Cooperative Scheduling for Stateful Systems - Michael Youssef & Zhantong Shang, LinkedIn
Description:
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Grab it
Learn how LinkedIn developed and implemented advanced stateful system support for Kubernetes in this technical conference talk. Explore the evolution beyond StatefulSet to handle critical stateful workloads running on bare metal infrastructure across LinkedIn's massive datacenter deployment. Discover the custom LiStatefulSet API implementation, the ApplicationClusterManager's role in enabling customizable safety checks and deployment policies, and the coordination protocol that manages workload lifecycles. Gain insights into how LinkedIn successfully handles planned/unplanned maintenance, hardware swaps, and custom rollout policies while maintaining system availability and data durability across tens of thousands of machines.
Cooperative Scheduling for Stateful Systems in Kubernetes