We are proud to announce our v0.5.1 release. Deis itself is now a set of cooperating Docker containers and Deis can build and deploy Dockerfile-based projects as well as Heroku-style apps.
Deis is a Django/Celery API server, Python CLI and set of Chef cookbooks that combine to provide a Heroku-inspired application platform for public and private clouds. Your PaaS. Your Rules.
Containerization of Deis Components. Deis' moving parts are now a set of trusted docker images. This will result in a more modular system that is easier to deploy and to manage.
Dockerfile Apps. In addition to Deis' broad support for Heroku-style apps, deis/builder can now build any application based on a
Binary CLI. While developers may prefer to run the Deis client from pypi or a local repository, the
make client_binary command will build an executable version with no additional requirements. Executables for some platforms are available for download:
We are proud to welcome @bacongobbler as a new maintainer. Formerly one of our most productive external contributors, Matthew is now working on Deis full time!
We want to thank the following Deis community members for creating GitHub issues, providing support to others, and working on various Deis branches:
Whew--we hope that's everyone! If we slighted your contributions, please let us know. We will gladly add it to these release notes.
As part of moving Deis into Docker containers, we had to change the exposed ports for some core platform services. For example, the Django API server is now exposed on 8000/tcp rather than 80/tcp. We will soon distribute a new proxy service that exposes the platform components on standard ports.
We have seen some issues with timeouts due to the public Docker Index. This can manifest as timeout errors when bootstrapping nodes. Typically re-running the provisioning operations will resolve the issue. We are working with the Docker team on a more robust resolution.
Update documentation to catch up with changes from the switching to a Docker-based system. We would like to add a description of the new container architecture as well as instructions for using the new
Dockerfile build workflow.
We are now prototyping a new scheduler based on Apache Mesos, an enterprise-grade cluster manager. This change to the scheduler implementation will be transparent to end users, but should allow Deis to scale to 10,000+ nodes.
In order to facilitate a streamlined CI/CD process, we need an ability to promote existing Docker images as builds (bypassing the
git push process). We are currently investigating the best workflow.
For more details, see issues tagged security on GitHub.
Though Chef will continue to be the recommended way to deploy Deis, we are moving away from requiring a Chef Server and using Data Bags for cluster configuration. Once we implement the new scheduler, we can remove the Chef dependency and provide support for other configuration management solutions (Puppet, Ansible, Salt).
We need to make it as easy for ops folks to publish a set of reusable backing services (databases, queues, storage, etc) and allow developers to attach those services to applications. This will be done in a loosely coupled way, following Twelve Factor best practices. You can review the initial implementation and follow progress on this GitHub issue.
Though we provide the ability to run admin commands inside containers, we don't currently support interactive shells into containers (i.e.
deis run bash). Once this infrastructure is in place, this will also allow us to implement log tailing and other real-time features.
Users should be able to push builds to Deis even if they don't use
git. We are investigating a
deis push command that would not require a specific VCS.
You can learn about other ways to get involved on our website.
Deis core maintainer and Software Engineer at Engine Yard
Follow me on Twitter @RealMattBoersma