How to Build a Blocking Queue with Two Condition Variables in Java

In Java, the Condition interface provides condition variables for monitors, allowing multiple distinct wait sets, unlike the single built‑in condition variable used with synchronized.

The combination of Lock and Condition therefore enables a monitor that can manage several condition variables, improving code readability and simplifying implementation.

Blocking queue requires two condition variables

To implement a blocking queue you need to guard two states:

Queue not empty : when the queue is empty, consumer threads must wait before dequeuing.

Queue not full : when the queue reaches its capacity, producer threads must wait before enqueuing.

The typical implementation uses two Condition objects obtained from a ReentrantLock:

ReentrantLock lock = new ReentrantLock();
Condition notEmpty = lock.newCondition();
Condition notFull = lock.newCondition();

Producer threads call await() on notFull when the queue is full and signal notEmpty after inserting an element. Consumer threads call await() on notEmpty when the queue is empty and signal notFull after removing an element.

The semantics of await(), signal() and signalAll() are equivalent to wait(), notify() and notifyAll() used with synchronized, but they must be used only with the Lock / Condition pair. Mixing wait() / notify() with a Lock / Condition monitor will cause runtime errors.

Attempting to call wait() or notify() inside a Lock / Condition based monitor is a common mistake that can lead to serious bugs and even job loss for developers unfamiliar with the correct usage.