Rook Deployed Scalable NFS Clusters Exporting CephFS
summary
Rook 1.0 and Ceph Nautilus now dynamically launch highly available, NFS-Ganesha active-active clusters that export CephFS subvolumes; operators create exports with Kubernetes CRDs and Rook launches Ganesha pods that store state in RADOS/etcd-free storage and use libcephfs for I/O. Nautilus adds client reclaim and faster failover so multiple NFS servers can safely share a filesystem, scaling per-export and migrating pod IPs on failover; the implementation is available today. Ceph Nautilus Rook 1.0
AI-written summary · about
chapters (16)
- 0:00Introduction
- 2:23Metadata Servers
- 3:16Origin Story
- 4:37Why Kubernetes
- 8:06Why Ceph
- 10:15NFS Ganesha
- 12:55Scale Out
- 14:33Passive Active
- 15:42NFS Clients
- 17:47Recovering State
- 20:50Coordinating Grace Periods
- 21:49Layering NFS
- 25:49Deployment Management
- 28:43Rook and CephFS
- 30:55Summary
- 32:29Future vision
related talks
37mRook Deployed Scalable NFS Clusters Exporting CephFS
34mRook Deployed Scalable NFS Clusters Exporting CephFS
29mSimplifying Storage in Kubernetes with Rook-Ceph
26mRook: Intro And Deep Dive With Ceph Storage
35mDistributed storage with Rook
27mRook: Intro and Ceph Deep Dive
19mUsing Rook in Production on National Research Platform
31mRook: Intro and Deep Dive With Ceph Storage