RedisLockRegistry improvement, using pub-sub for lock to without spinlock

[Meteorkor]

Expected Behavior

Remove spinlock for used by RedisLockRegistry
So, I thought about how to improve this part through redis pub-sub.

Current Behavior

Currently, spinlock is used to acquire redis-lock.

https://github.com/spring-projects/spring-integration/blob/main/spring-integration-redis/src/main/java/org/springframework/integration/redis/util/RedisLockRegistry.java#L231-L233

while (!obtainLock()) {
    Thread.sleep(100); //NOSONAR
}

https://github.com/spring-projects/spring-integration/blob/main/spring-integration-redis/src/main/java/org/springframework/integration/redis/util/RedisLockRegistry.java#L291-L298

long expire = now + TimeUnit.MILLISECONDS.convert(time, unit);
boolean acquired;
while (!(acquired = obtainLock()) && System.currentTimeMillis() < expire) { //NOSONAR
    Thread.sleep(100); //NOSONAR
}

If a vm instance acquires a lock and persists for a long time, or if multiple vm instances try to acquire a lock, I think the part where obtainLock() is called continuously is wasteful.

Context

I would like to contribute to the project.
If it's okay with you, I'd like to write a PR and get a review.

[artembilan]

I’m not fully sure what solution you suggest, but opening PR and continue discussion over there won’t harm. If you have time to code your idea, of course…

thank you!

[vipul-mykaarma]

ever since we have upgraded srping-integration-redis to 5.5.8 we are encountering the following error
Connection failure occurred. Restarting subscription task after 5000 ms
Although lock functionality is working fine . The error is coming from RedisMessageListenerContainer present within redis lock registry

[artembilan]

@vipul-mykaarma ,

that means you have some connectivity issue, although it is self-recovering.
The logic is like this:

if (ex instanceof RedisConnectionFailureException) {
			if (isRunning()) {
				logger.error("Connection failure occurred. Restarting subscription task after " + recoveryInterval + " ms");
				sleepBeforeRecoveryAttempt();
				lazyListen();
			}
		} else {

So, all should be OK with a couple retries.

There is nothing RedisLockRegistry-related though.