Effective migration from Newton to Queens with minimal downtime
summary
A 6‑month migration moves ~2,000 VMs and 500 TB from a vendor OpenStack to an internal Queens Ceph-based cloud using a custom Python toolset that orchestrates OpenStack APIs, RBD mirroring, a port‑sync daemon, and a Scalr REST proxy. Root disks stream via a rescue‑image agent with per‑64KB checksums for minimal writes and simple rollback; process scales toward 100 VMs/day, achieves RBD mirroring at ~2 TB/hour and saves ~200 TB of Ceph storage by avoiding full image copies.
AI-written summary · about
chapters (12)
- 0:00Intro
- 0:26The migration scenario
- 1:22Business requirements
- 6:43Why to reinvent the wheel?
- 8:14Toolset overview
- 10:25VM cloning steps
- 12:06Root disk migration workflow
- 13:57Ceph REST proxy
- 15:40Port sync daemon
- 16:37Scalr migration tool and REST proxy
- 18:22Use of the toolset after migration
- 20:08Lessons learned
related talks
38mMigrating workloads from VMware to OpenStack
35mMigrating a live site across the country without downtime
40mCoriolis How to easily migrate your IT workloads to Charmed OpenStack
33mEscaping the Vendor Trap: A Journey for Migrating Legacy Infrastructure To Op...
45mCloud Migration
36mOpenPOWER and GPU Porting Platform for FOSS Using OpenStack at OSUOSL
43mSupporting NASA Science with OpenStack
29mJourney through the center of migrations