Building a K8s Operator for a Distributed Database
summary
An Aerospike Kubernetes operator that claims 100% uptime for high‑load databases actually implements live patching, cluster replacement of tens of nodes, and degraded‑node recovery via the Operators API. It encodes rack awareness and XDR topology, uses init‑container flags to avoid accidental data wipes on persistent volumes, requeues reconciliations and requests delayed API responses to allow safe aborts of rolling updates, and is tested against customers running half a trillion objects and petabyte SLAs with sub‑millisecond latency.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:25Kubernetes operators overview
- 3:52Custom resources and controllers
- 6:41Aerospike database
- 8:48Operator design for Aerospike
- 10:14Persistent data handling challenge
- 11:40Rolling update and rollback
- 13:04Scaling challenges and limits
- 14:54Large-scale deployment tips
- 15:35Summary
related talks
21mBuilding a K8s Operator for a Distributed Database
35mOrchestrating Resilient Data: Harnessing the Strength of Kubernetes with Operators
44mDatabases on Kubernetes: Why You Should Care!
41mDatabases on Kubernetes Using a Custom Operator: Day 1, Day 2, and Beyond
34mIntroduction to Kubernetes Operators for Databases
15mHow We Solved the K8s Cluster Communication Challenge
29mFlipkart: Scaling Aerospike at Flipkart with Kubernetes
34mDatabases on Kubernetes