Implementing Continuous Delivery with Chef

my #chefconf 2014 presentation

Transcript

1. Implementing Continuous Delivery in Chef
2. Why am I here?
3. Opinionated
4. and yet…
5. Options
6. @thommay
8. Why Continuous Delivery?
9. Continuous is a side effect
10. Consistency
11. Correctness
12. Reproducibility
13. Repeatability
14. Why Chef?
15. Availability
16. Has all the bits we need
17. Let’s get started
18. Out of the box
19. Production quality?
20. The deploy resource
21. git clone …
22. Callbacks
23. Duplicates work
24. Reproducible?
25. Application cookbooks
26. :(
27. Assumptions galore
28. More duplicate work
29. Slow
30. Recap
31. Pipelining
32. CD can’t work in isolation
33. Continuous Integration
34. build once
35. test what you deploy
36. Packages
37. (I don’t care about Windows)
38. leverage your OS
39. easy to get a repo up
40. fpm
41. but…
42. atomic
43. config files
44. Something Custom
45. “artiballs”
46. package-lite
47. tarball
48. chef-solo
49. pretty similar to omnibus
50. “deployable”
51. a manifest file
52. some number of: artifacts
53. some number of: config file templates
54. generate resources
55. Honourable mentions
56. docker/lmctfy chef-metal decker Flynn ProjectAtomic Mesos/Marathon
57. Recap
59. baton
60. https://github.com/digital-science/baton
61. amqp messaging
62. not mcollective
63. build agents trivially
64. Chef Push Jobs
65. require on-premise enterprise chef
66. I don’t have one
67. looks nifty though
68. AutoScaling Groups
69. “a/b” deploys
70. add new ASG
71. grow new one
72. shrink old one
73. Bamboo
74. deploy tasks
75. Recap
76. Conclusions
77. Any Questions? Images: Paula Marina - Pipeline! https://flic.kr/p/4V3wLC Kihong Kim - Orchestra Miniature https://flic.kr/p/dM32mX