Hacker News Reader

Putting them in the same container means I have to manage the processes myself with something like supervisord. A docker-compose file makes this trivial because it is just a matter of a few lines of YAML that link the services together. It also means I can ship updates to my application and restart it without touching the rock-solid Postgres image.

View on HN

What do you mean by "manage the processes"?

Why would you need to touch postgres when updating your application?

Typically the process to update an application is to push up a new image and then tell docker to stop the old one, and load the new one in its place (yes, I know there are 0-downtime ways, but lets keep it simple).

If you put both processes in the same container, then you cannot simply upload a new image and restart the container. This is where it gets very complicated, error prone, and not in line with what I understand to be current best practices with docker.

I would not use Docker this way. I would treat the docker container like a server. I would update the application inside the running container.

Reply on news.ycombinator.com