Downloaded 33 times
Uploaded byIgor Moochnick
Continuous delivery workflow with Docker
This document discusses continuous delivery workflows using Docker. It covers topics like what Docker is, how to containerize applications and their dependencies, managing logs and failures, and sample tooling like Docker Compose and Jenkins for automating builds and deployments. Containerizing full stacks along with metadata and orchestration allows for repeatable, standardized environments that can be deployed across different clouds and machines.
More Related Content
Similar to Continuous delivery workflow with Docker
Continuous delivery workflow with Docker
- 1. By Igor Moochnick IgorShareConsulting igor@igorshare.com Continuous Delivery workflow with Docker
- 2. a) What isDocker? b) None at all c) Development d) CI/CD infrastructure e) Testing f) Production How are you using Docker?
- 3. Where do Icome from?
- 4. Use case: Real-Timehigh-volume data processing system running on any cloud and on-prem physical machines Datapre-processing: cleanup/sanitizationInject Push/pull Inject Push/pull Inject Push/pull Partitioning EventSourcing Sorting l l l l l l l l l l l l l l Processing Graph
- 5. • NetflixOSS /SpringCloud • Spark • Cassandra • Kafka • ELK Stack • Mesos • Jenkins • Docker Technology stack (some of the components…)
- 6. System Components inProduction (multi-host) l l l l l l Multi-host
- 7. Sandbox environments l l l ll Longevity/Failure/Integration/etc…
- 8. • Everything: ▫ Applications ▫Services. Ex: Nginx, Consul, Mesos, etc… ▫ Operational components. Ex: logstash, PuppetD, DataDog, NewRelic, etc… ▫ CI/CD infrastructure. Ex: Jenkins, Artifactory, Docker Registry, etc… ▫ And more … What is going into containers? Except …
- 9. • Everything else: ▫Data ▫ Secrets (Ex: tokens, keys, certificates, etc…) ▫ Configuration/Settings Unless versioned with the container ▫ Any other volatile piece of data that can change from environment to environment What is not going into containers?
- 10. Branching model forCD Deploy to Prod Deploy to Prod Main / trunk brunch1 brunch2
- 14. Release process asartifact https://github.com/constantcontact/jenkins_pipeline_builder Jenkins Pipeline Builder
- 15. Jenkins Workflow Plugin(newcomer)
- 16. Anatomy of aContinuous Delivery Release process Commit Acceptance Release Production Environment Code Artifact Artifacts Configuration Orchestration Orchestration Trigger Sandbox Environment
- 17. Commit + AcceptanceStage version = 0.0.1 ./gradlew build gradle.properties compile test Dockerfile.template Everything is a release candidate !!! ./gradlew e2e mysvc-0.0.1.jar ./gradlew docker mysvcdojo/mysvc:0.0.1 ADD {name}-{version}.jar Orchestration + health-check Tear-down Commit stage Acceptance stage Sandbox environment
- 18. • Make thecontents of containers unified ▫ As repeatable as possible across different types of applications ▫ Same ports ▫ Same file locations (logs, traces, configs …) ▫ Unified way of applying metadata Anatomy of a Dockerfile OS+Security+Environment = service environment service environment + service = operational service Dockerfile.template
- 19. Docker layering andinheritance hierarchy 102bb6bb1f36 ce70f38a21dd d36afbf40e7e 4b74a8a8f2fb 7af2e2ece3a9 c2f61748d146 529b51a5f129 d6bd0ffc4755 FROM img:1.0 EXPOSE 8080 EXPOSE 8080 ADD app-0.3.jar app.jar ENRYPOINT java –jar app.jar ENRYPOINT java –jar app.jar ADD app-0.1.jar app.jar ADD app-0.2.jar app.jar
- 20. • Environment variables •EXPOSE ports • Labels: https://docs.docker.com/engine/userguide/labels-custom- metadata/ Metadata
- 21.
- 22. Why not Docker-compose •Dependency-graph-aware • Validation checks gates • Scriptable / template-driven Dependency Check gate
- 23. • Most likelyit’s not a single container • It’s consist of: ▫ Application container ▫ Sidecar ▫ Log shipping ▫ Monitoring ▫ Etc… ▫ Sometimes: DB Memcache Etc… What is a containerized application? Pod
- 24. Log management Container 1 Container2 Container 3 Container 4 Log Volume Filebeat Logstash ElasticSearch Kibana 1) File-based logging Container 1 Gelf Logstash ElasticSearch Kibana 2) File-less logging Container 2 Gelf Container 3 Gelf Container 4 Gelf Worker Host Worker Host Log Volume Log Management Cluster Log Management Cluster
- 25. • Mesos scheduling ▫Marathon ▫ Chronos ▫ Custom scheduler Failure and Recovery
- 26.
- 27. Micro-services Dojo http://accordance.github.io/microservice-dojo/ Use it! Givefeedback! Contributions are welcome!
- 28.
Editor's Notes
- #4 3