Inspect, logs & exec

Look inside a running container.

When something is wrong, three commands let you see inside a running container without stopping it: logs to read what it printed, exec to get a shell inside it, and inspect to dump its full configuration. Reach for them in that order most of the time — the answer is usually in the logs.

$ docker logs web              # everything the container wrote to stdout/stderr
$ docker logs -f web           # -f follows, like tail -f
$ docker logs --tail 20 web    # just the last 20 lines

Exec into a container

docker exec runs a new command inside an already-running container; with -it (interactive + a terminal) it gives you a shell to look around. This is different from docker run, which starts a new container — exec attaches to one that is already up. It is the everyday way to poke at a live container: check a file, run a quick command, see what the process sees.

$ docker exec -it web sh       # a shell inside the running container
/ # ls /etc/nginx
conf.d  nginx.conf  ...
/ # exit                       # leaves the shell; the container keeps running

Inspect the details

docker inspect prints a container’s complete state as JSON — its IP address, mounts, environment, restart policy, and more. It is verbose, so use -f with a Go template to pull out just the field you want. This is where you go when you need a fact about the container that the logs will not tell you, like which IP it got or exactly which volumes are mounted.

$ docker inspect -f '{{ .NetworkSettings.IPAddress }}' web
172.17.0.2
$ docker inspect -f '{{ .State.Status }} ({{ .State.ExitCode }})' web
running (0)