945)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:907)
at java.lang.Thread.run(Thread.java:662)
“TIMED_WAITING (parking)”中的 timed_waiting 指等待状态,但这里指定了时间,到达指定的时间后自动退出等待状态;parking指线程处于挂起中。
“waiting on condition”需要与堆栈中的“parking to wait for 0x00000000acd84de8 (a java.util.concurrent.SynchronousQueue$TransferStack)” 结合来看。首先,本线程肯定是在等待某个条件的发生,来把自己唤醒。其次,SynchronousQueue 并不是一个队列,只是线程之间移交信息的机制,当我们把一个元素放入到 SynchronousQueue 中时必须有另一个线程正在等待接受移交的任务,因此这就是本线程在等待的条件。
示范三
“RMI RenewClean-[172.16.50.182:4888]” daemon prio=10 tid=0x0000000040d2c800 nid=0x97e in Object.wait() [0x00007f9ccafd0000]
java.lang.Thread.State: TIMED_WAITING (on object monitor)
at java.lang.Object.wait(Native Method)
– waiting on 0x0000000799b032d8 (a java.lang.ref.ReferenceQueue$Lock)
at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:118)
– locked 0x0000000799b032d8 (a java.lang.ref.ReferenceQueue$Lock)
at sun.rmi.transport.DGCClient$EndpointEntry$RenewCleanThread.run(DGCClient.java:516)
at java.lang.Thread.run(Thread.java:662)
怎样使用jstack诊断Java应用程序故障
1. 首先来了解一下jstack这个命令的作用,jstack 是一个可以返回在应用程序上运行的各种各样线程的一个完整转储的实用程序,您可以使用它查明问题。jstack [-l] pid,jpid可以通过使用jps命令来查看当前Java程序的jpid值,-l是可选参数,它可以显示线程阻塞/死锁情况。
/**
* 死锁例子
* @author crane.ding
* @since 2011-3-20
*/
public class DeadLock {
public static void main(String[] args) {
final Object obj_1 = new Object(), obj_2 = new Object();
Thread t1 = new Thread(“t1”){
@Override
public void run() {
synchronized (obj_1) {
try {
Thread.sleep(3000);
} catch (InterruptedException e) {}
synchronized (obj_2) {
System.out.println(“thread t1 done.”);
}
}
}
};
Thread t2 = new Thread(“t2”){
@Override
public void run() {
synchronized (obj_2) {
try {
Thread.sleep(3000);
} catch (InterruptedException e) {}
synchronized (obj_1) {
System.out.println(“thread t2 done.”);
}
}
}
};
t1.start();
t2.start();
}
}
2. 以上DeadLock类是一个死锁的例子,假使在不知情的情况下,运行DeadLock后,发现等了N久都没有在屏幕打印线程完成信息。这个时候就可以使用jps查看该程序的jpid值和使用jstack来生产堆栈结果问题。
$ java -cp deadlock.jar DeadLock
$
$ jps
3076 Jps
448 DeadLock
$ jstack -l 448 deadlock.jstack
结果文件deadlock.jstack内容如下:
2011-03-20 23:05:20
Full thread dump Java HotSpot(TM) Client VM (19.1-b02 mixed mode, sharing):
“DestroyJavaVM” prio=6 tid=0x00316800 nid=0x9fc waiting on condition [0x00000000]
java.lang.Thread.State: RUNNABLE
Locked ownable synchronizers:
– None
“t2” prio=6 tid=0x02bcf000 nid=0xc70 waiting for monitor entry [0x02f6f000]
java.lang.Thread.State: BLOCKED (on object monitor)
at com.demo.DeadLock$2.run(DeadLock.java:40)
– waiting to lock 0x22a297a8 (a java.lang.Object)
– locked 0x22a297b0 (a java.lang.Object)
Locked ownable synchronizers:
– None
“t1” prio=6 tid=0x02bce400 nid=0xba0 waiting for monitor entry [0x02f1f000]
java.lang.Thread.State: BLOCKED (on object monitor)
at com.demo.DeadLock$1.run(DeadLock.java:25)
– waiting to lock 0x22a297b0 (a java.lang.Object)
– locked 0x22a297a8 (a java.lang.Object)
Locked ownable synchronizers:
– None
“Low Memory Detector” daemon prio=6 tid=0x02bb9400 nid=0xa6c runnable [0x00000000]
java.lang.Thread.State: RUNNABLE
Locked ownable synchronizers:
– None
“CompilerThread0” daemon prio=10 tid=0x02bb2800 nid=0xcb8 waiting on condition [0x00000000]
java.lang.Thread.State: RUNNABLE
Locked ownable synchronizers:
– None
“Attach Listener” daemon prio=10 tid=0x02bb1000 nid=0x7f4 waiting on condition [0x00000000]
java.lang.Thread.State: RUNNABLE
Locked ownable synchronizers:
– None
“Signal Dispatcher” daemon prio=10 tid=0x02bd2800 nid=0xd80 runnable [0x00000000]
java.lang.Thread.State: RUNNABLE
Locked ownable synchronizers:
– None
“Finalizer” daemon prio=8 tid=0x02bab000 nid=0xe1c in Object.wait() [0x02d3f000]
java.lang.Thread.State: WAITING (on object monitor)
at java.lang.Object.wait(Native Method)
– waiting on 0x229e1148 (a java.lang.ref.ReferenceQueue$Lock)
at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:118)
– locked 0x229e1148 (a java.lang.ref.ReferenceQueue$Lock)
at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:134)
at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:159)
Locked ownable synchronizers:
– None
“Reference Handler” daemon prio=10 tid=0x02ba6800 nid=0xbe0 in Object.wait() [0x02cef000]
java.lang.Thread.State: WAITING (on object monitor)
at java.lang.Object.wait(Native Method)
– waiting on 0x229e1048 (a java.lang.ref.Reference$Lock)
at java.lang.Object.wait(Object.java:485)
at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:116)
– locked 0x229e1048 (a java.lang.ref.Reference$Lock)
Locked ownable synchronizers:
– None
“VM Thread” prio=10 tid=0x02b6a400 nid=0x568 runnable
“VM Periodic Task Thread” prio=10 tid=0x02bc8400 nid=0x75c waiting on condition
JNI global references: 878
Found one Java-level deadlock:
=============================
“t2”:
waiting to lock monitor 0x02baaeec (object 0x22a297a8, a java.lang.Object),
which is held by “t1”
“t1”:
waiting to lock monitor 0x02baa2bc (object 0x22a297b0, a java.lang.Object),
which is held by “t2”
Java stack information for the threads listed above:
===================================================
“t2”:
at com.demo.DeadLock$2.run(DeadLock.java:40)
– waiting to lock 0x22a297a8 (a java.lang.Object)
– locked 0x22a297b0 (a java.lang.Object)
“t1”:
at com.demo.DeadLock$1.run(DeadLock.java:25)
– waiting to lock 0x22a297b0 (a java.lang.Object)
– locked 0x22a297a8 (a java.lang.Object)
Found 1 deadlock.
3.从这个结果文件看到发现了一个死锁,具体是线程t2在等待线程t1,而线程t1在等待线程t2造成的,同时也记录了线程的堆栈和代码行数,通过这个堆栈和行数就可以去检查对应的代码块,从而发现问题和解决问题。