CoreOS Overview

Transcript

2. About Me Architect CoreOS @bakins github.com/bakins
4. Why build CoreOS?
5. run and isolate apps containers
6. what is it exactly? containers
7. libc python django app.py
8. $ /usr/bin/python run app.py
9. libc python django app.py
10. libc python django app.py example.com/myapp
11. $ container fetch example.com/myapp $ container run example.com/myapp
12. pid ns isolated pid 1
13. user ns isolated uid 0
14. network ns isolated netdev
15. mount ns isolated /
16. cgroups manage resources
17. cgroups count resources
18. cgroups limit resources
19. docker engine
21. google lmctfy cloud foundry garden mesos containers lxc systemd-nspawn
22. how are they created? containers
24. super-powers containers
25. App independence from the OS. System to get container to the server. Resource isolation between apps.
26. reduced API contracts OS
27. kernel systemd etcd ssh docker python java nginx mysql openssl app distro distro distro distro distro distro distr
29. python java nginx mysql openssl app distro distro distro distro distro distro distr kernel systemd etcd ssh docker
30. python openssl-A app1 distro distro distro distro distro distro distr java openssl-B app2 java openssl-B app3 kernel systemd etcd ssh docker
31. manual updates
32. automatic updates
33. automatic updates
34. atomic with rollback auto updates
37. super-powers OS
38. Opportunity for automatic updates. Consistent set of software across hosts. Base OS independent from app.
39. Slight Detour cloud-init ignition toolbox
40. design for host failure clustering
42. /etc distributed
43. open source software sequentially consistent exposed via HTTP runtime reconfigurable
44. -X GET Get Wait -X PUT Put Create CAS -X DELETE Delete CAD
45. Available
46. Available
47. Available
48. Unavailable
49. Available Leader Follower
50. Leader Follower Available
51. Leader Follower Temporarily Unavailable
52. Leader Follower Available
53. super-powers etcd
54. Share configuration data across hosts. Resilient to host failures. Designed for consistency across hosts.
55. getting work to servers scheduling
56. You Scheduler API Scheduler Machine(s)
57. $ cat foo.service [Service] ExecStart=/usr/bin/sleep 500 $ fleetctl start foo.service Job foo.service launched on e1cd2bcd.../172.17.8.101
58. while true { todo = diff(desState, curState) schedule(todo) }
59. while true { todo = diff(desState, curState) schedule(todo) }
60. while true { todo = diff(desState, curState) schedule(todo) }
61. while true { todo = diff(desState, curState) schedule(todo) }
62. fleet mesos kubernetes swarm job scheduling
63. locksmith coordination
64. super-powers scheduling
65. Think about app capacity first. Take advantage of compute resources. Build for resilience to host failure.
66. skydns, discoverd, confd service discovery
67. magic proxies service discovery
68. overlay network
69. OS Containers Cluster Configuration Job Scheduling Service Discovery
71. Kubernetes Namespaces Containers Pods Replication Controllers Label Queries Services