docs(optimize): document the pre-existing-drift reconcile

- maintenance.md: add the reconcile bullet (metadata-only manifest catch-up for
  a table with HEAD>manifest and an empty plan), and correct the 'requires a
  recovered graph' note — that guard is what makes the reconcile safe (no
  sidecar-covered drift in flight), not a claim that no drift exists.
- AGENTS.md: restore the reconcile mention in the Compaction capability row.
- testing.md: the maintenance.rs row lists all three optimize tests (publishes /
  reconciles / defers); fix the stale failpoints ensure_indices test name
  (recovery_rolls_forward_ensure_indices_on_feature_branch).

AGENTS.md

@@ -237,7 +237,7 @@ omnigraph policy explain --actor act-alice --action change --branch main
 | Per-dataset versioning + time travel | ✅ | `snapshot_at_version`, `entity_at`, snapshot-pinned reads across many tables |
 | Per-dataset branches | ✅ | **Graph-level** branches (atomic across all sub-tables), lazy fork, system branch filtering |
 | Atomic single-dataset commits | ✅ | **Multi-table publish via three layers**, NOT a single Lance primitive: (1) per-table Lance `commit_staged` for the data write, (2) `__manifest` row-level CAS via `ManifestBatchPublisher` for cross-table ordering, (3) the open-time recovery sweep for the residual gap between (1) and (2). All three layers ship; the five migrated writers (`MutationStaging::finalize`, `schema_apply`, `branch_merge`, `ensure_indices`, `optimize_all_tables`) write a `__recovery/{ulid}.json` sidecar before Phase B and delete it after Phase C. The next `Omnigraph::open` (gated on `OpenMode::ReadWrite`) runs the sweep in `db/manifest/recovery.rs`: classify, decide all-or-nothing per sidecar, roll forward via single `ManifestBatchPublisher::publish` or roll back via `Dataset::restore` followed by a manifest publish of the restored version (so both directions converge to `manifest == HEAD` — no residual drift), and record an audit row in `_graph_commit_recoveries.lance` (queryable via `omnigraph commit list --filter actor=omnigraph:recovery`). Continuous in-process recovery (no restart needed between Phase B failure and recovery) is the goal of a future background reconciler. Engine writes route through a sealed `TableStorage` trait exposing `stage_*` + `commit_staged` as the canonical staged-write surface; documented inline-commit residuals (`delete_where`, `create_vector_index`, plus legacy `append_batch` / `merge_insert_batches` / `overwrite_batch` / `create_*_index`) remain on the trait until upstream Lance ships a public two-phase API ([#6658](https://github.com/lance-format/lance/issues/6658), [#6666](https://github.com/lance-format/lance/issues/6666)) and the migration of every call site completes. |
-| Compaction (`compact_files`) | ✅ | `omnigraph optimize` orchestrates over all node/edge tables, bounded concurrency; **publishes each compacted table's new version to `__manifest`** (so the manifest tracks the Lance HEAD — required for reads to observe compaction and for schema apply / strict writes to pass their HEAD-vs-manifest precondition), under the per-`(table, main)` write queue with `SidecarKind::Optimize` recovery coverage; **refuses on an unrecovered graph** (errors if a `__recovery` sidecar is pending — recovery may roll back a partial write, so optimize requires `manifest == HEAD` going in); **skips blob-bearing tables** (reported via `TableOptimizeStats.skipped`, not silent), gated on `LANCE_SUPPORTS_BLOB_COMPACTION` until the upstream blob-v2 compaction-decode bug is fixed (see [docs/dev/invariants.md](docs/dev/invariants.md) Known Gaps) |
+| Compaction (`compact_files`) | ✅ | `omnigraph optimize` orchestrates over all node/edge tables, bounded concurrency; **publishes each compacted table's new version to `__manifest`** (so the manifest tracks the Lance HEAD — required for reads to observe compaction and for schema apply / strict writes to pass their HEAD-vs-manifest precondition), under the per-`(table, main)` write queue with `SidecarKind::Optimize` recovery coverage; **reconciles pre-existing uncovered drift** (a table with `HEAD > manifest` and an empty compaction plan — left by a pre-fix optimize or external raw `compact_files` — gets a metadata-only `manifest = HEAD` catch-up: the producer-side heal for legacy/external drift, the strict write path otherwise 409s on it); **refuses on an unrecovered graph** (errors if a `__recovery` sidecar is pending — recovery may roll back a partial write, so optimize requires no in-flight sidecar going in, which is also what makes the reconcile safe: any `HEAD > manifest` it then sees is content-preserving, never a partial write); **skips blob-bearing tables** (reported via `TableOptimizeStats.skipped`, not silent), gated on `LANCE_SUPPORTS_BLOB_COMPACTION` until the upstream blob-v2 compaction-decode bug is fixed (see [docs/dev/invariants.md](docs/dev/invariants.md) Known Gaps) |
 | Cleanup (`cleanup_old_versions`) | ✅ | `omnigraph cleanup` with `--keep` / `--older-than` policy |
 | BTREE / inverted (FTS) / vector indexes | ✅ | `ensure_indices` builds them on every relevant column; idempotent; lazy across branches |
 | `merge_insert` upsert | ✅ | `LoadMode::Merge`, mutation `update`/`insert`/`delete` lowering |

