# Agent Note: Shared persistence write coordinator Status: implemented English | [中文](2026-06-18-shared-persistence-write-coordinator.zh.md) ## Problem The JSONL provider needs correctness-heavy write orchestration around its storage primitives: per-Session state, `session/created` adoption, prefix reads, write-behind control, per-id operation serialization, HMR seeding, and dispose drains. Keeping that lifecycle in the Service Definition prevents an out-of-tree provider from copying it. The removed first-party database provider demonstrated the duplication cost; the [JSONL-only persistence decision](../simplification/2026-08-30-jsonl-only-session-persistence.md) owns its removal. ## Decision `dsh-session-persistence` exports a backend-agnostic `PersistenceCoordinator`. The JSONL provider composes one (`new PersistenceCoordinator(ctx, this)`), implements the small `PersistenceBackend` hook interface, and delegates its stateful public methods (`create`/`append`/`prepare`/`load`/`inspect`/`readFrom`) to it. Backend-owned metadata and revision listing bypass the coordinator. Composition, not inheritance. The coordinator is a concrete class the backend holds, not a base class the backend extends. The risk that a coordinator makes unusual backends fight an inheritance hierarchy is avoided: a backend exposes only the hooks and cannot reach the coordinator's private orchestration state. A third-party backend MAY still implement the abstract service directly without the coordinator, including immutable logical inspection and the default preparation fallback through `load`. The coordinator holds one lifecycle entry for each exact live `Session`: initialization plus a package-private write controller that owns pending events, a fixed batching deadline, the active write, failure retention, and the shared flush barrier. Each `session/event` enters that bounded write path, and `session/flush` bypasses the wait to observe quiescence. The [flush-controller simplification](../simplification/2026-07-23-collapse-persistence-flush-state.md) owns controller consolidation; the [bounded batching decision](2026-08-08-bounded-session-persistence-write-batching.md) owns scheduling cadence. Creation borrows the exact `Session.events` snapshot as its persistence seed. `Session` has already detached, validated, and deeply frozen every event, and the snapshot array remains stable when later appends replace the cached view. The coordinator and its backend hooks only read this typed in-process value, so cloning the complete log again would duplicate the ownership work described by the [agent-scope runtime decision](2026-07-12-agent-scope-runtime-design.md#session-append-materialize-validate-commit-notify). Public persistence `append()` still snapshots caller-owned input at its API boundary. Prepared-session suffixes and events admitted to the write-behind queue retain their existing copies. Those paths establish asynchronous queue ownership one suffix or event at a time and have no measured whole-log clone cost; removing their copies remains a separate ownership audit rather than part of creation-seed borrowing. The coordinator retires a session from `session/disposed`: it waits for the controller's initialization and current flush, serializes a final drain, and removes the controller and owned per-id state only after success. A failure leaves the controller discoverable for backend teardown to retry. Settled per-id chain tails remove themselves only when they are still current, so a completion cannot erase a newer operation for the same id. Backend teardown unregisters write-path listeners, flushes every remaining controller, awaits per-id operations, and then closes the backend. ### The hook interface (`PersistenceBackend`) Five required members plus optional empty-materialization and lifecycle hooks form the only boundary between the coordinator and storage: - `name` — backend label for the dispose-failure `AggregateError`. - `loadStored(id)` — read one stored prefix by id across every storage scope. Preparation, logical load/inspection, physical suffix reads, live adoption, and the create-collision probe share this lookup. The coordinator asserts the returned id and rejects a stored/live cwd mismatch before repair or state publication. - `appendBatch(meta, events, isMaterialized)` — durably append a contiguous batch, lazily materializing the session ATOMICALLY when not yet materialized. Ordinary creation therefore cannot leave an abandoned materialized-but-empty session. - `materializeHeader?(meta)` — explicitly persist a header-only session for `SessionPersistence.ensureMaterialized(session)`. This is reserved for a lifecycle frontend that treats an empty session itself as a resumable durable resource; [standard ACP automation controls](../feature/2026-08-22-standard-acp-automation-controls.md) are the first consumer. Backends that support that lifecycle implement the hook; lazy creation remains the default. - `commitRepair(meta, tornMarker, closers)` — make a crash repair durable: truncate the torn tail (iff `tornMarker !== undefined`) and append `closers`. **NOT required to be atomic** — JSONL legitimately truncates then appends in two fsync'd steps. Used by `prepare`/`load` (truncate + synthetic closers) and live adoption (truncate only, `closers = []`). - `list()` — list all stored metadata. - `close?()` — optional lifecycle teardown for a provider with owned resources; JSONL omits it. The dispose effect awaits it after the quiescence drain so a close failure never masks a drain error. ### The opaque torn marker The single design choice that keeps the seam clean: the crash-repair "where is the torn tail" token is opaque to the coordinator. The coordinator computes the synthetic closers (it owns `interruptedTurnClosers` from `dsh-session`), but it only tests `tornMarker !== undefined` and passes the value straight back to `commitRepair`; it never inspects it. JSONL carries the byte offset to truncate to plus any complete events decoded from an incomplete final frame, while another provider may choose its own marker type. The coordinator therefore knows neither byte lengths nor frame recovery state. ## Testing The shared `runPersistenceContract` proves that JSONL `inspect` balances an interrupted logical view without changing storage or revisions before `prepare` or `load` commits recovery. `runCoordinatorContract` (`tests/coordinator-contract.ts`) covers adoption, HMR, collision, Session and provider disposal drains, and crash-tail repair through an in-memory reference and JSONL. `persistence.spec.ts`, `preparations.spec.ts`, and `write-behind.spec.ts` cover preparation reuse and reservation, bounded prepared-state eviction, fixed-window follow-up batches, live-controller cleanup, same-id chain-tail races, failed-batch retry, and close ordering. JSONL specs retain storage mechanics and the through-coordinator torn-tail case that exercises the opaque-marker branch. ## Alternatives considered - **A base class the backends extend** — rejected for composition: a backend exposes only the hooks, cannot reach the coordinator's private orchestration state, and a third-party backend may still implement the abstract service directly without the coordinator at all. - **A wider hook API** — each candidate hook folds away: there is no scope-specific live lookup because `loadStored` plus the coordinator's cwd check preserves the collision boundary, no storage-locator generic because validated JSONL metadata reproduces its path, no separate `materialize` hook because the first batch must commit atomically with materialization, no separate create-collision probe because it is `loadStored(id) !== undefined`, and no coordinator pass-through for `list()` because listing needs none of the orchestration. ## Consequences The coordinator adds one indirection, an opaque torn marker, detached Session-retirement tasks, and bounded prepared Session state, but centralizes correctness-heavy orchestration for the JSONL provider and future implementations. Session disposal remains an observe-only event, so the Session owner does not await persistence retirement; the coordinator contains failures, preserves pending events in the live controller, and makes provider teardown the quiescence boundary. Its hook surface stays narrow: identity, adoption, collision checks, preparation, and immutable inspection reuse `loadStored`; materialization stays atomic inside `appendBatch`; and listing bypasses the coordinator. Read models use `inspect` rather than `load`, so observing a persisted open turn does not commit interruption closers; the [Session preparation decision](2026-08-05-session-preparation.md) owns reuse, reservation, and publication. A new provider implements storage primitives rather than copy the bounded write lifecycle.