Sergio Carracedo 
Recently, I had the need to optimize the space required for my team’s backups. They are mainly websites, and currently, I was dumping all project folders to another disk and to Dropbox, but Dropbox has a serious performance issue indexing a large number of files.
Analyzing the backup process and talking to other programmers about their way of doing backups, I realized that part of the problem was already solved beforehand: GIT.
I’m already using git in the projects, and this means that all the content of the repo is already safeguarded on the server (Bitbucket or Github in my case) and on my other work machines, so I only had to solve the other part of the problem: the files that are not in the repo, for example, in the case of Drupals, the files folder, which contains files uploaded from the backend and which doesn’t make sense to have in the repo.
Well, after having watched Félix Gómez’s talk on creating command-line apps in PHP (https://www.youtube.com/watch?v=mGNgT6y_8NY), I set out to create my own to perform this task.
The app is very simple for now. Given a starting directory and a destination directory, it searches the starting directory for all folders containing a git repo and queries all the files that are in the repo; that list of files is stored in a temporary text file that is later passed as a parameter to rsync in the —exclude-list option so that rsync does not copy those files.
I have also added the —cvs-exclude option so that rsync ignores git’s own files, such as the .git folder.
With this simple tool that I have uploaded to a public repo https://github.com/sergiocarracedo/backup-tools in case anyone wants to try it or contribute their ideas, I have solved the backup problem.
As to-dos for this tool, there is the optimization of that exclude list since it can generate a file of several megabytes, creating a second command to check which repos have uncommitted changes, etc…