为什么阿里不允许用Executors创建线程池
共 10275字,需浏览 21分钟
·
2020-08-25 19:43
1 文章概述
《阿里巴巴JAVA开发手册》有这样一条强制规定:线程池不允许使用Executors去创建,而应该通过ThreadPoolExecutor方式,这样处理方式更加明确线程池运行规则,规避资源耗尽风险。本文我们从资源和排查问题两个角度进行分析,同时参考DUBBO线程池声明方式创建一个符合规范的线程池。
2 资源角度
《阿里巴巴JAVA开发手册》从资源角度对这个问题进行了分析
FixedThreadPool SingleThreadPool
允许请求队列长度为Integer.MAX_VALUE可能会堆积大量请求从而导致OOM
CachedThreadPool ScheduledThreadPool
允许创建线程数量为Integer.MAX_VALUE可能会创建大量线程从而导致OOM
以下两个线程池使用链表实现的阻塞队列,不设大小理论上队列容量无上限,所以可能会堆积大量请求从而导致OOM
public static ExecutorService newFixedThreadPool(int nThreads) {
return new ThreadPoolExecutor(nThreads, nThreads,
0L, TimeUnit.MILLISECONDS,
new LinkedBlockingQueue
()); }
public static ExecutorService newSingleThreadExecutor() {
return new FinalizableDelegatedExecutorService
(new ThreadPoolExecutor(1, 1,
0L, TimeUnit.MILLISECONDS,
new LinkedBlockingQueue
())); }
以下两个线程池maxSize使用Integer最大值,所以可能会创建大量线程从而导致OOM
public static ExecutorService newCachedThreadPool() {
return new ThreadPoolExecutor(0, Integer.MAX_VALUE,
60L, TimeUnit.SECONDS,
new SynchronousQueue
()); }
public static ScheduledExecutorService newSingleThreadScheduledExecutor() {
return new DelegatedScheduledExecutorService
(new ScheduledThreadPoolExecutor(1));
}
public ScheduledThreadPoolExecutor(int corePoolSize) {
super(corePoolSize, Integer.MAX_VALUE, 0, NANOSECONDS,
new DelayedWorkQueue());
}
3 排查问题角度
如果使用Executors创建线程池,大家应该最常使用如下语句
public void testThread() throws Exception {
ExecutorService fixedExecutor = Executors.newFixedThreadPool(10);
for (int i = 0; i < 1000; i++) {
fixedExecutor.execute(new Runnable() {
public void run() {
System.out.println("公众号互联网公园");
}
});
}
}
上述语句在功能层面是没有问题的,但是在生产环境中有可能遇到CPU飙高,线程数持续增加,内存溢出等问题,我们时常需要通过线程快照进行观察。我们通过jstack命令观察上述代码线程快照
#525 prio=5 os_prio=0 tid=0x00006f6561039100 nid=0xdaa waiting on condition [0x00006f64e646d000]
WAITING (parking) :
at sun.misc.Unsafe.park(Native Method)
parking to wait for <0x00000006e6f3e230> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java:165)
at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1066)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1126)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:616)
at java.lang.Thread.run(Thread.java:645)
我们发现从线程快照看不出任何业务信息,只有类似pool-1-thread-2这种编号信息,不利于排查问题,我们需要给线程命名。
4 为线程进行命名
在并发编程中我们一定要为线程命名,这样有助于排查问题,关于如何命名我们可以参考DUBBO源码,分析FixedThreadPool线程池会发现其使用命名工厂为生产者和消费者线程进行命名
public class FixedThreadPool implements ThreadPool {
public Executor getExecutor(URL url) {
// 线程名称
String name = url.getParameter(Constants.THREAD_NAME_KEY, Constants.DEFAULT_THREAD_NAME);
// 线程个数默认200
int threads = url.getParameter(Constants.THREADS_KEY, Constants.DEFAULT_THREADS);
// 队列容量默认0
int queues = url.getParameter(Constants.QUEUES_KEY, Constants.DEFAULT_QUEUES);
// 队列容量等于0使用阻塞队列SynchronousQueue
// 队列容量小于0使用无界阻塞队列LinkedBlockingQueue
// 队列容量大于0使用有界阻塞队列LinkedBlockingQueue
// NamedInternalThreadFactory为线程命名
return new ThreadPoolExecutor(threads, threads, 0, TimeUnit.MILLISECONDS,
queues == 0 ? new SynchronousQueue
() : (queues < 0 ? new LinkedBlockingQueue
() : new LinkedBlockingQueue
(queues)), new NamedInternalThreadFactory(name, true), new AbortPolicyWithReport(name, url));
}
}
生产者默认线程名DubboServerHandler
public abstract class AbstractServer extends AbstractEndpoint implements Server {
protected static final String SERVER_THREAD_POOL_NAME = "DubboServerHandler";
}
生产者线程快照信息如下
"DubboServerHandler-1.1.1.1:20881-thread-20" #511 daemon prio = 5 os_prio = 0 tid = 0x00001f153121f200 nid = 0xd1a waiting on condition [0x00001f14edcdf000]
java.lang.Thread.State: WAITING (parking)
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x00000001e1f3abc0> (a java.util.concurrent.SynchronousQueue$TransferStack)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java : 115)
at java.util.concurrent.SynchronousQueue$TransferStack.awaitFulfill(SynchronousQueue.java : 452)
at java.util.concurrent.SynchronousQueue$TransferStack.transfer(SynchronousQueue.java : 312)
at java.util.concurrent.SynchronousQueue.take(SynchronousQueue.java : 924)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java : 1011)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java : 1121)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java : 111)
消费者默认线程名DubboClientHandler
