还是按照常规思路,先查GC日志,定位为缓慢累计的内存泄露问题。然后把dump文件加载出来,最后的结果如下
两个关于线程池的类的对象和一个关于业务代码的内部类。很快定位到了如下代码
getScheduledThreadPoolExecutor().scheduleWithFixedDelay(new Runnable() {
@Override
public void run() {
...
}
}, 0, 1, TimeUnit.MINUTES);
上面代码在执行中会被多次执行(到OOM实际上已经执行了上千万次)。这里显然发现程序员犯了望文生义的错误。scheduleWithFixedDelay不是说延迟一段时间后执行一次,而是延迟一段时间后执行完毕后,再延迟相同的一段时间,再执行。推荐他使用搭配DelayQueue的线程池。实例代码如下
public class TestMain {
public static void main(String[] args) {
DelayQueue queue = new DelayQueue<>();//延迟队列
ThreadPoolExecutor threadPoolExecutor = new ThreadPoolExecutor(3,10,1000l, TimeUnit.MILLISECONDS,queue);
threadPoolExecutor.prestartAllCoreThreads();
threadPoolExecutor.execute(new TaskInfo(1000,1));
threadPoolExecutor.execute(new TaskInfo(2000,2));
threadPoolExecutor.execute(new TaskInfo(3000,3));
threadPoolExecutor.execute(new TaskInfo(4000,4));
}
}
class TaskInfo implements Delayed,Runnable {
//标记任务
private int id;
//具体时间
private long excuteTime;
public TaskInfo(long delay,int id) {
this.excuteTime = TimeUnit.NANOSECONDS.convert(delay, TimeUnit.MILLISECONDS)+System.nanoTime();
this.id = id;
}
@Override
public long getDelay(@NotNull TimeUnit unit) {
return unit.convert(this.excuteTime- System.nanoTime() , TimeUnit.NANOSECONDS);
}
@Override
public int compareTo(@NotNull Delayed o) {
TaskInfo msg = (TaskInfo)o;
return this.excuteTime>msg.excuteTime?1:( this.excuteTime<msg.excuteTime?-1:0);
}
@Override
public void run() {
System.out.println("run task:"+id);
}
}
—————————————————我是求真务实的分界————————————————————————
我们来具体分析产生OOM的原理。
public ScheduledFuture<?> scheduleWithFixedDelay(Runnable command,
long initialDelay,
long delay,
TimeUnit unit) {
if (command == null || unit == null)
throw new NullPointerException();
if (delay <= 0)
throw new IllegalArgumentException();
ScheduledFutureTask<Void> sft =
new ScheduledFutureTask<Void>(command,
null,
triggerTime(initialDelay, unit),
unit.toNanos(-delay));
RunnableScheduledFuture<Void> t = decorateTask(command, sft);
sft.outerTask = t;
delayedExecute(t);
return t;
}
其中ScheduledFutureTask的run方法是在执行后会把这个任务再放回任务队列。这样保证任务能够反复进行。
private class ScheduledFutureTask<V>
extends FutureTask<V> implements RunnableScheduledFuture<V> {
ScheduledFutureTask(Runnable r, V result, long ns, long period) {
super(r, result);//内部会调用Executors.callable(runnable, result);把Runnable对象转化为Callable对象(RunnableAdapter)。这里解释了为什么RunnableAdapter对象会有很多。
this.time = ns;
this.period = period;
this.sequenceNumber = sequencer.getAndIncrement();
}
public void run() {
boolean periodic = isPeriodic();
if (!canRunInCurrentRunState(periodic))
cancel(false);
else if (!periodic)
ScheduledFutureTask.super.run();
else if (ScheduledFutureTask.super.runAndReset()) {
setNextRunTime();
reExecutePeriodic(outerTask);
}
}
}
所谓的RunnableAdapter,顾名思义,就是把一个Runnable对象通过适配器模式转化为一个Callable的对象。
public static Callable<Object> callable(Runnable task) {
if (task == null)
throw new NullPointerException();
return new RunnableAdapter<Object>(task, null);
}
delayedExecute会把这个task追加到自己的workQueue中。
/**
* Main execution method for delayed or periodic tasks. If pool
* is shut down, rejects the task. Otherwise adds task to queue
* and starts a thread, if necessary, to run it. (We cannot
* prestart the thread to run the task because the task (probably)
* shouldn't be run yet.) If the pool is shut down while the task
* is being added, cancel and remove it if required by state and
* run-after-shutdown parameters.
*
* @param task the task
*/
private void delayedExecute(RunnableScheduledFuture<?> task) {
if (isShutdown())
reject(task);
else {
super.getQueue().add(task);
if (isShutdown() &&
!canRunInCurrentRunState(task.isPeriodic()) &&
remove(task))
task.cancel(false);
else
ensurePrestart();
}
}
初始化ScheduledThreadPoolExecutor时,可以看到没有选择,使用的内部类DelayedWorkQueue。该队列的视线实际上就是DelayQueue,但是区别在于没有设置容量设置!!
Using a custom queue (DelayedWorkQueue), a variant of unbounded DelayQueue.
public ScheduledThreadPoolExecutor(int corePoolSize) {
super(corePoolSize, Integer.MAX_VALUE, 0, NANOSECONDS,
new DelayedWorkQueue());
}
这个该死的没有容量限制,最终导致了本次OOM的发生。