-
Notifications
You must be signed in to change notification settings - Fork 2
Description
It's pretty difficult at the moment to trace through what code is being run when a deployment happens. This is because it's assembled across three repositories then compiled into a single Docker container.
The code that does our deployment is currently spread across the deeson-piplines, pipeline-ci-tools and git-relay repositories. In reality, I think we could probably just extract all of this out into a single deploy.sh script that we put in the quickstart repo and run on a standard alpine container.
Does anyone actually use any of the relay types other than snapshot? Especially since the introduction of artifacts within bitbucket pipleines, you can assume the repo is clean between builds unless specific artifacts are defined so we don't really need mirror any more. I'm not sure what subtree-snapshot is or does I'll confess. Does it apply to any more than one of our projects?