Orchestrating Docker with Terraform and Consul by Mitchell Hashimoto

1. Orchestrating Docker with Consul and Terraform

2. Mitchell Hashimoto @mitchellh

3. HashiCorp Powering the software-managed datacenter. hashicorp.com

4. What’s the problem? “Orchestration” and why it is needed in a Docker world

5. Docker Solves a lot of Problems Packaging Docker Image Docker Registry Docker Host Distribution Execution

6. Other Pieces to the Puzzle • Infrastructure lifecycle and provisioning • Monitoring • Discovery • Orchestration • Transitioning to Docker

7. Infrastructure Lifecycle and Provisioning Datacenter Server Server Server Server Server Server Server Server Server DNS Database CDN

8. Infrastructure Lifecycle and Provisioning Datacenter Server Server Server Server Server Server Server Server Server DNS Database CDN • Docker hosts • Networking • Storage • Service providers (ex. DNS, CDN, etc)

9. Infrastructure Lifecycle and Provisioning Datacenter Server Server Server Server Server Server Server Server Server DNS Database CDN • Changing/Iterating • Rolling deploys • Canaries

10. Monitoring Container Container Container Container Container Container Virtual Virtual Virtual Physical (Machine) • Node • Container • Service • Propagation of this information Service Service Service

11. Discovery • “Where is service foo?” (database, web app, api, etc.) • Multi-datacenter • Service providers (Database as a Service, etc.) • Physical nodes + containers + non-Docker • Service configuration: “What is my configuration?”

12. Transitioning to Docker • Transition isn’t a single step, takes time • How to handle legacy + Docker and interaction?

13. Problems Not Inherent to Docker • Not new, not caused by Docker, will exist after Docker • Higher density exacerbates problems at earlier stage • New aspect: core datacenter functionality is shrinking! • Number of servers and containers will only grow, but the number of servers/containers we run/manage will shrink over time. • Infrastructure lifecycle, discovery, monitoring, orchestration problems still exist yesterday, now, and in the future.

14. Consul and Terraform Infrastructure lifecycle, service discovery, monitoring, and orchestration at scale for all infrastructures.

15. terraform.io

16. Build, combine, and launch infrastructure safely and efficiently. terraform.io

17. Terraform Goals • Unified view of modern datacenter with infrastructure as code • Compose multiple tiers (IaaS to PaaS to SaaS) • Safely change/iterate infrastructure • One workflow, technology agnostic • “No more dashboards.”

18. Infrastructure as Code DigitalOcean Droplet with DNS in DNSimple resource "digitalocean_droplet" "web" { name = "tf-web" size = "512mb" image = "centos-5-8-x32" region = "sfo1" } resource "dnsimple_record" "hello" { domain = "example.com" name = "test" value = "${digitalocean_droplet.web.ipv4_address}" type = "A" }

22. Infrastructure as Code • Human friendly config, JSON compatible • Text format makes it version-able, VCS-friendly • Infrastructure as code on a level not before possible

23. Compose Multiple Tiers / Providers

24. Providers • Integration point • Expose resources (“aws_instance”, “dnsimple_record”, etc.) • CRUD API • Core vs. Providers simplifies logic for integrators

25. Provider Composition • “Layer cake” • Provider for each layer • Unified configuration Container Container Container Container Container Container Virtual Virtual Virtual IaaS (OpenStack, etc.) • One step: “terraform apply” Physical (Machine)

26. Safely Change/Iterate Terraform Plan + digitalocean_droplet.web backups: "" => "<computed>" image: "" => "centos-5-8-x32" ipv4_address: "" => "<computed>" ipv4_address_private: "" => "<computed>" name: "" => "tf-web" private_networking: "" => "<computed>" region: "" => "sfo1" size: "" => "512mb" status: "" => "<computed>" + dnsimple_record.hello domain: "" => "example.com" domain_id: "" => "<computed>" hostname: "" => "<computed>" name: "" => "test" priority: "" => "<computed>" ttl: "" => "<computed>" type: "" => "A" value: "" => "${digitalocean_droplet.web.ipv4_address}"

29. Safely Change/Iterate • Plan shows you what will happen • Save plans to guarantee what will happen • Plans show reasons for certain actions (such as re-create) • Prior to Terraform: Operators had to “divine” change ordering, parallelization, rollout effect.

30. Other Features • Modules • Provisioners (remote-exec, local-exec, etc.) • Parameterization • Outputs • One command run: terraform run github.com/foo/bar

31. consul.io

32. Service discovery, configuration, and orchestration made easy. Distributed, highly available, and datacenter-aware.

33. Service Discovery Where is service foo?

34. Service Discovery Service Discovery via DNS or HTTP $ dig web-frontend.service.consul. +short 10.0.3.89 10.0.1.46 $ curl http://localhost:8500/v1/catalog/service/web-frontend [{ “Node”: “node-e818f1”, “Address”: “10.0.3.89”, “ServiceID”: “web-frontend”, … }]

35. Service Discovery • DNS is legacy-friendly. No application changes required. • HTTP returns rich metadata. • Discover both internal and external services (such as service providers)

36. Failure Detection Is service foo healthy/available?

37. Failure Detection

38. Failure Detection • DNS won’t return non-healthy services or nodes. • HTTP has endpoints to list health state of catalog.

39. Key/Value Storage What is the config of service foo?

40. Key/Value Storage Setting and Getting a Key $ curl –X PUT –d ‘bar’ http://localhost:8500/v1/kv/foo true $ curl http://localhost:8500/v1/kv/foo?raw bar

41. Key/Value Storage • Highly available storage of configuration. • Turn knobs without big configuration management process. • Watch keys (long poll) for changes • ACLs on key/value to protect sensitive information

42. Multi-Datacenter

43. Multi-Datacenter Service Discovery $ dig web-frontend.singapore.service.consul. +short 10.3.3.33 10.3.1.18 $ dig web-frontend.germany.service.consul. +short 10.7.3.41 10.7.1.76

44. Multi-Datacenter Setting and Getting a Key $ curl http://localhost:8500/v1/kv/foo?raw&dc=asia true $ curl http://localhost:8500/v1/kv/foo?raw&dc=eu false

45. Multi-Datacenter • Local by default • Can query other datacenters however you may need to • Can view all datacenters within one UI

46. Orchestration Events, Exec, Watches

47. Events, Exec, Watches Dispatching Custom Events $ consul event deploy 6DF7FE … $ consul watch -type event -name deploy /usr/bin/deploy.sh … $ consul exec -service web /usr/bin/deploy.sh …

48. Events, Exec, Watches • Powerful orchestration tools • Pros/cons to each approach, use the right tool for the job • All approaches proven to scale to thousands of agents

49. Operational Bullet Points • Leader election via Raft • Gossip protocol for aliveness • Three consistency models: default, consistent, and stale • Encryption, ACLs available • Real world usage to thousands of agents per datacenter

50. Consul and Terraform + Docker

51. Consul + Docker • progrium/consul (docker-consul) • registrator • Two options: Consul agent in or out of another Docker container. Both are easy, well supported.

52. Consul: To Docker or Not to Docker • Consul in Docker: Health checks must be network based due to isolation. • Consul on host: Can run health checks that enter container, invoke Docker for rich checks (file tests, pid tests, etc.) • Up to you, both ways work with the above caveats.

53. Terraform + Docker • Create all infrastructure components, not just Docker hosts: Load balancers, DNS, networking, storage, external service providers, etc. • Deploy Docker with static assignment (Docker provider, remote-exec), or schedulers (ECS, Mesos, Kubernetes, etc.)

54. Thanks! hashicorp.com