Tales from the Azure Migration Trenches
summary
Claims a live on‑prem to Azure migration with zero data loss and minimal downtime and shows how it actually happens: uses SQL Data Sync (hub/member plus a local sync agent) for online DB sync, robocopy for files, Azure App Services + Azure SQL, GitHub + Azure DevOps pipelines, Terraform for infra, and a feature flag shim to swap SignalR backplanes. Initial bulk copy of ~200 GB takes ~10 hours, file initial load 3–4 days, and the sync tooling has hard limits like 500 tables per sync group and 24 MB row size. The end state is bidirectional sync, DNS flip to Cloudflare, and a successful roll‑out.
AI-written summary · about
chapters (8)
- 0:00Azure migration challenges
- 3:12Discovery assessment phase
- 15:16Cloud migration strategy
- 21:54Azure SQL migration tools
- 36:38SQL Data Sync setup
- 48:30Migrating code and pipelines
- 56:28Networking and SignalR issues
- 1:00:42Lessons learned and rollout
related talks
26mTransition to the Cloud Return
56mLittle Fluffy Clouds
01h 00mLet's Get Scary and Continuously Deploy Database Changes!
38mSync Work Items Between Multiple Team Projects on Azure DevOps: the Good, the Bad, the Ugly
59mBuilding Resilient Applications In Microsoft Azure
50mBringing DevOps to the Database
25mMass migrations to the cloud
28mCitus Cluster: our zero-downtime migration story