Milvus Metadata Management (2)
Dec 31, 2019milvus

Summary
This post details the fields within the `Tables` and `TableFiles` metadata tables in Milvus versions 0.5.0 and 0.6.0. It explains the significance of fields like `index_file_size`, `engine_type`, `nlist`, `metric_type`, `state`, `created_on`, `owner_table`, `partition_tag`, and `version`. It also describes fields in `TableFiles` such as `file_type`, `file_size`, and the number of vector rows, and clarifies the relationship between data files and their corresponding index files. The post also highlights the evolution of the data file cleanup mechanism in v0.6.0, moving from a fixed interval hard-delete to a reference counting approach for improved reliability and reduced disk usage.
More from the AI Data Storage Engine trail
- Milvus Snapshots: Point-in-Time Collection Views Without Copying Data
- Announcing Milvus 3.0: Lake-Native Vector Search and a More Powerful Retrieval Engine
- How to Recover Milvus Data After etcd Corruption Without a Backup
- Open Tag: An Open-Source Claude Tag for Claude Code and Codex
- How Force Merge Compaction Nearly Doubled Milvus Search QPS