Writing Mutex Logic for Concurrency Control

Yes. In a distributed system, there's a possibility of this happening. That's why it is good as a best practice to implement a robust exception handling technique to gracefully cater for this. Including an efficient retry mechanism to serialize the conflicting threads. But generally, and even in a distributed system, the possibility is almost zero. Considering that the redis SETNX operation used to acquire the lock is atomic.