At the conclusion of June 2015, a conference for Docker developers was held in San Francisco, USA. We contribute to your attention the story of EMC orchestrate Ryan Wollner about how DockerCon 2015 went, about the party and the ecosystem that developed around the league.

DockerCon15 Logo

I have been participating in the open source profession for to a greater extent than 5 years, I am engaged in network technologies, PaaS infrastructure and containerization technologies. Participation in projects such as OpenFlow / SDN, OpenStack was held with varying success, just Docker is objectively collecting maximum attention from the skillfulness community.

Developers approximately the world pay important attention to embedded architectures, journaling, monitoring, data migration, networks and running stateful services in containers) Before I plunge deeper into the topic of persistence volumes and show the results of my team, I lack to highlight few related topics of the conference and the community as a whole.

One of the themes of DockerCon 2015 was the use of Docker in product. Conditional who you deman about your experience with Dock worker you bet it was used, the answers will vary. Production using containers with microservices give notice glucinium "hell," as Brian Cantrill put information technology (if you haven't seen Brian speak about Unix philosophy and debugging at runtime, I highly recommend observation any videos of his reports, especially from the O'Reilly conference, which took place on the eve of DockerCon). But, on the other hand, it helps to split applications into limited domains with a high-level degree of manageableness, and development teams get the opportunity to effectively conduct the CI / CD debugging / delivery / deployment action. Netflix raw reference on githubis confirmation of this, and Andrian Cockcroft (Andrian Cockcroft) rather deeply rummaged in a subject. You can also see my send on microservices, which will help you understand the context.

Is Docker good for a sale? Below I will key the points that you need to know about before using Docker in production. I would also like to add that each of the items is a separate big topic of DockerCon.

Network

The interaction 'tween Docker and startup Socketplane resulted in the possibility of victimization Libnetwork technology in projects , which allows copulative Docker daemons from different hosts, which makes it easier for the container to transference IP traffic over the network link layer. Libnetwork is gimbaled from outside the main Docker daemon, is independent of the network implementation and, most importantly, hides implementation inside information from the drug user, allowing the uncastrated system to "just work".

Security

I talked with Eric Windish from Docker, who said that the security of the Docker engine was always granted key tending, since any vulnerability in it would mean a menace to the container. Comprehensive work is underway to improve the code in the security subsystem, and also use other tools supported AppArmor and SELinux . Information technology is very rum how the Docker security subsystem will unfold in concurrence with projects such as VMWare Lightwave .

Journaling, monitoring, manageability

Containers are, naturally, cool, but by launching at the same time a thousand or ten m containers, you realize the need for a tool that would allow you to efficiently debug, supervise and manage them. And it looks the likes of a solution awaits us. Project Orca, announced away Docker founder Solomon Heiks , aims to connect the Docker Locomotive, Dock-walloper Swarm, GUI, Docker Write out, as well as tools for security department, installing, deployment and configuration. Of course, it is not necessary to run Docker through Orca, but it would be great to hold a room to acquire all of the in a higher place without losing focal ratio and manageability. Separate tools (loggly, cadvisor , ruzit, datalog, backlog entries) compete for a place on this list, and it's just great!

Extensibility

Docker gives the initiative to the community, declaring that it wants to personify an extensible tool that allows individual plug-ins to figure out with the network, authentication, data warehouses. Currently there is VMWare Lightwave for authentication and EMC ScaleIO for persistence storage. Not so bad considering Docker is only two years sure-enough!

Persistent services with inward state preservation

EMC and ClusterHQ did a great Job together, creating the Flocker open origin project . It allows you to manage container volumes and provides data migration and high availability of volumes in situations where it is necessary to restore containerised applications.

Also of great interest at the conference was the announcement of extensions to the native volumes of Docker. Disdain the experimental status, start with version 1.7.x, these extensions provide various options for managing services while maintaining the internal state. Flocker itself bum work with Docker volume extensions exploitation the flag loudness-driver = flocker if the Flocker-Docker plugin is utilized .

Result of interactionc ClusterHQ, became the publication happening GitHub generator drivers that permit you to work with EMC ScaleIO and EMC XtremIO . At the Important Science laborator in San Francisco, we demonstrated all the driver features. As a result, people gathered to watch demos for beer and conversations in an friendly setting. This link contains the ScaleIO demo, which we showed at the meeting devoted to the work of Flocker + ScaleIO, deployed along Amazon AWS using the MEAN stack of applications that downloaded data from Chitter and placed in MongoDB. Meeting slides are also gettable here .

Not all DockerCon guests had plenty seating; many listeners stood or sat along the floor for the intact session. We hoped that we would have a petite more time sol that the guys with laptops had the opportunity to hack our demonstrate, but the 40 proceedings allotted to us were clearly non plenty. They did everything they could!

Through a inborn Go implementation called RexRay , EMC integrates persistence volume management paths, but without the automatic data migration that Flocker produces. The possibilities of RexRay are very broad, for instance, you can use several backends concurrently, enunciat, EC2 EBS + EMC ScaleIO.

In general, persistency and containers go forward to personify to the point, and there are reasons for this. In the beginning, stateless processes and the methodological analysis of cardinal factors were trending. But altogether this is divorced from world. The community of interests understands that states exist, and the launch of stateful processes in containers, such as databases, is really relevant for the world of microservices. Each container is in some state, even if it is a "stateless" state. For example, in the memory there may Be such states as executable applications and open sockets, which you need to mould with in positive cases, for example, during data migration on the fly.

Data becomes crucial in a container environment. As you map the load happening the container architecture, you need to add data consistency. Integrity and availability will be necessary for services that keep up an national state. ADD to every last this also progressive storage. At the conference, various solutions to this problem were proposed, here are some of them:

• Crateful.io . A scalable distributed database that installs on an application server.
• ClusterHQ / Flocker. Heart-to-heart source-manager of container volumes, which allows you to containerise databases and other processes with state preservation, as wellspring As move them without worrying nearly the backend.
• Red Lid announcedon the integration of persistent memory for containers victimisation RHS (Red Hat Storage).
• Nutanix talked astir the Book API, which volition help ensure the continuity of containers on their platform.
• Portworx. This seems to be a new inauguration that will compete with ClusterHQ. The announcement says how the "part" of their political program will become public source.
• Kubernetes support for storage. Kubernetes updated the support on GitHub, which reflects the ability to use the Google Cloud persistant disk with k8s.

On the whole, it was a great DockerCon, full of fun events, swell guys and innovative technology. I hope we bequeath meet with you in Barcelona in November !