I am working on a web application that is implemented in Mercurial and deployed to Amazon Web Servies. We are in the process of planning our repository structure, and I would like to know how other people handled this.
We need a separate storage stableand devfor bug fixes and new features, respectively. At Amazon-land, we have separate environments for the live, test, and Dev environments to run code, code that is going to live, and what we're just trying to do. The dev environment is likely to be built when we need it and then shuts down again, so its IP addresses can change.
Ideally, we would like hg pushto restore the chain from our local developer, right down to life. However, due to server security and the fact that IP servers (especially the transition device environment) can change, we may need the servers pulling out when they were created. We will also have cases where autoscaling spawns new servers, and we need to get the latest, verified code somewhere.
I am interested to know how you solved this problem (these problems), or if you have any suggestions on how we can do this.
source
share