Arch9 - A cloud based continuous delivery implementation
Download as pptx, pdf0 likes395 views
Presentation shows the evolution of different configuration management tool from imperative cod on a mutable infrastructure through declarative code back to the imperative code with immutable infrastructure.
More Related Content
What's hot (19)
Viewers also liked (10)
Similar to Arch9 - A cloud based continuous delivery implementation (20)
More from Pavel Chunyayev (12)
Recently uploaded (20)
Arch9 - A cloud based continuous delivery implementation
- 1. A Cloud Based Continuous Delivery Implementation Pavel Chunyayev Amsterdam, 22-04-2015
- 2. Continuous Delivery in 2015 Development • Commit • Merge • Trigger pipeline Build • Create artifact • Assign version Testing • Journey across environments • Quality criteria Production • Canary deployment • Release != deployment
- 3. Continuous Delivery in 2015 Development • Commit • Merge • Trigger pipeline Build • Create artifact • Assign version Infrastructure • Infrastructure as a code • Automated • On-demand Testing • Journey across environments • Quality criteria Production • Canary deployment • Release != deployment
- 4. Declarative vs Imperative Declarative • I want this server in this state Imperative • I want you to apply this sequence of actions
- 5. Mutable (reusable) vs Immutable (disposable) Reusable • I want to use the same server during the life of the project. Immutable • I want to provision a new server when I need to change something.
- 6. Configuration Management tools Declarative Imperative Mutable Immutable #!/bin/bash
- 8. Case 1 Challenges • Projects require environments to be created • A lot of environments are similar • Takes time for admins to provision and configure them
- 9. Private cloud enables automation What we did • VMware vSphere to provide on-demand VMs • Puppet to configure environments • Hiera to separate data from the code • Jenkins to control the lifecycle Continuous Delivery • Huge time savings and code reuse • Developers are enabled to create configurations • Deployments are scripted Apache New server Java Zabbix agent MySQL
- 12. Case 2 Challenges • Control the state of the environments • Release new versions of applications safely • Automate the procedures to save time
- 13. Case 2 What we did • Puppet to control the state of the servers • Package artifact and assign version • Deploy specific version to specific environment • Rebuild the whole environment on deploy Continuous Delivery • Visibility • Repeatable process • Scripted deployments Production 1.3-127 Staging 1.3-140 Testing 1.3-156
- 16. Case 3 Challenge • Software needs to be tested in different configurations • Test environments are multi-server • Environments take days to provision manually
- 17. Environments as a service Solution • Amazon Web Services to provision infrastructure • Ansible to create procedures for each configuration • REST API service to coordinate requests Continuous Delivery • Ready for both human and robots • 20 minutes to create any environment • Reduced feedback time REST API Ansible
- 18. Environments as a service Declarative Imperative Mutable Immutable
- 20. Key takeaways • Optimize for Continuous Delivery • Automate infrastructure pavel@levi9.com @PavelChunyayev Any questions?
Editor's Notes
- #2: Hello, my name is Pavel Chunyayev, I’m a Continuous Delivery architect. With more than 10 years of experience, more than 5 of them I worked for Levi9 helping our clients to successfully run their applications in production as well as establishing and improving their Continuous Delivery pipelines in order to deliver software in time and of certain quality.
- #3: Let’s have a look at the CD process to aim for in 2015.
- #7: It just happens that configuration management is hard with bash scripts.
- #18: Pass configuration directly through service Any configurations are supported
- #20: Cloud For best case CD Automated, repeatable, on-demand infrastructure Automation Cycle time, feedback loop Change QA/Sysadmin Quality