Longhorn: Intro, Deep Dive and Q&A
summary
Longhorn ships a production-ready V2 data engine (SPDK, NVMe-oF) in 1.12.0 released May 2026, claiming multi× higher IOPS and lower latency by assigning dedicated CPU cores and hugepages; rebuild speed and health monitoring improved; DR uses incremental backups to pull a standby DR volume. Planned in 1.12.1 are copy‑on‑write cloning and sharding/erasure parity (k/n) to cut replica storage overhead. Deployments run everywhere Kubernetes runs, including a reported largest cluster of about 2,500 nodes.
AI-written summary · about
chapters (16)
- 0:00Longhorn intro and agenda
- 1:45Longhorn overview and features
- 4:40Longhorn release policy
- 6:04Longhorn architecture
- 8:11Control and data plane
- 10:18V1 vs V2 performance
- 13:30Enterprise features
- 14:55Snapshots and backup
- 18:09Disaster recovery
- 19:13Live upgrade and migration
- 21:31Read-only many volumes
- 22:55Stability and testing
- 23:58Longhorn 1.11 and 1.12 new features
- 26:53V2 performance benchmarks
- 27:14Planned features 1.12.1 and 1.13
- 29:44Community updates
related talks
32mLonghorn: Intro, Deep Dive and Q&A
38mLonghorn: Intro, Deep Dive And Q+A
32mLonghorn: Intro, Deep Dive and Q&A
35mImproving Longhorn Performance With SPDK
47mLonghorn: Intro, Deep Dive and Q+A
48mIntro to Longhorn: Open Source Cloud-Native Distributed Block Storage Built On and For K8s
37mBuild a Kubernetes Based Cloud Native Storage Solution From Scratch
27mThe Storage Architecture of Intel's Data Management Platform (DMP)