Building and Deploying apps on Apache Mesos and Marathon

Slides of a my presentation on building and deploying applications on Apache Mesos and Marathon at Hacker's Avenue hosted at Indix.

Transcript

1. Building & Deploying Applications on Apache Mesos & Marathon bit.ly/2gqEIKb
2. dev on ops duty @indix oss contributor (oss.indix.com) aayushKumarJarvis @_iameveryware aayush kumar
3. quick intro to mesos marathon framework a short demo agenda
4. Mesos : A Platform for Fine-Grained Resource Sharing in the Data Center http://mesos.berkeley.edu/mesos_tech_report.pdf Google Borg : https://www.youtube.com/watch?v=0W49z8hVn0k Google Omega : https://www.youtube.com/watch?v=XsXlm4wmB6o origin
5. life without mesos
6. static partitioning
7. static partitioning IS BAD
8. static partitioning does NOT scale (elastically)
9. failures === downtime
10. imagine all this happening at SCALE* * maybe a few hundreds to even thousands of servers
11. mesos : data center’s operating system
12. how it works : master slave setup
13. how it works : framework & resource offers
14. quick facts and checks: • Scalable to 10,000 nodes • High Availability • Native support for launching Docker Containers • Cross platform compatible, runs on Linux, OSX, Windows • Has a Web UI and also API’s for building third party apps
15. marathon Cluster wide init system for services/apps running on the cluster Orchestrator for Apache Mesos
16. constraints Constraints control where apps run to allow optimizing for either fault tolerance (by spreading a task out on multiple nodes) or locality (by running all of an application tasks on the same node). Fields • Hostname - Slaves are matched with their hostnames • Attributes - Slaves are matched through attributes defined as part of config
17. examples : constraints
18. examples : constraints
19. running an app on marathon Example marathon config (JSON) : More example configs: https://mesosphere.github.io/marathon/docs/application-basics.html
20. demo time !
21. advantages • colocate diverse workloads : Same infrastructure can support analytics, stateless microservices, distributed data services and traditional apps to improve utilisation and reduce costs. • provide evergreen extensibility : To run new application and technologies without modifying the cluster manager or any of the existing applications built on top of it. • elastically scale : Scale application and the underlying infrastructure from a handful, to tens, to tens of thousands of nodes.
22. questions ? also a feedback please :)