Cloud Boot Camp is a series of training, exercises, and
projects for HUIT to bring a team of admins and engineers up to speed and ready to
support a large, highly visible online Harvard application.
This repository holds the centralized documentation, planning, and coordination for
Cloud Boot Camp.
Most of the content for this repository will be captured in the wiki. It will also serve as an "index" to other repos -- the digital artifacts of this project-based training -- that will be captured here in th HUIT GitHub org. By capturing this in a central place, it'll be not only straightforward to support the platform, but will bootstrap other support efforts and resources in the future.
The projects this work is targeting are:
- Deploying a Fault Tolerant LAMP-based Service
- Downtime-Free Database Upgrades in AWS
- Front a Website with CloudFront CDN
- Automated Continuous Integration: From GitHub Commit to Running Test Service
- Comprehensive Monitoring, Instrumentation and Logging for a Dynamic Site
- Continual Resiliency Testing
Our goals with this initiative and set of exercises are:
- Develop an ability within HUIT to support a set of high-profile sites in AWS by July 26
- Create a cross-discpline team that can handle support concerns at varying levels in the "stack," and of varying complexity
- Engineer supporting services and refine current technology to reduce to a minimum the hands-on support needed for production sites
- Collaborately develop a support "playbook" for the sites in question
- Develop a model for supporting production sites in the cloud that will inform ongoing planning around CLoud & Devops in HUIT
Not only do we want to "catch up," at the end we want to lead the way!
This effort represents not only the support of a new platform, but also a new approach to IT. So to properly support the platform, you will need to change gears in the way you approach systems and software.
Many think of the cloud as a new type of service born of the need to reduce power for storage and servers, and to reduce cost. While that plays into it, the primary reason that cloud was developed is to enable automation & agility at the infrastructure level, the sort of things that a DevOps approach looks for. In fact, it more useful to think of cloud as the logical outcome to the pressures applied by DevOps. I.e. Cloud is a set of technologies that are built to implement DevOps Processes.
Given that, the natural way to use the cloud – and to support apps in the cloud – is with DevOps approaches and toolsets.
So ... time to switch gears.
Here is a list of tools you'll want to install and have available as we move forward. (This is not exhaustive, and perhaps not even suffucient. Feel free to contribute). Most of these work better on Mac or Linux, since over the past few years they have become the prefered OS's for devops engineering.
- Homebrew for Mac: Linux/Unix tools
- From Git or using
brew install git - SSH
- Python This is the runtime that the AWS tools are written in
On the Mac, Homebrew let's this installation happen quickly & easily. After installing homebrew, make sure that /usr/local/bin is the first in your path, then brew install ... the rest. Takes 3-5 minutes.
- Vagrant ... get the latest (>= v1.6)
- Virtual Box
- AWS commandline tools These are best installed from the command-line using
pip install awscli - Elastic Wolf GUI for visibility into your AWS resources.
- HipChat We'll use this for communciation and tracking changes to code and builds in real time.
- Mailing List: TBD
If you want to hang with the cool kids, take a look at these tools as well:
- MOSH The MObile SHell, let's you reconnect to SSH sessions without interruption. Use (brew to install)
- SublimeText
- Docker The next big thing in app deployment (built-in with Vagrant >= v1.6).
Each of these projects are intended to mimic or augment the application, release tooling, or supporting technology around the application (such as logging. monitoring, instrumentation, and resiliency testing). We break each aspect into self-contained components, define a useful exercise -- simplified in complexity but still transferable to the real world -- around each area. We'll then break into teams and implement.
The artifacts of this work -- all code, data, etc. -- will be captured into git repos within the GitHub org, along with accompanying documentation.
- Deploying a Fault Tolerant LAMP-based Service
- Downtime-Free Database Upgrades in AWS
- Front a Website with CloudFront CDN
- Automated Continuous Integration: From GitHub Commit to Running Test Service
- Comprehensive Monitoring, Instrumentation and Logging for a Dynamic Site
- Continual Resiliency Testing
- [Building a Control Plane for a Dynamic Site]
We will need more here ... please contribute!
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent