How to build an HA container orchestrator infrastructure for production – Giulio De Donato
Download as PPTX, PDF0 likes686 views
The document discusses container orchestration tools including Docker Swarm, Kubernetes, DC/OS Mesosphere, and Amazon ECS. It provides descriptions of key features such as how each handles tasks, services, volumes, networking and load balancing. It notes that Swarm is simplest for a few nodes but lacks a UI, Kubernetes is widely used and supported but more complex, DC/OS provides a good UI and supports multiple teams, and ECS is very easy to use within AWS but doesn't provide full orchestration capabilities. The document aims to help understand the different approaches taken by these orchestration platforms.
How to build an HA container orchestrator infrastructure for production – Giulio De Donato
- 1. Nome Speaker@twitter Agile innovation Giulio De Donato @liuggio
- 3. Today’s path devops Container orchestrators principles DEMO THE BEST CONTAINER ORCHESTRATOR And how to make easy money Talks end here
- 4. devops
- 5. Devops dev and ops
- 6. devops
- 7. devops is about dev and ops devops is from to .
- 8. Devops misconceptionsDEVOPS is not a verb
- 9. New brain :)
- 13. IDEA Dev and Ops DEVELOPMENT PRODUCTION Code Infrastructure DOCKER IS EASY!!! ... CONTAINER ORCHESTRATION IS HARD
- 14. Docker-era killing feature ● Sharing infrastructure. ● Infrastructure as code also for dev/stage. ● Infrastructure as code derived from dev. ● Kills the “Worked on my computer” syndrome. ● Shorter Lead Time. ● Simple automation. ● Enable Immutable Pipeline ● Stateless ● Faster
- 16. This APP should be replicated at least 3 times. The APP should communicate over An overlay network called backend. The Simple Containers Needs APP (nginx) This APP need 3Gb Ram, 1 core, 20GB of SSD. The APP must be load-balanced internally on 9000, and health-checked on port 80 load-balanced. All of the logs of my AP should go there. APP (nginx) APP (nginx) 80 9000
- 17. Container Orchestration does Karl di mesosphere: https://www.youtube.com/watch?v=C_u4_l84ED8
- 18. Team needs App with source code must be built Dev ~= Staging env ~= Production Env Container must be Built/Updated/Pushed Production must be updated with no Downtime only if the App is healthy in the best node. Automatic Rollback Log and Alert Autoscaling Is the system healthy? Which version is running? Managed/non-managed services, networking. Security
- 19. Container Orchestration Application Development lifecycle Source repo, CI, Artifact repository Container lifecycle Container builder, Image repository Container Orchestration Scheduling, Resource management, Service management Day 2 Operations Debugging, Monitoring
- 20. Docker Swarmkit ● Facilissimo da installare ● Velocissimo ecosistema ● Velocissimi cambiamenti ● Facile per pochi nodi ● Stessi tool di dev ● Swarm si occupa solo di clustering e scheduling ● Solo Command Line / no UI ● Non appesantisce la computazione
- 23. kubernetes ● Opinionated ● Usatissimo ● Ecosistema floridissimo ● Supportato da Google, Redhat ... ● Approccio dev ● Ben integrato con i maggiori cloud provider ● Forte storico an open source system for automating deployment, scaling, and operations of applications
- 24. DC/OS mesosphere ● In produzione da mostri sacri ● Package manager ● UI ottima ● Supporta piu team ● Anche Big-data ● Espone Servizi interni (eg. Jenkins abilità l’OAuth2.0) ● Complesso ● Folders DC/OS is based on the production proven Apache Mesos distributed systems kernel, combining years of real-life experience with best practices for building and running modern applications in production.
- 25. Amazon AWS ECS AWS ECS ● Simple/Powerful ● Canary deployment ● Chip ● Very easy to work in the AWS ecosystem ● Fast ● Not OS :( ● (No networking, no Consensus, no computation for the orchestration)
- 27. Server Docker Swarmkit Container Server1 Container Redis Task==container Service == group of task of the same image Network Volume Route mesh volume volume user admin
- 28. agents managers Docker Swarmkit Server Container Server1 Container Redis Task==container Service == group of task of the same image Network Volume Route mesh volume volume Port: 80 networks Replica:2 user admin a p i
- 29. Kubernetes Container server2 Container Server1 Container Redis Pod == group of containers that shares volumes and … Service == Abstraction and access ReplicaSet == Handle the replica count Volumes! volume user admin
- 30. Kubernetes Container server2 Container Server1 Container Redis Pod == group of containers that shares volumes and … Service == Abstraction and access ReplicaSet == Handle the replica count Volumes! volume Port: 80 Replica:2 user admin a p i LOAD BALANCER agents managers
- 33. DC/OS mesosphere Server Container Server1 Container Redis folder Service == group of container/command/script of the same type Network Volume Internal VPN volume Load balancer is a Containe r DNS/networks/VIP user admin agents managers a p i
- 34. AWS ECS
- 35. Service TASK::v1 Task==group of containers Service == a specific tag version, scaled, e load bal. Cluster a group of services. TASK::v2 containerB containerA ECS cnt === docker cnt Volumes Link and aliasing (still) No networks (AWS VPI) Task is a definition of cnts,volumes,constraints Task has a version AUTOSCALELoad balancer (target group) Load balancer Cloud watch Scale AWS- ECS Service is mostly about deploy, scaling task, and reconfigure load balancer.
