How One Iceberg Optimization Forced Me to Rebuild My Stock Analysis Database
Same Job, Two Engines: InnoDB and Iceberg Internals
After five years running analytics on InnoDB, the author switched to an Iceberg-on-MinIO lakehouse for one key optimization: columnar reads. The post explains InnoDB's three guarantees and why they hurt write-heavy ingestion, then details Iceberg's three-layer query planning—data files, manifests, and manifest lists—and how Polaris and MinIO fit in. A practical deep dive into OLTP vs. OLAP internals.
After all the technical jargon, the English translation is the MVCC allows inserts and updates to happen without blocking reads.