docs/dev/testing.md

@@ -34,8 +34,8 @@ The engine's `tests/` is the principal coverage surface; most graph-shaped behav
 | `s3_storage.rs` | S3-backed graph (skipped unless `OMNIGRAPH_S3_TEST_BUCKET` is set) |
 | `lance_version_columns.rs` | Per-row `_row_last_updated_at_version` behavior |
 | `validators.rs` | Schema constraint enforcement (enum, range, unique, cardinality) across JSONL, insert, update paths |
-| `maintenance.rs` | `optimize` (compaction) + `cleanup` (version GC): empty/idempotent/no-op edges, policy validation, head preservation; `optimize` publishes the compacted version so the manifest tracks the Lance HEAD and a subsequent schema apply succeeds (`optimize_publishes_compaction_to_manifest_so_schema_apply_succeeds`), and reconciles a pre-existing manifest-behind-HEAD drift forged via raw Lance compaction (`optimize_reconciles_preexisting_manifest_head_drift`) |
-| `failpoints.rs` | Failure-injection coverage (gated on `failpoints` feature). Includes the five per-writer Phase B → recovery integration tests (`recovery_rolls_forward_after_finalize_publisher_failure`, `schema_apply_phase_b_failure_recovered_on_next_open`, `branch_merge_phase_b_failure_recovered_on_next_open`, `ensure_indices_phase_b_failure_recovered_on_next_open`, `optimize_phase_b_failure_recovered_on_next_open`). |
+| `maintenance.rs` | `optimize` (compaction) + `cleanup` (version GC): empty/idempotent/no-op edges, policy validation, head preservation; `optimize` publishes the compacted version so the manifest tracks the Lance HEAD and a subsequent schema apply succeeds (`optimize_publishes_compaction_to_manifest_so_schema_apply_succeeds`), reconciles a pre-existing manifest-behind-HEAD drift forged via raw Lance compaction (`optimize_reconciles_preexisting_manifest_head_drift`), and refuses while an unresolved recovery sidecar is pending (`optimize_defers_when_recovery_sidecar_is_pending`) |
+| `failpoints.rs` | Failure-injection coverage (gated on `failpoints` feature). Includes the five per-writer Phase B → recovery integration tests (`recovery_rolls_forward_after_finalize_publisher_failure`, `schema_apply_phase_b_failure_recovered_on_next_open`, `branch_merge_phase_b_failure_recovered_on_next_open`, `recovery_rolls_forward_ensure_indices_on_feature_branch`, `optimize_phase_b_failure_recovered_on_next_open`). |
 | `recovery.rs` | Open-time recovery sweep — sidecar I/O, classifier dispatch (NoMovement / RolledPastExpected / UnexpectedAtP1 / UnexpectedMultistep / InvariantViolation), all-or-nothing decision, roll-forward via `ManifestBatchPublisher::publish`, roll-back via `Dataset::restore`, audit row in `_graph_commit_recoveries.lance`, `OpenMode::ReadOnly` skip path |
 | `composite_flow.rs` | Compositional/narrative end-to-end stories — multi-step flows that compose mechanics covered by other test files. Catches integration regressions where individual operations all pass their unit tests but their composition breaks (sequential merges, post-merge main writes, time-travel through merge DAG, reopen consistency over multi-merge histories, post-optimize and post-cleanup strict writes). |
 

