Container Orchestration @Docker Meetup Hamburg
20 likes3,426 views
This document discusses container orchestration and provides an overview of various container orchestration tools and concepts. It describes schedulers that manage resource allocation and deployment of containers across clusters as well as tools for configuration management, service discovery, and maintaining a consistent cluster state. Examples of specific container orchestration systems like Mesos, Marathon, CoreOS, Kubernetes, and Docker libswarm are outlined.
Container Orchestration @Docker Meetup Hamburg
- 1. Container Orchestration Timo Derstappen @teemow
- 2. Why Orchestration? ● Microservice Architectures ● Polyglot Persistence ● Fault tolerance / High Availability ● Scalability ● Utilization
- 6. Overview ● Schedulers ● Configuration Management ● Service Discovery
- 7. Scheduler ● Resource management ● Decide which job runs where ● Monitor jobs ● Dependency resolution
- 8. Scheduler Concepts ● Google Borg (Monolithic) ● Apache Mesos (Two Level) ● Hadoop on Demand (Two Level) ● Google Omega (Shared State) Read: “Omega: flexible, scalable schedulers for large compute clusters” http://research.google.com/pubs/pub41684.html
- 11. Configuration Management ● State which container runs where ● Configure your containers ● Basis for service discovery
- 12. Consistent Cluster State ● Etcd (Raft) ● Consul (Raft) ● Doozer (Paxos) ● Zookeeper (Paxos) ● Spanner (Paxos) http://raftconsensus.github.io/
- 13. Service Discovery ● Ambassador Pattern ● Consul (DNS) ● Etcd ● Haproxy / confd
- 15. Docker libchan ● Link containers from everywhere ● Unix Socket ● HTTP2 / TLS ● Messaging / Raw TCP ● Like Go Channels over the network
- 16. Many Choices ● CoreOS ● Kubernetes ● Apache Mesos ● Consul / Serf ● Project Atomic / Geard ● Panamax
- 18. Docker libswarm ● Primitives of container orchestration ● With adaptors to actual implementations ● Docker Remote API abstraction ● Not yet ready for prime time
- 19. Mesos ● Primitives for distributed scheduling ● Meta Scheduler ● Writing your own scheduler is “easy” ● Docker Containerizer (since 0.20.0)
- 21. Mesos Apps
- 22. Marathon ● Framework for long running processes ● REST API / Webinterface ● Meta Framework
- 23. Marathon ● Constraints ● Health Checks ● “Service Discovery” / Load balancing ● Rolling update
- 24. CoreOS ● Small linux distribution to manage container clusters ● Automatic updates ● Using docker for packet management
- 25. CoreOS ● Fleet ● Etcd
- 26. Fleet ● Systemd for the cluster ● Simple scheduling via Metadata ● Resolves service dependencies ● Health Check / Failover
- 27. Kubernetes ● Container cluster management ● Created by Google
- 28. Kubernetes Concepts ● Pods ○ A group of containers on the same host ○ They share the same network namespace ○ Eg. presence, ambassador, logging, monitoring, caching, configuring sidekicks ● Labels ● Service ● Replication Controller
- 30. Inception anyone? ● Kubernetes runs on CoreOS ● Mesosphere works on Kubernetes integration ● Conair - run CoreOS/Kubernetes/Panamax components in containers
- 31. Thanks for listening! Timo Derstappen @teemow http://giantswarm.io We’re hiring!