Redisson 提供的分布式锁
image.png使用实例
private void redissonDoc() throws InterruptedException {
//1. 普通的可重入锁
RLock lock = redissonClient.getLock("generalLock");
// 拿锁失败时会不停的重试
// 具有Watch Dog 自动延期机制 默认续30s 每隔30/3=10 秒续到30s
lock.lock();
// 尝试拿锁10s后停止重试,返回false
// 具有Watch Dog 自动延期机制 默认续30s
boolean res1 = lock.tryLock(10, TimeUnit.SECONDS);
// 拿锁失败时会不停的重试
// 没有Watch Dog ,10s后自动释放
lock.lock(10, TimeUnit.SECONDS);
// 尝试拿锁100s后停止重试,返回false
// 没有Watch Dog ,10s后自动释放
boolean res2 = lock.tryLock(100, 10, TimeUnit.SECONDS);
//2. 公平锁 保证 Redisson 客户端线程将以其请求的顺序获得锁
RLock fairLock = redissonClient.getFairLock("fairLock");
//3. 读写锁 没错与JDK中ReentrantLock的读写锁效果一样
RReadWriteLock readWriteLock = redissonClient.getReadWriteLock("readWriteLock");
readWriteLock.readLock().lock();
readWriteLock.writeLock().lock();
}
如果拿到分布式锁的节点宕机,且这个锁正好处于锁住的状态时,会出现锁死的状态,为了避免这种情况的发生,锁都会设置一个过期时间。这样也存在一个问题,加入一个线程拿到了锁设置了30s超时,在30s后这个线程还没有执行完毕,锁超时释放了,就会导致问题,Redisson给出了自己的答案,就是 watch dog 自动延期机制。
Redisson提供了一个监控锁的看门狗,它的作用是在Redisson实例被关闭前,不断的延长锁的有效期,也就是说,如果一个拿到锁的线程一直没有完成逻辑,那么看门狗会帮助线程不断的延长锁超时时间,锁不会因为超时而被释放。
默认情况下,看门狗的续期时间是30s,也可以通过修改Config.lockWatchdogTimeout来另行指定。
另外Redisson 还提供了可以指定leaseTime参数的加锁方法来指定加锁的时间。超过这个时间后锁便自动解开了,不会延长锁的有效期
watch dog 核心源码解读
// 直接使用lock无参数方法
public void lock() {
try {
lock(-1, null, false);
} catch (InterruptedException e) {
throw new IllegalStateException();
}
}
// 进入该方法 其中leaseTime = -1
private void lock(long leaseTime, TimeUnit unit, boolean interruptibly) throws InterruptedException {
long threadId = Thread.currentThread().getId();
Long ttl = tryAcquire(-1, leaseTime, unit, threadId);
// lock acquired
if (ttl == null) {
return;
}
//...
}
// 进入 tryAcquire(-1, leaseTime, unit, threadId)
private Long tryAcquire(long waitTime, long leaseTime, TimeUnit unit, long threadId) {
return get(tryAcquireAsync(waitTime, leaseTime, unit, threadId));
}
// 进入 tryAcquireAsync
private <T> RFuture<Long> tryAcquireAsync(long waitTime, long leaseTime, TimeUnit unit, long threadId) {
if (leaseTime != -1) {
return tryLockInnerAsync(waitTime, leaseTime, unit, threadId, RedisCommands.EVAL_LONG);
}
//当leaseTime = -1 时 启动 watch dog机制
RFuture<Long> ttlRemainingFuture = tryLockInnerAsync(waitTime,
commandExecutor.getConnectionManager().getCfg().getLockWatchdogTimeout(),
TimeUnit.MILLISECONDS, threadId, RedisCommands.EVAL_LONG);
//执行完lua脚本后的回调
ttlRemainingFuture.onComplete((ttlRemaining, e) -> {
if (e != null) {
return;
}
if (ttlRemaining == null) {
// watch dog
scheduleExpirationRenewal(threadId);
}
});
return ttlRemainingFuture;
}
scheduleExpirationRenewal 方法开启监控
private void scheduleExpirationRenewal(long threadId) {
ExpirationEntry entry = new ExpirationEntry();
//将线程放入缓存中
ExpirationEntry oldEntry = EXPIRATION_RENEWAL_MAP.putIfAbsent(getEntryName(), entry);
//第二次获得锁后 不会进行延期操作
if (oldEntry != null) {
oldEntry.addThreadId(threadId);
} else {
entry.addThreadId(threadId);
// 第一次获得锁 延期操作
renewExpiration();
}
}
// 进入 renewExpiration()
private void renewExpiration() {
ExpirationEntry ee = EXPIRATION_RENEWAL_MAP.get(getEntryName());
//如果缓存不存在,那不再锁续期