docs/user/maintenance.md

@@ -4,10 +4,11 @@
 
 ## `optimize_all_tables(db)` — non-destructive
 
-- Lance `compact_files()` on every node + edge table on `main`, then **publishes the compacted version to the `__manifest`** so the manifest's `table_version` tracks the compacted Lance HEAD. Reads pin the manifest version, so without this publish compaction would be invisible to readers *and* would break the HEAD-vs-manifest precondition of the next schema apply / strict update/delete ("stale view … refresh and retry"). The publish advances the graph version (a system-attributed commit) only for tables that actually compacted.
+- Lance `compact_files()` on every node + edge table on `main`, then **publishes the compacted version to the `__manifest`** so the manifest's `table_version` tracks the compacted Lance HEAD. Reads pin the manifest version, so without this publish compaction would be invisible to readers *and* would break the HEAD-vs-manifest precondition of the next schema apply / strict update/delete ("stale view … refresh and retry"). The publish advances the graph version (a system-attributed commit) for tables that actually compacted.
+- **Reconciles pre-existing benign drift.** A table whose Lance HEAD already sits ahead of its manifest pin with nothing left to compact — left by a pre-fix `optimize` that never published, or by an external raw `compact_files` — is caught up by publishing `manifest = HEAD` (a metadata-only commit; `committed: true` with `fragments_added == 0`, no fragment changes). This is the producer-side heal for *uncovered* drift, and it is safe because the "requires a recovered graph" guard below ensures no recovery sidecar is in flight, so any `HEAD > manifest` it sees is necessarily content-preserving — never a partial write. Strict writes / schema apply on such a table 409 with "stale view" until an `optimize` (or any insert/merge, which heals it as a side effect) reconciles it.
 - Rewrites small fragments into fewer large ones; old fragments remain reachable via older manifests until `cleanup` runs.
 - Each table's compact→publish runs under its per-`(table, main)` write queue (serializing with concurrent mutations — compaction is a Lance `Rewrite` op that retryable-conflicts with a concurrent merge/update/delete on overlapping fragments). The Lance-HEAD-before-manifest-publish gap is covered by a `SidecarKind::Optimize` recovery sidecar (loose-match): a crash in that window rolls the compacted version forward on the next `Omnigraph::open` (compaction is content-preserving, so roll-forward is always safe).
-- **Requires a recovered graph.** `optimize` refuses (errors) when an unresolved recovery sidecar is present under `__recovery` — operating on an unrecovered graph could publish a partial write the open-time recovery sweep would roll back. Reopen the graph to run the recovery sweep, then re-run `optimize`. (Recovery roll-back now publishes its restored version, so a recovered graph always satisfies `manifest == Lance HEAD` going in; there is no leftover drift for `optimize` to interpret.)
+- **Requires a recovered graph.** `optimize` refuses (errors) when an unresolved recovery sidecar is present under `__recovery` — operating on an unrecovered graph could publish a partial write the open-time recovery sweep would roll back. Reopen the graph to run the recovery sweep, then re-run `optimize`. (This is exactly what makes the reconcile above safe: a recovered graph has no *sidecar-covered* drift going in, so any `HEAD > manifest` `optimize` then sees is uncovered, content-preserving drift — never a partial write.)
 - Bounded by `OMNIGRAPH_MAINTENANCE_CONCURRENCY` (default 8).
 - Returns `[TableOptimizeStats { table_key, fragments_removed, fragments_added, committed, skipped }]`.
 - **Blob tables are skipped.** A table that declares any `Blob` property is not compacted: it is reported with `skipped: Some(BlobColumnsUnsupportedByLance)` (and logged via `tracing::warn`) instead of compacted, and the rest of the sweep proceeds normally. The current Lance `compact_files` mis-decodes blob-v2 columns under its forced `BlobHandling::AllBinary` read; **reads and writes are unaffected** — only compaction is. This is gated by `LANCE_SUPPORTS_BLOB_COMPACTION` (`db/omnigraph/optimize.rs`) and removed when the upstream Lance fix lands (see [docs/dev/lance.md](../dev/lance.md)). Consequence: fragment count and deleted-row space on blob tables are not reclaimed until then; query results are never affected.