Kubernetes on-premises, why you don't want to do it?
summary
A practical guide for running Kubernetes on-premises that walks through making a cluster production-ready by installing MetalLB for load balancing, cert‑manager plus nginx ingress for external traffic, an NFS or local CSI driver for persistent volumes, and Prometheus for observability. Covers common on‑prem problems (DNS resolver misconfigurations, NFS permission issues, CSI driver quirks and immature operators) and recommends planning for slow or failing upgrades via VM snapshots or a standby cluster and tested restore procedures.
AI-written summary · about
chapters (22)
- 0:00Kubernetes on-prem intro
- 1:07Why avoid on-prem Kubernetes
- 2:11Admin tasks for on-prem
- 5:43Choosing network plugin
- 7:33DevOps role in maintenance
- 9:20Load balancer and volumes
- 10:49Custom cluster needs
- 13:18Production cluster features
- 14:01Core service components
- 15:49On-prem deployment problems
- 18:27MetalLB and Ingress setup
- 20:48Volume driver and cert manager
- 22:19Default volume class issues
- 24:49Debugging NFS permissions
- 26:14Self-solving on-prem issues
- 27:41Beta components risk
- 29:10Basic on-prem deployment summary
- 29:56Insurance industry on-prem
- 31:19Cluster maintenance challenges
- 32:45Cluster update problems
- 34:33On-prem backup and restore
- 35:37Q&A
related talks
42mKubernetes on-premises
02h 42mSecuring Your On-Premise Kubernetes Cluster from Scratch
36mCan on-prem be also a cloud Networking made easy
19mBuilding an On-Premise Kubernetes...
32mHow to make kubernetes rhyme with prod readiness
40mThe Good, the Bad, and the Ugly of Managing Production Systems with Kubernetes
49mThe state of on-prem Kubernetes
29mHow to make Kubernetes rhyme with prod readiness