public abstract class AbstractClient extends AbstractEndpoint implements Client {
protected static final String CLIENT_THREAD_POOL_NAME = "DubboClientHandler";
}
消费者线程快照信息如下
20881-thread-10" #688 daemon prio=1 os_prio=0 tid=0x00001f6114004800 nid=0x14d8 waiting on condition [0x00001f63e131a000] :
TIMED_WAITING (parking) :
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x00000006e21df0d0> (a java.util.concurrent.SynchronousQueue$TransferStack)
at java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:111)
at java.util.concurrent.SynchronousQueue$TransferStack.awaitFulfill(SynchronousQueue.java:460)
at java.util.concurrent.SynchronousQueue$TransferStack.transfer(SynchronousQueue.java:361)
at java.util.concurrent.SynchronousQueue.poll(SynchronousQueue.java:141)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1066)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1111)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:611)
at java.lang.Thread.run(Thread.java:141)
5 创建一个规范线程池
我们参考DUBBO线程池定义命名工厂
public class NamedInternalThreadFactory extends NamedThreadFactory {
public NamedInternalThreadFactory() {
super();
}
public NamedInternalThreadFactory(String prefix) {
super(prefix, false);
}
public NamedInternalThreadFactory(String prefix, boolean daemon) {
super(prefix, daemon);
}
public Thread newThread(Runnable runnable) {
String name = mPrefix + mThreadNum.getAndIncrement();
InternalThread ret = new InternalThread(mGroup, runnable, name, 0);
ret.setDaemon(mDaemon);
return ret;
}
}
public class NamedThreadFactory implements ThreadFactory {
protected static final AtomicInteger POOL_SEQ = new AtomicInteger(1);
protected final AtomicInteger mThreadNum = new AtomicInteger(1);
protected final String mPrefix;
protected final boolean mDaemon;
protected final ThreadGroup mGroup;
public NamedThreadFactory() {
this("pool-" + POOL_SEQ.getAndIncrement(), false);
}
public NamedThreadFactory(String prefix) {
this(prefix, false);
}
public NamedThreadFactory(String prefix, boolean daemon) {
mPrefix = prefix + "-thread-";
mDaemon = daemon;
SecurityManager s = System.getSecurityManager();
mGroup = (s == null) ? Thread.currentThread().getThreadGroup() : s.getThreadGroup();
}
public Thread newThread(Runnable runnable) {
String name = mPrefix + mThreadNum.getAndIncrement();
Thread ret = new Thread(mGroup, runnable, name, 0);
ret.setDaemon(mDaemon);
return ret;
}
public ThreadGroup getThreadGroup() {
return mGroup;
}
}
再定义一个线程池,在线程池执行方法开放一个业务名称参数供调用方设置
public class ThreadPoolStarter {
public static ThreadPoolExecutor getExecutor(String threadName) {
if (executor == null) {
synchronized (ThreadPoolStarter.class) {
if (executor == null) {
int coreSize = Runtime.getRuntime().availableProcessors();
BlockingQueue
queueToUse = new LinkedBlockingQueue (QUEUE_SIZE); executor = new ThreadPoolExecutor(coreSize, POOL_CORE_SIZE, MAX_SIZE, TimeUnit.SECONDS, queueToUse, new NamedInternalThreadFactory(threadName, true), new AbortPolicyDoReport(threadName));
}
}
}
return executor;
}
}
public class ThreadExecutor {
public static void execute(String bizName, Runnable job) {
ThreadPoolStarter.getExecutor(bizName).execute(job);
}
public static Future> sumbit(String bizName, Runnable job) {
return ThreadPoolStarter.getExecutor(bizName).submit(job);
}
}
编写一个实例进行测试
public void testThread() throws Exception {
for (int i = 0; i < 10000; i++) {
ThreadExecutor.execute("BizName", new Runnable() {
public void run() {
System.out.println("公众号互联网公园");
}
});
Thread.sleep(1000L);
}
}
}
再观察线程快照可以清晰查看业务名
#262 daemon prio=5 os_prio=0 tid=0x0000000023b5c000 nid=0x31d4 waiting on condition [0x000000003c0be000]
WAITING (parking) :
at sun.misc.Unsafe.park(Native Method)
- parking to wait for <0x00000006c35781f0> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175)
at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2039)
at java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:442)
at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1074)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1134)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
6 文章总结
本文首先介绍了《阿里巴巴JAVA开发手册》不允许使用Executors创建线程池这个规定,然后从资源和排查问题两个角度分析了为什么这么规定,最后我们参考DUBBO线程池声明方式创建了一个规范线程池,这样使用线程池有助于快速定位和排查问题。
长按二维码关注更多精彩文章