Linux Administration Cookbook

How it works...

All we've done here is highlight that containers are ephemeral by nature, and while you can stop and start the same container (minus the --rm to the docker run command), you're running in a transient state until you tag your container and upload it to a registry somewhere. 

It's generally not good practice to build a container by starting one and then installing a bunch of software inside it, before leaving it and saving it for later. The better method is to use a Dockerfile or some other automated and reproducible way of building containers. 

What we've also done is point out that while docker containers should be a self-contained little entity, that doesn't mean you can't hop inside them to see what's going on, and even ...