OpenStack with Puppet and MongoDB Ops Manager Automates Database Provisioning Across Multiple Data Centers
Cisco uses MongoDB in multiple projects across the company. At MongoDB World this year, two of its software developers presented how they had built a fully automated Continuous Integration / Continuous Delivery (CI/CD) pipeline to enable one of their largest customers to rapidly push new features and upgrades into production.
Cisco develops software used by some of the largest Pay-TV companies on the planet. MongoDB is used as the database in the Cisco solution to power critical services such as the program guide, subscriber information, entitlements, registered devices, and more. One of Cisco’s largest customers wanted to orchestrate the creation of sharded MongoDB clusters across multiple data centers, on top of its OpenStack private cloud.
MongoDB Ops Manager allows administrators to automate the deployment of MongoDB. Before that can happen, the administrator must first provision the physical hardware and Virtual Machines (VMs), install agents, create cluster users, and mount storage volumes. In Cisco’s session, the software developers demonstrated how they were able to combine all of these steps into a single process, using:
- OpenStack Heat for orchestration of VM and networking resources
- Puppet for configuration scripts and installation of Ops Manager agents
- Ops Manager and its programmatic API for provisioning MongoDB across multiple regions to enable active/active clustering
The presentation walks through each step in the pipeline, with the relevant code samples.
Download the MongoDB Operational Best Practices guide to learn more on making your operations teams more productive.