【从问题开始】
在使用线程池的时候往往会看到线程池reject相关日志,例如以下日志:
[Running, pool size = 80, active threads = 3, queued tasks = 3, completed tasks = 1674]
遂产生疑问:为什么线程池大小为80,队列长度为3的线程池,只有3个active threads就会触发reject机制?
第一篇结尾处,根据execute方法的流程发散了两个问题:
- RejectExecution的日志输出为什么会将alive threads输出为3?
- 核心线程与非核心线程的添加流程是什么?为什么会添加失败?
本篇就问题1探寻线程池的拒绝策略。
目录
线程池的拒绝处理机制 - RejectedExecutionHandle
线程池的拒绝处理机制 - RejectedExecutionHandler
线程池的构造函数中指定了线程池的拒绝处理机制:
public ThreadPoolExecutor(int corePoolSize,
int maximumPoolSize,
long keepAliveTime,
TimeUnit unit,
BlockingQueue<Runnable> workQueue,
ThreadFactory threadFactory) {
this(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue,
threadFactory, defaultHandler);
}
public ThreadPoolExecutor(int corePoolSize,
int maximumPoolSize,
long keepAliveTime,
TimeUnit unit,
BlockingQueue<Runnable> workQueue,
RejectedExecutionHandler handler) {
this(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue,
Executors.defaultThreadFactory(), handler);
}
public ThreadPoolExecutor(int corePoolSize,
int maximumPoolSize,
long keepAliveTime,
TimeUnit unit,
BlockingQueue<Runnable> workQueue,
ThreadFactory threadFactory,
RejectedExecutionHandler handler) {
if (corePoolSize < 0 ||
maximumPoolSize <= 0 ||
maximumPoolSize < corePoolSize ||
keepAliveTime < 0)
throw new IllegalArgumentException();
if (workQueue == null || threadFactory == null || handler == null)
throw new NullPointerException();
this.acc = System.getSecurityManager() == null ?
null :
AccessController.getContext();
this.corePoolSize = corePoolSize;
this.maximumPoolSize = maximumPoolSize;
this.workQueue = workQueue;
this.keepAliveTime = unit.toNanos(keepAliveTime);
this.threadFactory = threadFactory;
this.handler = handler;
}
根据线程池的构造函数可知,传入的handler被保存至成员属性中。
private volatile RejectedExecutionHandler handler;
如果不传,默认使用defaultHandler。
再看handler属性的调用点:
final void reject(Runnable command) {
handler.rejectedExecution(command, this);
}
可知,在execute流程中,拒绝策略触发时,实际调用的就是RejectedExecutionHandler::rejectedExecution方法。
RjectedExecutionHandler是一个接口,只提供了一个接口方法:rejectedExecution方法。在ThreadPoolExecutor中有四个实现:
它们分别是:
CallerRunsPolicy
如果线程因队列阻塞无法执行,直接让调用线程执行这个任务
public void rejectedExecution(Runnable r, ThreadPoolExecutor e) {
if (!e.isShutdown()) {
r.run();
}
}
AbortPolicy和RejectedExecutionException
如果被调用线程因队列阻塞无法执行任务,抛出RejectedExecutionException异常
public void rejectedExecution(Runnable r, ThreadPoolExecutor e) {
throw new RejectedExecutionException("Task " + r.toString() +
" rejected from " +
e.toString());
}
当拒绝策略为AbortPolicy时,会抛出线程池拒绝异常。输出信息类似:
[Running, pool size = 80, active threads = 3, queued tasks = 3, completed tasks = 1674]
DiscardPolicy
处理不了,直接静默丢弃
public void rejectedExecution(Runnable r, ThreadPoolExecutor e) {
}
DiscardOldestPolicy
新来任务时塞不进去时,从队列头拉取任务丢弃,执行新来的任务,丢弃老的
public void rejectedExecution(Runnable r, ThreadPoolExecutor e) {
if (!e.isShutdown()) {
e.getQueue().poll();
e.execute(r);
}
}
其中,默认实现是AbortPolicy
private static final RejectedExecutionHandler defaultHandler =
new AbortPolicy();
在本例中,代码环境使用的即默认策略,抛出异常。
看AbortPolicy的实现:
public void rejectedExecution(Runnable r, ThreadPoolExecutor e) {
throw new RejectedExecutionException("Task " + r.toString() +
" rejected from " +
e.toString());
}
输出实际上是ThreadPoolExecutor重写的toString方法
public String toString() {
long ncompleted;
int nworkers, nactive;
final ReentrantLock mainLock = this.mainLock;
mainLock.lock();
try {
ncompleted = completedTaskCount;
nactive = 0;
nworkers = workers.size();
for (Worker w : workers) {
ncompleted += w.completedTasks;
if (w.isLocked())
// 这里可知nactive实际取的是加锁成功的worker数量,而非有一个算一个
++nactive;
}
} finally {
mainLock.unlock();
}
int c = ctl.get();
String rs = (runStateLessThan(c, SHUTDOWN) ? "Running" :
(runStateAtLeast(c, TERMINATED) ? "Terminated" :
"Shutting down"));
return super.toString() +
"[" + rs +
", pool size = " + nworkers +
", active threads = " + nactive +
", queued tasks = " + workQueue.size() +
", completed tasks = " + ncompleted +
"]";
}
toString方法对Worker的处理过程通过ReentrantLock加锁,加锁成功后才做workers状态的统计。从toString方法中可以看出,输出的几个核心指标:
- nworkers:线程池大小,取自成员变量workers的大小,这个大小既不是maxPoolSize也不是corePoolSize,而是实际创建出来的worker的数量。既然触发了reject,往往这个值是大于等于maxPoolSize的,正常应该不会比它小。
- nactive:活跃线程数量,实际上是取的worker中处于isLocked状态的数量。
- workQueue.zise():根workers的size不同,这个是线程池的阻塞队列长度,根据第一篇中分析的reject的触发场景,一般发现是满了,那就是因为阻塞队列满而触发reject,如果没满,那多半是因为线程的状态已经是非运行态了,例如Executors.newSingleThreadExecutor.execute()在没有任务执行时恰好触发GC后,可能输出阻塞队列不满就触发reject的日志。
- ncompleted:所有worker中completedTaskCount属性之和。
这就可以解释问题:为什么active thread还不满就触发reject了。但是同样,今天的分析也引出了以下几个问题:
由此我们提出以下几个问题,作为后续分析的方向:
- Worker是什么?它的运行机制是什么样的?
- 对Worker信息的统计为什么要加锁?
本篇结论:
- reject策略在ThreadPoolExecutor中有四种实现,其中默认实现是Abort Policy,即触发reject时,抛出RejectExecution。
- Abort Policy的错误信息实际是ThreadPoolExecutor中toString方法的输出信息,在输出内容中实际是对Worker的统计。而active thread实际代表了处于isLocked状态的worker数量。