July 17, 2015

Jenkins - Getting Jenkins Cloud Ready - Part 1

Jenkins is a continuous integration tool for software development. You can define jobs, integration and deployment and schedule their execution. It's one of the most important tools in software development. The developer does not have to wait for a feedback but will be informed with detailed infos on how his new feature has fared inside the environments.

Jenkins tend to get bulky in time. Projects have to set up their unique demands, e.g. software versions and installations (Java 6-8, Tomcat, JBoss, Websphere).  As a result, a master-slave setup can ease the complexity of setting up Jenkins but setting up the master-slave environment itself remains.

The perfect solution is the cloud where you can setup any complex environment with containers and setup templates. Especially the Google Compute Engine looks interesting as you pay for use and you do not have any setup costs.

Therefore, I aim to define a setup template for cloud deployment and then explode my setup into the cloud. Which means setting up a secure Jenkins master with a scalable amount of build slaves. This will enable teams to setup a build environment per project rather than per team or company. The setup can be deployed and removed on demand and enable indefinite staging environments for software development.

The first step is building the appropriate cloud container using Docker. Docker is a container framework where you setup applications inside a linux container and being able to fire it up in any cloud environment. Jenkins is by its history an on-premise solution and setup is usually done by hand. A Jenkins instance is not by default able for deploy-and-use inside the cloud. Appropriate means:

  • The container must include a ready-to-use Jenkins.
  • The container must have a default security setup.
  • The container must accept slave connections.
  • The container must support backup & restore.

In my Github repository and project blacklabelops/jenkins I implemented a cloud-ready Jenkins master inside a Docker container. The project has elaborate documentation, feel free to roam and try it out. Managing Jenkins inside Docker required the following steps:

  1. Setting up a Dockerfile that installs Jenkins, Java and required software packages.
  2. Extending the Dockerfile by configuring Jenkins to use my pre-defined Docker Volume for persisting its settings and using it as the default workfolder.
  3. As a result I could write bash scripts in order to backup & restore the Docker Volume. Now I was able to resume my immutable container in any state.
  4. Define environment variables to be able to configure my container for specific use cases. Being able to control security setup, ports for slave connection and plugin installations.
  5. Environment variables are interpreted at startup and they use the docker-entrypoint script for applying their intend. The script itself is hooked inside the Dockerfile.
  6. Implementing several Groovy Scripts to execute the settings inside Jenkins, e.g. settings for security and the java virtual machine. Blacklabelops Groovy Scripts.
The container is build and deployed by the public Docker registry Docker Hub and checked by Continuous Integration by Circle-CI. Docker Hub and Circle-CI are free for public Github Repositories. You can always check my repository for the current state of my project as build badges provide a good overview of my implementation state.

Dockerhub Build Badge: 
Circle-CI Build Badge:


Dockerhub Info Badge:

This is now my cloud-ready Jenkins! I have already over 600 downloads, several clones of my repository and first positive feedback. Feel free to try! I would be happy to get some stars and comments on my Github and Dockerhub repositories!

The next article will be about deploying this container inside the Google Container Engine and I will have a look at the costs of its execution. Also I will have an article about securing Jenkins on the network layer with https. Last article will be about deploying and scale slaves in addition to Jenkins. Stay tuned!

Interesting? I also have a Jira cloud container! blacklabelops/jira

June 18, 2015

Github - Automatic VM Creation with Packer and Virtualbox


You always wanted to spin up VM's from latest available ISO images automatically? You want full control of what's installed on your VM? You like CentOS? Then this is the perfect place for you!

June 13, 2015

Release and Version Vagrant Boxes with Atlas

Vagrant is an amazing flexible management tool for VM (Virtual Machines). You can easily build, run and deploy Virtualbox, VMWare, AmazonWS and more VMs. Atlas is a platform for uploading and managing Vagrant boxes. I can build and upload new boxes and download my boxes to any environment and deploy them when needed.

