Docker bind: address already in use

What this error means

bind: address already in use means Docker tried to publish a container port onto a host port that is already occupied. The conflict is on the host machine, so the fix is to find the process or older compose stack using that port, then stop it or map the container to a different host port.

Why this happens

Port publishing uses the host network namespace. If another process already owns localhost:3000, 0.0.0.0:5432, or any other requested host port, Docker cannot bind that address for the new container.

For Docker bind: address already in use, focus on host port ownership and compose state. Docker daemon health, disk usage, and image contents are usually secondary unless other Docker commands also fail.

Common causes

• Host port is already used by another local process
• An old Docker Compose stack is still running
• Another container already publishes the same host port
• Compose maps multiple services to the same host port

Quick fixes

1. Read the error line and identify the host port Docker tried to bind.
2. Run lsof -i :3000 with the actual port from the error.
3. Stop the old process or compose stack that owns the port.
4. If both services must run, change the left side of the compose port mapping, such as 3001:3000.

Copy-paste commands

List running containers with published ports

docker ps --format "table {{.Names}}\t{{.Ports}}\t{{.Status}}"

Check compose services in this project

docker compose ps

Find the process using port 3000

lsof -nP -iTCP:3000 -sTCP:LISTEN

Stop the current compose stack

docker compose down

Check whether Docker itself is reachable

docker info

Real-world fixes

• A previous docker compose up may still be running from another terminal. Run docker compose ps in the project before starting a second stack.
• Local dev servers commonly occupy ports like 3000, 5173, 8000, and 8080. Stop the dev server or remap the Docker host port.
• For databases, avoid deleting volumes just to free a port. Stop the container first; the volume may contain local data.

Step-by-step troubleshooting

1. Find the port mentioned near bind: address already in use; it is usually shown as the host-side bind address.
2. Run docker ps --format "table {{.Names}}\t{{.Ports}}\t{{.Status}}" to see whether another container already publishes that port.
3. Run lsof -nP -iTCP:<port> -sTCP:LISTEN or the platform equivalent to find non-Docker processes.
4. Stop the old container, compose stack, or local process that owns the port.
5. If the port is intentionally in use, change only the host side of the mapping, for example 3001:3000, then retry docker compose up.

Platform-specific fixes

macOS

• Use lsof -nP -iTCP:<port> -sTCP:LISTEN to find apps, dev servers, or Docker Desktop proxies holding the port.

Linux

• Use ss -ltnp | grep :<port> when lsof is unavailable, then stop the owning service with the normal service manager.

Windows

• Use netstat -ano | findstr :<port> in PowerShell or Command Prompt, then match the PID in Task Manager.

How to prevent it

• Reserve predictable host ports per project and document them in the compose file.
• Run docker compose down when switching branches or projects that publish the same ports.
• Use project-specific compose project names when running multiple stacks at once.

Related errors

• Docker cannot connect to the daemon
• Docker Compose port is already allocated
• Docker no space left on device
• Docker permission denied

FAQ

What should I check first?

Check the host port shown beside bind: address already in use, then run docker ps and lsof for that port. The fastest fix is usually stopping the older process or remapping the host port.

Is this caused by the Dockerfile?

Usually no. A Dockerfile can expose a container port, but this error happens when Docker publishes that port to a host address that is already taken.

Can I run both services at the same time?

Yes, if they use different host ports. Keep the container port the same and change the host side, such as 3001:3000.

How do I know the fix worked?

Rerun the same docker run or docker compose up command. The fix worked when Docker starts the container without another bind: address already in use message and docker ps shows the expected port mapping.