Continuous delivery of Mesos and Docker import articles

650) this.width=650; "src=" http://s5.51cto.com/wyfs02/M02/89/EF/wKioL1gig_XheDNfAAXo2WPPVMw355.jpg "style=" height : auto;vertical-align:middle;border:0px; "title=" 2015100813410005937156.jpg "alt=" Wkiol1gig_ Xhednfaaxo2wppvmw355.jpg "/>

The word change is commonplace in today's digital age, and it technology will revolutionize every time, from WEB2.0, virtualization, cloud computing, Big Data, microarchitecture, DevOps to today's container Docker and Mesos.

docker facilitates the testing, deployment, and upgrading of applications, which package various applications and the running environment on which they depend, into standard container/image, and then publish to different platforms to run. Docker's lightweight, fast-moving, easy-to-migrate feature facilitates devops, borrowing containers and allowing developers to easily integrate into the product delivery process.

mesos is the best practice for software-defined data centers, and its philosophy, in the most popular words, is to allow operations personnel to operate the data center, even if the operation of a server, the data center in the CPU, memory, storage and other resources in a single server for scheduling and management. Sounds very big on, if the technical language to describe, Mesos is defined, a unified cluster resource management and scheduling platform, the production environment in a variety of service framework, deployed in a common cluster, sharing the resources in the cluster, by Mesos to the resources of the unified scheduling, provided to the service framework for use. The advent of Mesos provides a great convenience for the management of IaaS, PAAs, and operations.

In practice, Mesos and Docker are the best companions, the former provides a unified management of resources, the latter to achieve the isolation of the use of resources, there are points, at different levels to play a different performance. At the same time, Mesos and Docker have their own technology ecosystem, both of which promote and promote each other's ecological circle.

§marathon guarantees a long-running framework of services on the Mesos, ensuring that services running on Mesos are running and automatically migrated to other servers when a Mesos host fails.

§chronos service loop Run job scheduling framework, can be set repeatedly at a specific time to trigger the operation of the Mesos service

§mesos-dns the DNS-based service discovery mechanism for Mesos, it is convenient to find the container running location and manage the dispatch.

§kubernetes cluster-level container Orchestration management system, can easily manage the container application running across the cluster, provide container scheduling, deployment, service discovery, extension mechanism and other functions

§swarm tools for creating clusters of Docker hosts (servers running the Docker daemon), using swarm to operate the cluster, will make the user feel as if they are operating on a single host.

The best companion for Mesos and Docker is a great convenience for the continuous delivery of applications, providing great convenience for operational management of applications, both of which play a critical role in continuous delivery solutions, such as demonstrating the practice of Mesos and Docker in application continuous delivery. Its specific process can be described as follows:

o The developer of the product according to the requirements of the implementation of code, the implementation of the code to submit to GIT/SVN and other code management warehouse;

o The tester compiles, packages, integrates, and so on, through continuous integration tools such as Jenkins/hudson, and packages the code into Dockerimage submitted to the Docker image Warehouse;

o The tester deploys the test environment with the new dockerimage through the Mesos resource scheduling and marathon platform, and the tester validates the product requirements functionality in the test environment. If you do not meet the validation feedback to the developer, the developer continues to fix it and notifies the OPS personnel to deploy the code to the pre-release or production environment if it has met the requirements;

o the operator receives a notification that the test has been validated, and deploys the newly released Docker Image to a production environment managed by the Mesos Dispatch;

o during the operation of production environment, operation and maintenance personnel can through the marathon and other platforms to the environment for the expansion of the business and the reduction of capacity, and even some failure of automatic recovery management.

650) this.width=650; "src=" http://s5.51cto.com/wyfs02/M02/89/EF/wKioL1gig_2iZR-kAADYo28oKI0046.jpg "style=" height : auto;vertical-align:middle;border:0px; "title=" 4.jpg "alt=" Wkiol1gig_2izr-kaadyo28oki0046.jpg "/>

Mesos and Docker Delivery application cases

The above process description is based on the comparison of traditional standards of organizational structure of the company to describe a clear role in the division of development, testing, operations and other role personnel; Some companies have already practiced devops management, development and operation of the integrated, the organizational structure of the role division is more simple and concise, its use of Mesos, The continuous delivery process for Docker implementations will be somewhat similar, with the exact procedure being the same, except that the person who operates may be different.

The advent of Mesos and Docker application of fast, continuous, automated delivery to the ground eliminates a lot of barriers, bringing great convenience. It technology is constantly changing, we can choose only a constant positive embrace. Tomorrow will be better, this is our application of continuous delivery series of articles, in the follow-up will continue to update this series according to practice, together accelerate the Internet agile operations.

Continuous delivery of Mesos and Docker import articles