Taking Your Database Beyond the Border of a Single Kubernetes Cluster
Talkat KubeCon + CloudNativeCon North America 2021 · on Thursday, 2021-10-14· at 14:30· Petree Hall D + Online·
summary
Runs Apache Cassandra as a six‑node cluster (2 nodes per zone across 3 zones) and expands to Cassandra across two regions by creating a second Kubernetes cluster per region to double capacity, add geo locality and HA. Key pitfalls are overlapping pod CIDRs and DNS: plan non‑overlapping CIDR ranges and choose between VPC routing, VPN/BGP, or overlay networks. Production options include Istio and psyllium for pod‑IP routing, node‑local DNS or managed Cloud DNS, and an operator that reconciles resources across clusters (already in use).
AI-written summary · about
chapters (17)
- 0:00Intro
- 0:26What are we trying to do?
- 1:19Seems simple...
- 1:35Right?
- 2:56Networking
- 4:54Address Collisions
- 7:06Flatten It Out
- 7:55It depends...
- 9:40Overlay Networks
- 10:23Service Discovery
- 11:25DNS Stubs
- 12:59Managed DNS
- 14:42Strong Alternatives
- 17:25Added Benefits
- 20:54Deployment
- 23:41So what you're saying is...
- 25:13References
related talks
29mPre-cloud Cassandra to Multi-Cloud Kubernetes
42mDistributed Data Stores on Kubernetes
35mKubernetes Persistent Data Challenges – AZ, Region and Multi-Clou...
33mWhat We’ve Learned Building a Multi-Region DBaaS on Kubernetes
45mMulti-Kubernetes, Multi-Region — Help!
34mBringing Apache Cassandra Closer to Kubernetes
32mRunning Databases in Istio's Service Mesh
31mAn Adventure in Multi Cloud with Kubernetes and Distributed Data