2019-11-17, 17:00–17:30, The Theater
Disaster management is a mission-critical function that most startups today don’t plan for. Recent disruption in Cloud providers demonstrates that no infrastructure can avoid the inevitable downtimes caused by the catastrophe. All infrastructures should have fail-safe measures to ensure “nothing” is lost in the event of service disruption. This proposal demonstrates a simple system for ensuring entire Kubernetes clusters can be replicated on-demand. The use-cases are many and obvious including data security, cost optimization and reducing the downtime of workloads.
In this talk, Ajesh and Vishnu will present a methodology for using OpenEBS custom controllers to replicate/migrate Kubernetes cluster into another Kubernetes cluster. They will demonstrate data migration on the fly and discuss both use cases and hurdles faced.
This talk will benefit DevOps attendees who support stateful applications and manage the configuration of Kubernetes clusters. Using this approach users can easily replicate/migrate stateful applications from one Kubernetes cluster to another.
The cluster migration pattern is useful in situations like canary deployment, blue-green deployment, cluster upgrade, disaster recovery.
This talk will be beneficial SREs and DevOps users who are concerned about data privacy and do not trust third-party vendors. This approach also reduces time and cost overhead.