Distributed systems use various coordination patterns beyond simple distributed locking. Initially, distributed locks ensure exclusive access to resources, but they can be improved with leases to handle node failures and fencing tokens to prevent stale operations. Leader election is introduced when the goal shifts from exclusive ownership of an operation to having a single node coordinate the rest of the cluster. Unlike short-lived locks, leadership is a long-lived role, making leader election suitable for tasks like job scheduling or cluster management. However, leaders can also fail, necessitating a mechanism to detect their unavailability and elect a new one, which still relies on concepts like leases and timeouts. Leader election alone does not guarantee agreement among nodes; that task falls to consensus mechanisms. Consensus ensures multiple nodes agree on the same state, which is crucial for maintaining consistency in configurations or cluster membership. Modern coordination platforms like ZooKeeper or etcd combine leases, fencing, leader election, and consensus into a single infrastructure. Understanding the distinct problems each coordination primitive solves—locking for ownership, leader election for coordination, and consensus for agreement—is key. While distributed locks are foundational, many systems achieve correctness through alternative methods like optimistic concurrency or queue-based processing. Each coordination mechanism addresses specific challenges, building a layered toolkit for reliable distributed systems.