1 / 36

Yale Build and Deploy

Yale Build and Deploy. E Camden Fisher Yale University ITS. Sneak Preview. Who am I?. E Camden Fisher Technical Lead Unix Infrastructure and Virtualization camden.fisher@yale.edu @fishnix. What is Jenkins?.

mjeanette
Download Presentation

Yale Build and Deploy

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Yale Build and Deploy E Camden Fisher Yale University ITS SneakPreview

  2. Who am I? E Camden Fisher Technical Lead Unix Infrastructure and Virtualization camden.fisher@yale.edu @fishnix

  3. What is Jenkins? Jenkins is an application and a framework that manages and monitors the execution of repeated tasks.

  4. Who Uses Jenkins @Yale?

  5. Overview Before Jenkins After Jenkins "fire and forget", consistent! Testing is automated! Code coverage is easy! Bugs caught early and often! Devs worry about code! Change control in the right places for deployments! Rapid progress! Agility! Identical artifact per environment! Identical configuration per environments! Deployments are “click” easy Integration Nirvana • Building is slow, error prone • Testing is onerous • Code coverage is onerous • Bugs caught later • Devs worry about servers • No change control for deployments • Slow progress • Different artifact per environment • Inconsistent configuration per environment • Deployments are "hard" • Integration Hell

  6. Let’s Not Forget Continuous Integration! • Maintain a Single Source Repository • Automate the Build • Make Your Build Self-Testing • Every Commit Should Build the Mainline on an Integration Machine • Keep the Build Fast • Test in a Clone of the Production Environment • Make it Easy for Anyone to Get the Latest Executable • Everyone can see what's happening • Automate Deployment • Avoid "Integration Hell!!”

  7. Why did we chose Jenkins? • Easy! • Extensible • Scalable • Flexible • Open Source • Supported!

  8. Yale History Lesson Application lifecycle is a progression • Source Code Management • Maven and Artifactory • Building and Testing with Jenkins • Container Configurable artifacts • Runtime Configurable Containers • Managed deployments with Jenkins

  9. SCM

  10. MVN Maven • Project Object Model (POM) • Simplifies dependency resolution ("oops I forgot that .jar!") • Makes the build process easy and uniform Artifactory (Maven Repository) • Where do I put my built artifacts? • Makes it easy for everyone to get the latest build!

  11. Builds: Before Jenkins • We have a single source repo + a place to store built artifacts, but….. • Builds still take a long time • Testing takes longer! • Code coverage take even longer! • Build environments are not standardized • Mistakes are caught later, hard to debug

  12. Builds: Enter Jenkins • SCM commits automatically kick off a build • Testing and code coverage is automated and is run on every commit. • Broken builds immediately notify the team and the committer • Tests run in a Clone of the Production Environment! • Everyone can easily see what's happening! • Developers can get back to coding instead of building and testing. • Releases are quick and easy

  13. Deployment: The problem. • Configuration must be different per environment • Choice between loss of control or loss of agility • Often devs configure the container • Either devs can edit deployables or they can't • If they can... things are bad • Code is deployed, edited and removed w/out Change Control • Moving quickly… "I'll clean it up later!" • Dev environments quickly diverge from Production • Security is compromised • If they can't... things are worse • Change requests are "slow” • Systems groups must do everything

  14. Deployments: How do we fix them? • Standardize the process! • Promote SANE change control • Normalize deployables/artifacts • Single location for app configuration data • Standardize the container • Externalize what makes an environment unique and special

  15. Yale Application Installer Plugin • Standardize the process! • Why? • SSH creds + Delegation • Manage keys outside • Node name substitution • How? • Installer is co’d, scp’d • install.properties on dest • Pre-Install • SSH exec installer • Post-Install

  16. Container Configurable Artifacts • Artifacts were built with embedded configuration. • ie. datasources, CAS service endpoints, etc • Artifacts are different per environment! What!? • "Oops, I forgot to update that parameter!" • Externalization of configuration parameters • deployable XML • Apps self configure with JNDI • Now we have Container Configurable artifacts • The SAME artifact migrates between environments • XML configuration data (stored as build parameters in Jenkins) is all that differentiates environments

  17. Yale Standard Java Container • JBoss EAP (5.1 now) • Minimal customization to externalize configuration into runtime • Packaged in RPM • Configuration Management to install + manage what’s “special” • Meets 100% of use cases (so far) • Runtime configurable container is key!

  18. Yale Standard Java Container $JBOSS_HOME/deploy/jbossweb.sar/server.xml <!-- Connector for SSL 20100624 ECF --> <Connector protocol="HTTP/1.1" SSLEnabled="true" port="${jboss.bind.httpsport}" address="${jboss.bind.address}” secure="true" clientAuth="false" scheme="https” proxyName="${jboss.proxyname}" proxyPort="${jboss.proxyport}” SSLCertificateFile="${jboss.server.home.dir}/conf/server.crt" SSLCertificateKeyFile="${jboss.server.home.dir}/conf/server.pem” URIEncoding="UTF-8” SSLProtocol="TLS”> /etc/init.d/jboss_nodexx # JBoss variables export JBOSS_OPTS=“-Djboss.bind.httpport=${HTTPPORT} \ -Djboss.bind.httpsport=${HTTPSPORT} ${ADDLJBOPTS}” ${JBOSS_HOME}/bin/run.sh ${JBOSS_OPTS} -c ${JBOSS_NODE} -g ${JBOSS_PARTITION} -b ${JBOSS_BIND} -u ${CLSTR_ADDR}

  19. Deployments: The result. • Consistency! • Jenkins writes configuration XML: $human_error-- • Eliminate shells, and elevated privileges on servers • Container is managed by infrastructure with the O/S • Developers can deploy to DEV at will • Empowers developers to GTD • Puts gates at appropriate places • Changes to the Jenkins jobs, containers + deploys to Test/Prod require change control • Frees Systems folks to work on more interesting things • Standard Container + Cloudy IaaS + Container Configurable artifacts = Vendor-lock-in-less PaaS!

  20. Drupal Continuous Deployment • Yale has a large shared Drupal infrastructure • Shared = needs change control • Migration process looks similar to Java • Some end users want to edit themes! • Jenkins to the rescue!

  21. What’s New? • Workflow integration through web services • Build + Release of Apache Servicemix bundles and “features” • Spawning and Destroying Servicemix child instances • Deployment of Features and OSGI bundles to Apache Servicemix • Deploying Drupal 7 • git + pull

  22. Other Languages • Php • Code coverage and unit testing available • Ruby • Automated unit testing • Automated code coverage • Automated deployment coming soon • .Net • Build, unit test, archive creation

  23. The Future • RBAC + folders • delegate responsibility to other systems groups • More ruby • unit testing and code coverage • deployments • Enterprise Service Bus • Centralized SSH mangagement • the tools are better now • Managing/Provisioning with cfgmgmt • Cloud deployments

  24. Questions? E Camden Fisher camden.fisher@yale.edu @fishnix

More Related