distributed_index

Overview ¶

Command distributed_index demonstrates the v0.2 distributed write and distributed index-build workflow:

1. A coordinator seeds the dataset schema with a small base write.
2. Two "workers" each write disjoint data as uncommitted fragments with WriteFragments, returning opaque Transactions.
3. A driver commits both transactions in one batch (CommitBuilder. ExecuteBatch merges the Append transactions into a single new version).
4. Each worker builds a per-fragment IVF index segment pinned to shared centroids (CreateIndexUncommitted with WithFragments(...)). The driver merges the segments (MergeIndexSegments) and commits one logical index (CommitIndexSegments).
5. A vector search runs against the committed distributed index.

In production the workers run on separate machines/processes. The Transaction and IndexMetadata objects are shipped to the driver as bytes (see their Bytes() accessors).

Usage: go run ./examples/distributed_index