Dev Ops Geek Fest: Automating the ForgeRock Platform
Download as PPTX, PDF3 likes5,382 views
This document discusses using DevOps tools and practices with ForgeRock identity platforms. It begins by explaining the needs of different roles that DevOps aims to address. It then covers topics like elastic scaling of ForgeRock, different DevOps tools available, and ForgeRock's role in supporting DevOps. The document demonstrates configuring all ForgeRock components like OpenIDM, OpenAM, OpenDJ and OpenIG using Ansible. It discusses experiences using tools like Docker and the benefits of containers. Finally, it presents a proposal to run OpenAM on Kubernetes and leverage its capabilities for container orchestration.
Ad
More Related Content
What's hot (20)
Viewers also liked (20)
Ad
Similar to Dev Ops Geek Fest: Automating the ForgeRock Platform (20)
Ad
More from ForgeRock (20)
Recently uploaded (20)
Dev Ops Geek Fest: Automating the ForgeRock Platform
- 1. Adventures in DevOps Warren Strange Director, Sales Engineering [email protected]
- 2. DevOps in a nutshell…. source: HTTP://XKCD.COM/974/
- 3. Why DevOps? Copyright © Identity Summit 2015, all rights reserved. • Developer: “I want a development box” • QA tester: “I want to test a complex configuration that mirrors production” • Sys Admin: “I want a reliable, repeatable production configuration” • Potential Customer: “I want a demonstration of how your product works” • ForgeRock University “I want to quickly create lab environments for 30 students”
- 4. Elasticity Copyright © Identity Summit 2015, all rights reserved. • The ForgeRock platform scales extremely well vertically with a small number of nodes • Easy to scale up / down through virtualization, adding more CPU, RAM, etc • OpenAM 13 stateless sessions provide new horizontal scaling options
- 5. Which tool?
- 6. What role can ForgeRock Play? • Make our products more “DevOps” friendly. E.g: – OpenAM 13 REST configuration service – Reduce file system dependencies – Commons project to implement keystore in OpenDJ – More flexible logging options (e.g. syslog) • Longer term: move towards 12factor architecture • What we can’t do is pick a “winner” in the DevOps tools game • Community: How can we facilitate more sharing?
- 7. Enough Talk. Let’s see some DevOps • https://github.com/ForgeRock/frstack • Ansible / Vagrant project to install all of the ForgeRock components – OpenIDM - identity lifecycle management – OpenAM - access management – OpenDJ - directory services – OpenIG - identity gateway – OpenAM Agent - policy enforcement point J
- 8. Demo of frstack ( 5 min)
- 9. Things I learned so far... • Normalizing environments is painful e.g. Apache on CentOS/RHEL is not quite the same as Ubuntu/Debian • More flexible == more brittle e.g. OpenDJ CLI arguments changed slightly from 2.x to 3.x. • Not a lot of sharing right now... – Are DevOps assets too specific to an organization? – Takes too much time to clean up and document DevOps assets?
- 10. Containers gone wild J • Docker = “Micro VMs” – Includes all dependencies – One process per container – Similar to BSD Jails, Solaris Zones • Docker in Production? – Still not for the faint of heart...
- 11. Kubernetes J • Containers alone are not sufficient. They need orchestration, container networking, service lookup, rolling upgrades, placement (affinity / non-affinity) • Created by Google, based on 10+ years of experience running containers at scale • Container agnostic (Docker, Rocket, etc) • Open source project
- 13. Demo of Docker (5 min)
- 14. Docker - What I learned • Great for developers and “throw away” environments • Docker fits best for 12factor, stateless applications • Externalize persistence - it’s a lot of work to “pull apart” applications • Docker “data volumes”: How do you guarantee your container is running on a node that has the data? • Kubernetes data volumes are a higher level abstraction. They are a network resource, not tied to a node implemented using Google Persistent Disk, NFS, iSCSI J
- 15. Questions? J
- 16. DevOps Resources Ansible http://www.ansible.com/resources Jake’s Amazing OpenIDM Vagrant project https://github.com/jakefeasel/openidm- boilerplate/ frstack project https://github.com/ForgeRock/frstack Puppet Module https://github.com/ConductAS/puppet- openam Kubernetes http://kubernetes.io
- 17. Thank You! Warren Strange Director, Sales Engineering [email protected]
- 18. Big Idea: OpenAM on Kubernetes • Strategy – Vanilla OpenAM / Tomcat Docker container, with no “personality” – External OpenDJ config/CTS store – K8 data volume holds ~/openam configuration directory • Keystore, logs, bootstrap, service definitions – Bootstrap script tweaks .openamcfg/ to point to the above k8 volume • Use static DNS names for cluster networking – openam-hosta.localdomain wired for SFO to openam- hostb.localdomain • Use realms, DNS aliases to “personalize” for target environment – realm /acme, dns alias: acme.com J