- 41. Ma...
- 42. Nome Speaker@twitter ●xpeppers.com/radar ● docker-italia.slack.com docker- italia.herokuapp.com ●We are hiring great people Giulio De Donato @liuggio
Editor's Notes
- #5: Riusciamo a dare una definizione di devops?
- #6: Riusciamo a dare una definizione di devops?
- #7: Riusciamo a dare una definizione di devops? non c’è una risposta definitiva, cisno molte opinioni su cosa c’è sotto il tetto del devops e cosa no. è una cultura? è una job title? è un modo di organizzarsi o solo un modo di pensare, Possiamo anche pensare ad un movimento che sia in evoluzione, quindi non fermiamoci troppo nel dare una defizione adesso, Invece possiamo parlare di temi comuni, strumenti e idee. Nato con l’idea di migliorare il service delivery agility, il devops enfatizza la comunicazione, la collaborazione e l’integrazione tra software developer e IT operations. Piu che guardare questi in silos separati e paralleli, il DevOps riconosce che l’interdipendenza del software dev e IT operation e aiuta una organizzazione al fine di render A perfect storm of converging adjacent methodology including Agile, Operations Management (Systems Thinking & Dynamics), Theory of Constraints, LEAN and IT Service management came together in 2009 through a smattering of conferences, talks and Twitter (#devops) debates worldwide that eventually became the philosophy behind DevOps. Agile software development paved the way, steering away from the waterfall method of software development toward a continuous development cycle. Ma senza includere le operation side so while development could be continuous, deployment was still waterfall-oriented. In a DevOps environment, cross functionality, shared responsibilities and trust are promoted. DevOps essentially extends the continuous development goals of the Agile movement to continuous integration and release. In order to accommodate continuous releases, DevOps encourages automation of the change, configuration and release processes.
- #8: Certo idea money suonava male
- #10: Ok ci siamo puliti un po’ il cervello da alcuni concetti che Ora prendete tutti questi concetti e buttateli via perche si parla di orchestratoriii uuiiiiii
- #11: Docker è facile
- #12: Golang -> mysql
- #13: Golang -> mysql
- #14: Al this and more with a fucking new unstable tool
- #15: Al this and more with a fucking new unstable tool
- #16: Golang -> mysql
- #17: Definizione per un orchestratore è dichiarativa
- #18: Definizione per un orchestratore è dichiarativa
- #19: Si ma chi copre allora i bisogni del team?
- #21: Perchè solo quelli opensource? Perche sono i driver dell’ecosistema.
- #22: Definizione per un orchestratore è dichiarativa
- #23: Golang -> mysql
- #24: Perchè solo quelli opensource? Perche sono i driver dell’ecosistema.
- #25: Perchè solo quelli opensource? Perche sono i driver dell’ecosistema.
- #26: Cover the business processes.
- #27: Golang -> mysql
- #28: Golang -> mysql
- #29: Golang -> mysql
- #30: Pod che sono effimeri Servizi che sono una astrazione a cui è delegata l’esposizione Un servizio scala insieme e contiene anche piu container I volumi sono ben fatti essendo partiti su gcloud lo storage è una richiesta che deve essere fatta a priori e ci sono diversi driver esempio Load balancer Kubernets ha un cronjob per schedulare operazioni Ecosistema valido
- #31: Pod che sono effimeri Servizi che sono una astrazione a cui è delegata l’esposizione Un servizio scala insieme e contiene anche piu container I volumi sono ben fatti essendo partiti su gcloud lo storage è una richiesta che deve essere fatta a priori e ci sono diversi driver esempio Load balancer Kubernets ha un cronjob per schedulare operazioni Ecosistema valido
- #32: Golang -> mysql
- #33: Golang -> mysql
- #34: Store Stateful container are really important Di giorno container di notte bigdata Monitoring Utile organizzare tramite folders
- #35: Perchè solo quelli opensource? Perche sono i driver dell’ecosistema.
- #36: Golang -> mysql
- #37: La sindrome di muchacha
- #38: La sindrome di muchacha
- #39: Creiamo i nodi Li connettiamo con n
- #40: Creiamo i nodi Li connettiamo con n
- #41: Creiamo i nodi Li connettiamo con n