有一個後臺程序需要這麼一個功能:
需要知道自己什麼時候進入穩定狀態。
我原來用一種沙雕方式實現的,即開一個線程讓一個計數器定時自增,自增到某個數即表示已經進入穩定狀態,如果中間有打破這種狀態的操作,那麼就讓計數器重置爲0:
public static AtomicInteger STABLE_TIMER = new AtomicInteger(0);
taskExecutors.execute(new Runnable() {
@Override
public void run() {
while (true) {
try {
TimeUnit.SECONDS.sleep(1);
if (STABLE_TIMER.getAndIncrement() == LC.STABLE_MIN_TIME) {
L.i(Thread.currentThread().getName() + "|進入穩定狀態,報告自身狀態");
//TODO 進入穩定狀態需要做的操作
}
} catch (InterruptedException e) {
e.printStackTrace();
STABLE_TIMER.decrementAndGet();
}
}
}
});
雖然勉強達到了效果,但實在是…
太醜了。
後來想了想,其實可以找到一些類比場景,比如輪詢,比如心跳;
再仔細想想,原來看過的netty心跳源碼中有一段和這個需求非常相似,比如讀取超時的判斷就依賴一個ReaderIdleTimeoutTask:
ScheduledFuture<?> schedule(ChannelHandlerContext ctx, Runnable task, long delay, TimeUnit unit) {
return ctx.executor().schedule(task, delay, unit);
}
private final class ReaderIdleTimeoutTask extends AbstractIdleTask {
ReaderIdleTimeoutTask(ChannelHandlerContext ctx) {
super(ctx);
}
@Override
protected void run(ChannelHandlerContext ctx) {
long nextDelay = readerIdleTimeNanos;
if (!reading) {
nextDelay -= ticksInNanos() - lastReadTime;
}
if (nextDelay <= 0) {
// Reader is idle - set a new timeout and notify the callback.
readerIdleTimeout = schedule(ctx, this, readerIdleTimeNanos, TimeUnit.NANOSECONDS);
boolean first = firstReaderIdleEvent;
firstReaderIdleEvent = false;
try {
IdleStateEvent event = newIdleStateEvent(IdleState.READER_IDLE, first);
channelIdle(ctx, event);
} catch (Throwable t) {
ctx.fireExceptionCaught(t);
}
} else {
// Read occurred before the timeout - set a new timeout with shorter delay.
readerIdleTimeout = schedule(ctx, this, nextDelay, TimeUnit.NANOSECONDS);
}
}
}
其中channelIdle方法就相當於進入了一種判定狀態,在這裏當然是判定爲超時了。
所以Netty的心跳就是這樣:定時自檢是否已經達到一種“超時狀態“,如果當前沒有,那麼就上次的時間進行計算,再決定下一次自檢的時間。
那麼可以直接仿寫一個task,整一個工具類出來:
public class StableStateHolder {
private static class Holder {
static StableStateHolder INSTANCE = new StableStateHolder();
}
public static StableStateHolder get() {
return Holder.INSTANCE;
}
private ScheduledExecutorService stableTimer;
private ScheduledFuture<?> stableTimeout;
private static final long DELAY = 10 * 1000;
private static final long DELAY_LONG = 20 * 1000;
private long currentTimeMills;
private StableStateHolder() {
stableTimer = new ScheduledThreadPoolExecutor(1);
stableTimeout = schedule(new StableTimeoutTask(), DELAY_LONG, TimeUnit.MILLISECONDS);
currentTimeMills = System.currentTimeMillis();
}
public void tick() {
currentTimeMills = System.currentTimeMillis();
}
private ScheduledFuture<?> schedule(Runnable task, long delay, TimeUnit unit) {
return stableTimer.schedule(task, delay, unit);
}
private final class StableTimeoutTask implements Runnable {
@Override
public void run() {
long prevDelay = System.currentTimeMillis() - currentTimeMills;
if (prevDelay > DELAY) {
//TODO 進入判定狀態
stableTimeout = schedule(this, DELAY, TimeUnit.MILLISECONDS);
} else {
stableTimeout = schedule(this, DELAY - prevDelay, TimeUnit.MILLISECONDS);
}
L.w("活躍線程數:" + Thread.activeCount());
}
}
}
這樣就好看多了,雖然還是有點沙雕。哈哈哈哈。