I have several vagrant boxes under configuration and adjust them when necessary. New boxes will be rolled out and each box is checked daily by a nightly build. Old version will be still available and replaced seamlessly. In this tutorial I want to show how I implemented a release cycle for Atlas.

The release cycle is described in this Tutorial.

Project and code can be found in the Github repository blacklabelops/packercentos.

June 12, 2015

BlackLabelOps Automatic VM Creation

I finished my first Packer project! It creates a Virtualbox CentOS 7.1 Image for Vagrant on the fly!

Repositories:
I use this project as a base box for:
  • Creating a Vagrant base box for my examples.
  • Creating a Docker base image for my examples.
  • Testing versioning and releasing of VMs on Github and Atlas.
  • Testing continuous integration with Packer.

April 27, 2015

SAFe isn't Safe

May be you have heard about SAFe, may be you haven't. Does not really matter:

"SAFe tells corporate dinosaurs that those little mammals downstairs are nothing to worry about. Agile, on the other hand, thinks mammals are the answer."

The Article: Issues with SAFe

Coding Tutorials ;)

February 28, 2015

Cross Functional Teams in Munich

A good article by Stylight on their approach on cross functional teams!

February 11, 2015

Page Listing Technology Stacks: Stackshare.io

Discover what top companies are using Learn about the tech stack behind some of the world's best software companies. See which tools and services they use, and how they're using them. Stackshare.io

The Three Rules of the DevOps Game

New essay by Mary Poppendieck: The Three Rules of the DevOps Game

January 18, 2015

Non-Proper Problem Solving

If this happens to you then you just need help:

January 13, 2015

Lean Startup Board Game

I really liked the Kanban Board Game, hope this Lean Game will become a reality:

January 12, 2015

"Lean Startup" - Leicht verständlich

Ein sehr guter Foliensatz, der Lean Startup anschaulich anhand von Startups, Unternehmen und Enterprises darstellt: Lean Startup - Auch in erwachsenen Unternehmen?

January 10, 2015

Technical Design Tradeoffs

Development teams should not think of it as making technical design decisions but managing tradeoffs. There is no optimal decision and the tradeoff does not automatically lead to technical debt.

Why Deployment Freezes Don't Prevent Outages

Think about this, every time you say "Never touch a running system!":

The Tyranny of the Ever Decreasing Timebox

A classic reference, a must have read in Agile Development: The Tyranny of the Ever Decreasing Timebox - David J Anderson

December 17, 2014

ThoughtWorks Technology Radar

ThoughtWorks has started a monthly review on technologies: Platform, Frameworks etc.
The PDF format is perfect! Technology Radar PDF

November 04, 2014

Problem Altsoftware

Der folgende Zitat aus dem Artikel fasst das Problem "Altsoftware" sehr gut zusammen:
"Die unterschiedlichen Web-Apps ermöglichen den Einsatz verschiedener Technologien innerhalb des Java-Universums. Wirklich genutzt wurde dies jedoch nicht. Im Gegenteil: Bewährtes wurde gerne wieder verwendet und übergreifende Konzepte in geteilten Bibliotheken hinterlegt. Eine zunehmend enge Verzahnung war das Ergebnis. Kleine Änderungen wurden aufwändig. Innovationen, wie z. B. eine Java-8-Einführung oder ein Spring-4-Update, erzeugen viel zu hohe Aufwände. Solche Themen dann zeitnah oder „nebenbei“ anzugehen, fällt nahezu aus. Unzufriedenheit über die veraltete Software und ein Verlust an „Spaß bei der Arbeit“ sind die subtilen wie ernstzunehmenden Konsequenzen.

Darüber hinaus traten weitere Symptome monolithischer Anwendungen auf: Eine technisch enge Kopplung von fachlich entkoppelten Themen resultiert in steigenden Lern- und Einarbeitungskurven bei den Entwicklern. Die Konsequenzen einer Änderung abzusehen wurde stetig schwieriger. "

September 29, 2014

August 10, 2014

May 30, 2014

Seen on Twitter