What You Need to Know Before Using Local Persistent Volumes
Talkat KubeCon + CloudNativeCon North America 2021 · on Wednesday, 2021-10-13· at 15:25· Concourse Hall 152 + Online·
summary
Local PersistentVolumes require explicit provisioning and operational work: choose manual, static using the local-pv static provisioner daemonset that creates PVs per disk, or dynamic via zfs-csi and topolvm which create exact-size volumes on demand. Recovery from an unrecoverable node needs the sequence delete PVC then pod (or automation) because PVs are node-bound; naive drains and cloud upgrades (GKE PDB timeout ~60 minutes) risk double migrations or data loss for 5–10 TB pods. Mitigations: automation, priority classes, taints, and Kubernetes scheduler/storage fixes in 1.21+.
AI-written summary · about
chapters (9)
- 0:00Intro
- 1:09What are persistent volumes
- 6:57Local vs Network attached volumes
- 10:23How to provision local volumes
- 16:10How to deal with operations
- 25:25Problems with kubernetes
- 27:43Problems with stateful sets
- 30:33Concurrent pod scheduling
- 32:56Storage capacity
related talks
28mKubernetes Storage Evolution: Enabling High Performance Distributed Datastores
36mPersistent Storage with Kubernetes in Production - Which Solution and Why?
32mYour Pods can be stateless, your VMs cannot! How to manage state in k8s
28mPersistent Storage with Kubernetes in Production - Which Solution and Why?
50mRunning Databases Efficiently at Scale - Case Study of Flipka...
29mStateful Workloads in Kubernetes: A Deep Dive
10m⚡ Lightning Talk: Navigating the Complexities of Stateful Workloads in Kubernetes
41mUsing External Persistent Volumes to Reduce Recovery Times and Achieve High Availability on DC/OS