What is flocker?

Flocker solves a specific headache for teams running applications in containers: what happens to your data when a container moves to a different server. Normally, if you're running something like a database inside a container, the data it stores stays stuck on that one server. If the container gets moved elsewhere, the data doesn't follow. This tool makes that data portable, so your storage travels with your container wherever it goes. The way it works is by creating what it calls "datasets" instead of standard storage volumes. When you use it to manage your applications, the dataset and the container are handled together as a pair. So if a container needs to move from one server to another in your cluster, the associated data volume moves with it automatically. You can also use it just to manage data volumes on their own if you prefer to handle your containers through other means. This is mainly aimed at operations teams who need to run stateful services like databases in production. For example, if you're running a database that needs to be relocated because a server is getting crowded or needs maintenance, normally that's a painful process of manually copying data around. With this tool, the data just follows the container, making those migrations much simpler. The project was built by ClusterHQ, a small team of engineers with deep experience in distributed systems. It's written in Python and was under active development, with version 1.0 released in June 2015. The README doesn't go into much detail about the specific technical architecture or setup process, pointing instead to separate documentation for installation guides and tutorials.