Docker Basics

A container is a program running in its own small, sealed box with its own files. Real systems are made of several of them working together, and when one stops, the first things you do are look at what is running, read what it said, and look around inside its neighbours. Your lab has three containers and one of them will not start. You will find out why, fix it and start it again, with the only four docker commands this terminal allows.

docker ps -a lists every container, stopped ones too. Read the STATUS column: web and cache say Up and (healthy), and worker says Exited (1), so it stopped as soon as it started. Plain docker ps would not have shown it at all. A stopped container still has its log, and the second command prints it: the first line begins worker: cannot start and says its settings file has no QUEUE_HOST line.

You cannot go inside a stopped container, but web shares the settings folder with it, a named volume. docker exec runs one command inside a running container: this one prints the settings file, four lines ending with POLL_SECONDS=5. The second adds the missing line, naming the service that keeps the queue, cache. (For a shell inside, type docker exec -it web sh in your own terminal.) Then docker restart worker starts it again and prints its name.

Now worker says Up, and for a few seconds (health: starting) before (healthy): run it again if it has not changed yet. Typing docker logs worker in your own terminal now ends with worker: finished job lines. Then press Stop the lab on the lab bar and Start the lab again, and run docker ps -a: the worker comes up healthy at once, because your line is in the named volume. Anything you had written inside a container outside /settings would be gone, which is how containers behave everywhere.