Not everyone can take advantage of Managed Platforms like MongoDB Atlas. For those that can’t, there are alternatives. We’ll demonstrate how we are using the Enterprise Kubernetes Operator to create containers running MongoDB in Red Hat’s OpenShift Platform on Azure at scale.
2. MongoDB Operator and OpenShift
Introduction
What is OpenShift
The Problem with Self Service
Databases in a Non-Persistent World
MongoDB Operator to the Rescue
Demo
Questions
3. A Little About Mike
I am a Database Engineer
I was a Software Engineer
I was a Helpdesk Support
I am not a Systems Engineer
I am not a Storage Engineer
I am not a Linux Admin
I am not a Kubernetes Admin
@mikegray831
mongomikeblog.wordpress.com
4. A Little About Mark
I am not a DBA
I am not a Developer
I was a Windows Admin
I am a Systems Engineer
I am a Storage Engineer
I am a Linux Admin
I am a Kubernetes Admin
@markdeneve
www.xphyr.net
5. Paychex
Paychex is a leading provider of integrated human capital management
solutions for payroll, HR, retirement, and insurance services. Backed by
46 years of industry expertise, Paychex serves approximately 605,000
payroll clients as of May 31, 2017, across more than 100 locations and
pays one out of every 12 American private sector employees.
6. #PaychexIsIT
• Paychex IT is:
• Linux
• Windows
• Docker
• Kubernetes
• OpenShift
• Azure
• MongoDB
• Kafka
• RabbitMQ
• Ansible
• Cassandra
• Redis
• Puppet
• OpenStack
• MySQL
• Microsoft SQL
Server
• Oracle
9. Kubernetes
• Container Management and
Orchestration
• Open Source Project
• 52k+ Github Stars
• 18k+ Github Forks
• Started by Google in 2014
• Backed by Multiple Vendors
10. OpenShift - Enterprise Kubernetes
• Developer focused user experience
• More Secure by design
• Polished End User Experience
Two Versions:
• OpenShift – Backed by Red Hat Support
• OKD – Upstream OpenSource Distribution
11. OpenShift @Paychex
• In Production since 2015
• 20 Clusters accross two regions
• 7 in-place upgrades
• OpenShift hosted applications are
responsible for moving well over $500
billion per year at Paychex
14. Developers are Clever!
• Shadow IT
• Tight Timelines
• Need to innovate Now
• “Just using it for testing”
• Always becomes PROD
Attribution: wocintechchat.com
19. Shadow IT in Kubernetes is not Ideal
The GOOD
• Self Service
• Quick to Spin Up
• Quick to scale
• Works for simple use cases
The BAD
• No easy way to Get
DBAs involved
• Black Box is NOT always a
good thing
• No control over data and where
it is going
• Database Sprawl
21. Persistent Storage in Kubernetes
• Persistent Storage in Kubernetes is
Hard
• Complexity introduced when
a container can run anywhere
• Multiple Vendors, Multiple Solutions
• No good ways to back up data
• Lack of Enterprise Solutions
• Long Term Archival / Retrieval is a
challenge
23. Kubernetes Operators
• A method of packaging, deploying and
managing a Kubernetes application.
• A Kubernetes application is
• deployed on Kubernetes
• managed using the Kubernetes native tooling
• Think of Operators as the runtime that
manages this type of application on
Kubernetes.
25. MongoDB Operator + MongoDB Ops Manager
• Features added by Mongo
• Deploy containers with MongoDB installed in
standard patterns
• Standalone
• Replica Set
• Sharded Cluster
• All the best of MongoDB Enterprise / Ops
Manager
• Security
• Backups/Restores
• Easy Upgrades
26. MongoDB Ops Manager / Cloud Manager
• Ops Manager/ Cloud Manager
• MongoDB Provided Monitoring /
Backup / Automation Solution
• Accepts API calls from Operator
to build MongoDB
27. Self Service, with Guide Rails
• How does MongoDB Operator Help guide
Devs?
• All the benefits of containerization, plus
• Security
• Standardization
• Same Deployment tools that are used in
production
• Easy remote management through Cloud
Manager
• "Push Button"
• Easy for Database Engineers / DevOps / SRE
teams to manage using current process
30. Demo Tasks?
• Create Database in OpenShift – prerecorded
• On the fly upgrade - live (but have recording)
• Scale Mongo Cluster up - live (but have recording)
32. MongoDB Operator @ Paychex
• Sandbox Environment First
• Used in Non-Prod/Development Envs
• Automate data restores for test
• Prod will remain Separate Infrastructure
33. Upcoming Features of the MongoDB Operator
• Check out the 3 PM session in Nassau presented by Jason Mimick and
Andrey Belik for the full Roadmap!
34. Summary
• Save Devs from Themselves!
• Simplify infrastructure
• Keep ops teams involved (they exist for a reason!)