What Playing Minecraft Taught Me About DevOps
summary
Describes converting huge partitioned heap tables to clustered column store indexes without the downtime SSDT’s rebuild-copy-drop would cause, using stored procedures, SQL Agent jobs, a logging table and per-partition high-watermarks, tested with DBFit and full pre-prod restores, moving data partition-by-partition over four months, monitoring CPU and log growth, then swapping and renaming tables — result: massive compression and a controlled rollout that avoids the “copy everything at once” failure mode.
AI-written summary · about
chapters (8)
- 0:00Introduction and setup
- 1:25Minecraft and DevOps basics
- 2:51Surviving the first night
- 5:59Real-world DevOps problem
- 9:48SSDT and creepers analogy
- 11:14Controlled data migration
- 13:59Deployment and testing
- 18:31Lessons learned and conclusion
related talks
23mShift-Left: A DevOps and Database Story
44mBreaking a Monolithic Devops
23mShift-Left: A DevOps and Database Story
44mChaos Engineering for SQL Server
48mDevOps Discovery: Navigating the Database Landscape
01h 00mDevOps for Your Relational Database Foundations
30mA Devops Approach to Data: A Tale Through the Trenches in Search of a Cloud Data Revolution
23mShift-Left: A DevOps and Database Story
