Continuous Delivery in Java - Things we learned hurting ourselves

Transcript

1. Continuous Delivery in Java Cose che abbiamo imparato facendoci del

   male Italian Agile Day ‘15

2. @pdincau Here I am

3. Continuous Delivery? What is

4. “Continuous Delivery is a software development discipline where you build

   software in such a way that the software can be released to production at any time. ” Martin Fowler* * martinfowler.com/bliki/ContinuousDelivery.html

5. “Continuous Delivery is nothing more than reducing the stress you

   get when you deliver business value to the customer ” My Team* * Come and ask!

6. n t Continuous Delivery... What is

7. Continuous Integration

8. Continuous Deployment

9. “Safety" Deploy rules

10. Technically Speaking

11. Deploy any specific working version to production with confidence

12. 3 ways

13. Rodeo Style

14. No tests Manual deploy Snowflake environments Non predictable outcome Few

    or very little collaboration

15. By the book

16. Fast Feedback Done is released Shared responsibility Everything is automated

    Repeatable and reliable process

17. Start small and iterate

18. Once upon a time...

19. We did know...

20. Previous deployment process was painful Database continuously change Different environments

21. Iteration n.1 No Automation

22. Environment replica on AWS Manual build Manual DB migrations Deploy

    with scp to remote JBoss

23. It was our own choice...

24. ...but we felt like this

25. No automation at all When was our last deploy?

26. A taste of Automation Iteration n.2

27. Set up a basic Jenkins installation Test and build triggered

    on push DB migration using Liquibase Task to deploy on AWS

28. This worked for a while

29. Errors in MyBatis DB mappers Wrong/Missing wirings

30. Yay! Environments are ready!

31. Deploy Everywhere Iteration n.3

32. New Jenkins job to run IT Track slowest tests First

    official deploy an all environments

33. …with different DB drivers !#$?!

34. Duplication

35. Incorrect use of mvn profiles N-builds for N-environments Long build

    and deploy time

36. Single Artefact Iteration n.4

37. Got rid of configurations using JBoss *Unique* artefact for all

    environments Deploy version X.Y.Z from Nexus

38. We started getting fast

39. But QA were not so happy Branches, branches everywhere!

40. Make them happy Iteration n.5

41. AWS is for nightly builds Tags over branches Track deployed

    version

42. Acceptance tests

43. Explained values and motivations to Product Owner

44. Remove Candidate Clean Up Clean Up Meet the pipeline Migrate

    Database Fetch Unit Tests Build Build Slow Tests Integration Tests Integration Deploy Publish UAT QA and Publish

45. Where to go next

46. Parallel builds Environments Provisioning

47. Things we would change...

48. Pipeline with “placeholders” Extract configurations Explain at the beginning the

    values behind

49. Things we would do again...

50. Build the pipeline iteratively Use of Liquibase Share pipeline responsibility

51. Recommended reading

52. That’s all!

53. Thanks @pdincau