Attaching to Running Containers

Imagine your container is a room.

docker logs is like watching security camera footage. You can see what happened, but you can’t interact.

docker attach is like picking up a phone that’s connected to someone already in the room. You can hear what they’re saying (see stdout), but you’re connected to whatever they’re doing. If you hang up (Ctrl+C), they might hang up too (container might stop).

docker exec is like walking into the room yourself. You’re a new person in the room, separate from whoever else is there. You can look around, touch things, run commands. When you leave, the original person stays.

For debugging and exploration, docker exec is almost always what you want.

docker exec — Running Commands Inside

# Run a single command
docker exec mycontainer ls /app

# Start an interactive shell
docker exec -it mycontainer /bin/bash
# or for Alpine (no bash)
docker exec -it mycontainer /bin/sh

The flags:

• -i (interactive): Keep STDIN open
• -t (tty): Allocate a pseudo-terminal

Together, -it gives you an interactive shell experience.

docker attach — Connecting to the Main Process

docker attach mycontainer

This connects your terminal to the container’s main process (PID 1). You see what it outputs.

Warning: If you press Ctrl+C while attached, you might kill the main process, stopping the container! Use Ctrl+P, Ctrl+Q to detach without stopping.

When to Use Each

Goal	Command
Run a quick command	docker exec mycontainer <cmd>
Interactive debugging shell	docker exec -it mycontainer /bin/bash
See live output of main process	docker attach mycontainer
Follow logs (safer)	docker logs -f mycontainer

Check what’s running inside

docker exec mycontainer ps aux

Check network connectivity

docker exec mycontainer ping -c 3 google.com
docker exec mycontainer curl localhost:8080

Check files

docker exec mycontainer cat /app/config.yaml
docker exec mycontainer ls -la /var/log/

Check environment variables

docker exec mycontainer env

Get a shell and explore

docker exec -it mycontainer /bin/bash
# Now you're "inside" — explore as needed
# Type 'exit' to leave

When you docker exec, Docker:

1. Uses the container’s namespaces (same view of filesystem, network, etc.)
2. Creates a new process in those namespaces
3. This process is separate from PID 1

This is why exec is safe — killing your exec process doesn’t kill the container.

Container (myapp)
├── PID 1: nginx (main process)
├── PID 20: nginx worker
├── PID 21: nginx worker
└── PID 42: /bin/bash (your exec session) ← separate process

---