if (ee == null) {
return;
}
Timeout task = commandExecutor.getConnectionManager().newTimeout(new TimerTask() {
@Override
public void run(Timeout timeout) throws Exception {
ExpirationEntry ent = EXPIRATION_RENEWAL_MAP.get(getEntryName());
if (ent == null) {
return;
}
Long threadId = ent.getFirstThreadId();
if (threadId == null) {
return;
}
//执行lua 进行续期
RFuture<Boolean> future = renewExpirationAsync(threadId);
future.onComplete((res, e) -> {
if (e != null) {
log.error("Can't update lock " + getName() + " expiration", e);
return;
}
if (res) {
//延期成功,继续循环操作
renewExpiration();
}
});
}
//每隔internalLockLeaseTime/3=10秒检查一次
}, internalLockLeaseTime / 3, TimeUnit.MILLISECONDS);
ee.setTimeout(task);
}
//lua脚本 执行包装好的lua脚本进行key续期
protected RFuture<Boolean> renewExpirationAsync(long threadId) {
return evalWriteAsync(getName(), LongCodec.INSTANCE, RedisCommands.EVAL_BOOLEAN,
"if (redis.call('hexists', KEYS[1], ARGV[2]) == 1) then " +
"redis.call('pexpire', KEYS[1], ARGV[1]); " +
"return 1; " +
"end; " +
"return 0;",
Collections.singletonList(getName()),
internalLockLeaseTime, getLockName(threadId));
}
关键结论
上述源码读过来我们可以记住几个关键情报:
- watch dog 在当前节点存活时每10s给分布式锁的key续期 30s;
- watch dog 机制启动,且代码中没有释放锁操作时,watch dog 会不断的给锁续期;
- 从可2得出,如果程序释放锁操作时因为异常没有被执行,那么锁无法被释放,所以释放锁操作一定要放到 finally {} 中;
看到3的时候,可能会有人有疑问,如果释放锁操作本身异常了,watch dog 还会不停的续期吗?下面看一下释放锁的源码,找找答案
// 锁释放
public void unlock() {
try {
get(unlockAsync(Thread.currentThread().getId()));
} catch (RedisException e) {
if (e.getCause() instanceof IllegalMonitorStateException) {
throw (IllegalMonitorStateException) e.getCause();
} else {
throw e;
}
}
}
// 进入 unlockAsync(Thread.currentThread().getId()) 方法 入参是当前线程的id
public RFuture<Void> unlockAsync(long threadId) {
RPromise<Void> result = new RedissonPromise<Void>();
//执行lua脚本 删除key
RFuture<Boolean> future = unlockInnerAsync(threadId);
future.onComplete((opStatus, e) -> {
// 无论执行lua脚本是否成功 执行cancelExpirationRenewal(threadId) 方法来删除EXPIRATION_RENEWAL_MAP中的缓存
cancelExpirationRenewal(threadId);
if (e != null) {
result.tryFailure(e);
return;
}
if (opStatus == null) {
IllegalMonitorStateException cause = new IllegalMonitorStateException("attempt to unlock lock, not locked by current thread by node id: "
+ id + " thread-id: " + threadId);
result.tryFailure(cause);
return;
}
result.trySuccess(null);
});
return result;
}
// 此方法会停止 watch dog 机制
void cancelExpirationRenewal(Long threadId) {
ExpirationEntry task = EXPIRATION_RENEWAL_MAP.get(getEntryName());
if (task == null) {
return;
}
if (threadId != null) {
task.removeThreadId(threadId);
}
if (threadId == null || task.hasNoThreads()) {
Timeout timeout = task.getTimeout();
if (timeout != null) {
timeout.cancel();
}
EXPIRATION_RENEWAL_MAP.remove(getEntryName());
}
}
释放锁的操作中 有一步操作是从 EXPIRATION_RENEWAL_MAP 中获取 ExpirationEntry 对象,然后将其remove,结合watch dog中的续期前的判断:
EXPIRATION_RENEWAL_MAP.get(getEntryName());
if (ent == null) {
return;
}
如果释放锁操作本身异常了,watch dog 还会不停的续期吗?不会,因为无论释放锁操作是否成功,EXPIRATION_RENEWAL_MAP中的目标 ExpirationEntry 对象已经被移除了,watch dog 通过判断后就不会继续给锁续期了。
网友评论