Stateful Containers: Flocker on CoreOS
0 likes•290 views
The document discusses the integration of Flocker with CoreOS to manage persistent data volumes in Docker containers. It highlights key functionalities such as volume sharing, persistence across container deletions, and the process for creating and mounting volumes using Flocker. Additionally, it addresses challenges related to EBS volume limitations and the need for efficient cloud storage solutions in a distributed system.
Stateful Containers: Flocker on CoreOS
- 2. Hi my name is Stephen I’m a developer evangelist at ClusterHQ I was previously at Iron.io advoca<ng stateless workloads
- 4. docker docs… • Volumes are ini.alized when a container is created. If the container’s base image contains data at the specified mount point, that exis.ng data is copied into the new volume upon volume ini.aliza.on. • Volumes can be shared and reused among containers. • Changes to a volume are made directly. • Changes to a volume will not be included when you update an image. • Volumes persist even if the container itself is deleted. External Volumes
- 5. docker run -v /data ubuntu docker run -v /data2 ubuntu host /DATA /DATA2 Volumes inside a container
- 6. docker rm <container_1> host /DATA /DATA2 Volumes inside a container
- 7. Mount a host directory as a data volume docker run -v $HOME:/data ubuntu host $HOME/data
- 8. Mount a host directory as a data volume host docker rm <container_id> $HOME/data
- 9. Mount a host directory as a data volume $HOME/data host docker run -v $HOME:/data ubuntu
- 10. Mount a host directory as a data volume /data docker run -v /data ubuntu coreos-‐host-‐node—1 coreos-‐host-‐node—2
- 11. That container is not guaranteed to be scheduled to that host /data coreos-‐host-‐node—1 /data My data is still here! fresh host volume :( coreos-‐host-‐node—2 docker run -v /data ubuntudocker run rm <container_id>
- 12. That container is not guaranteed to be scheduled to that host /host/logs/container/logs coreos-‐host-‐node—1 /host/logs/container/logs My data is still here! fresh host volume :( coreos-‐host-‐node—2 copy host volume?
- 13. • Keep track of Host Volumes? • What if the volume is extremely large? • BitTorrent Sync? • Use NFS? GlusterFS? OtherDistributedFileSystemFS Problems to address
- 14. Distributed Synchronized Filesystems coreos-‐host-‐node—N… /data /data /data /data /data from what I have read so far • Network uniformity • CPU intensive • Maintenance
- 17. Limita<on of ~40 EBS volumes to a Linux Host before Boot Failures become a thing coreos-‐host-‐node—1
- 18. • MounKng correct ebs volumes to the correct containers/host machine • Provisioning EBS volumes happens manually not dynamically • you need to interact with the the AWS EBS API or Console Problems to address
- 19. Enter Flocker
- 20. coreos-‐host-‐node—1 ectd docker flocker Redundant Cloud Storage! demo:/data redis:latest docker run -d -v demo:/data --volume-driver=flocker --name=redis redis:latest CREATE
- 21. Redundant Cloud Storage! coreos-‐host-‐node—1 demo:/data redis:latest ectd docker docker run -d -e USE_REDIS_HOST=redis --link redis:redis -p 80:80 --name=app stephenitis/bird-counter:latest flocker bird-‐counter
- 23. Redundant Cloud Storage! coreos-‐host-‐node—1 demo:/data redis:latest ectd docker flocker bird-‐counter coreos-‐host-‐node—2 redis:latest ectd docker flocker docker run -d -e USE_REDIS_HOST=redis --link redis:redis -p 80:80 --name=app stephenitis/bird-counter:latest remount volume moby-‐counter
- 25. demo?
- 26. coreos-‐host-‐node—1 ectd docker flocker Flocker Create and Mount a Volume demo:/data redis:latest docker run -d -v demo:/data --volume-driver=flocker --name=redis redis:latest CREATE
- 27. coreos-‐host-‐node—1 demo:/data redis:latest ectd docker docker run -d -e USE_REDIS_HOST=redis --link redis:redis -p 80:80 --name=app stephenitis/bird-counter:latest flocker bird-‐counter
- 29. Flocker Magic coreos-‐host-‐node—1 demo:/data redis:latest ectd docker $ docker run -d -v demo:/data --volume-driver=flocker --name=redis redis:latest $ docker run -d -e USE_REDIS_HOST=redis --link redis:redis -p 80:80 --name=app stephenitis/bird-counter flocker bird-‐counter coreos-‐host-‐node—2 redis:latest ectd docker flocker bird-‐counter
- 31. how it works
- 32. Magic.
- 33. • All Flocker services to be Containerized • run our containers in “privileged” mode • When the operator executes docker run -‐-‐privileged • We need to to communicate the state of mounted volumes to the host system back to our containerized flocker service Problems to address to get this to work on CoreOS
- 34. coreos-‐host-‐master-‐node docker flocker-‐docker-‐plugin flocker-‐container-‐agent flocker-‐dataset-‐agent flocker-‐control-‐service docker run -d -v demo:/data --volume-driver=flocker --name=redis redis:latest redis
- 35. Installing the tools across all cores nodes was brainlessly easy with our uft-flocker tool $ uft-flocker-install cluster.yml $ uft-flocker-config cluster.yml $ uft-flocker-plugin-install cluster.yml $ # boom shaka laka Making installa<on of all the pieces easier
- 36. wrapping flocker in nsenter wrap commands that interact with host with nsenter `mount`, `umount`, `lsblk`, `losetup` https://github.com/jpetazzo/nsenter
- 37. nsenter it can enter existing namespaces, or spawn a process into a new set of namespaces. https://github.com/jpetazzo/nsenter
- 38. syncing /etc/mtab for `psutil` /etc/mtab psutil
- 39. Installing the tools across all cores nodes was brainlessly easy with our uft-flocker tools /etc/mtab psutil /bin/nsenter --mount=/host/proc/1/ns/mnt -- cat /etc/ mtab > /etc/mtab.tmp - mv /etc/mtab.tmp /etc/mtab
- 41. Thanks to ClusterHQ team Madhuri Yechuri / Luke Marsden /Kai Davenport CoreOS (for teaching us how our services would work in a dockerized format) Giant Swarm (for pushing us forward) Jérôme Petazzoni (nsenter _w)
- 42. demo instruc.ons at goo.gl/Wymh74
- 43. Q’s @stepheniKs [email protected] San Francisco | Bristol | Remote Hiring like ClusterHQ.com/careers/