@serfab/cadre-core 1.14.0 → 1.16.0

package/src/strand-instance-manager.ts

@@ -1,1883 +1,2026 @@
-import debug from 'debug';
-import type { PrivateKey } from '@libp2p/interface';
-import { createLibp2pNode, type IRawStorage } from '@optimystic/db-p2p';
-import { wrapStorageWithCache, disposeStorageCache } from '@serfab/quereus-plugin-sereus';
-import { StrandDatabase } from './strand-database.js';
-import { PeerJoinBackfill, type PeerJoinBackfillConfig } from './peer-join-backfill.js';
-import { strandNetworkStatePersistence, type StrandNetworkStateStore } from './strand-network-state.js';
-import {
-  StrandRevocationEnforcer,
-  createRevocationConnectionGater,
-  readStrandRevocationRows,
-  type StrandRevocationEnforcementConfig
-} from './strand-revocation-enforcer.js';
-import {
-  StrandMembershipReconciler,
-  type PendingMembershipInviteSource,
-  type StrandMembershipReconciliationConfig
-} from './strand-membership-reconciler.js';
-import { removeMemberPeer } from './strand-membership-writer.js';
-import { strandMemberKeyPair } from './strand-member-key.js';
-import { assertSchemaSignature } from './schema-verification.js';
-import { assertStrandScopeKey } from './storage-scope.js';
-import {
-  StrandFirstSyncGate,
-  StrandAwaitingFirstSyncError,
-  strandFirstSyncComplete,
-  DEFAULT_STRAND_FIRST_SYNC_TIMEOUT_MS,
-  type StrandFirstSyncConfig
-} from './strand-first-sync-gate.js';
-import type {
-  StrandInstance,
-  StrandRow,
-  StorageConfig,
-  NetworkConfig,
-  LatencyHint,
-  NodeProfile,
-  SAppConfig,
-  SAppInfo,
-  RawStorageProvider,
-  Libp2pNodeWithRepo
-} from './types.js';
-import { DEFAULT_CONNECTION_MONITOR, resolveStrandClusterSize, strandClusterPolicy } from './types.js';
-import { strandNodeAddrs } from './strand-network-config.js';
-import { resolveRelayServer } from './relay-server.js';
-import { strandCohortTopicOption, type StrandReactivityConfig } from './strand-reactivity.js';
-import { superviseRelayReservation, type RelayReservationSupervisor } from './relay-reservation.js';
-import { isCohortUnreachableRead } from './control-read-retry.js';
-import { declaredCohortReadDeadlineMs, optimysticDialLimits, peerJoinPushBudget, relayReservationBudgetMs, resolveLinkRoundTripMs } from './link-budget.js';
-
-const log = debug('sereus:cadre:strand-manager');
-
-/**
- * How often a joining launch that could not build its database, because no seed peer had
- * answered, rebuilds its runtime (`StrandInstanceManager.deferUnreachableJoin`). Each attempt
- * starts a libp2p node and dials the seed, so it is spaced well above the first-sync gate's
- * probe cadence.
- */
-export const UNREACHABLE_JOIN_RETRY_MS = 5_000;
-const timing = debug('sereus:cadre:timing');
-
-/** One caller blocked in {@link StrandInstanceManager.whenWritable}. */
-interface WritableWaiter {
-  resolve: () => void;
-  reject: (error: Error) => void;
-}
-
-/**
- * Whether a tracked instance is launched but waiting for its first sync: the runtime is
- * up (`libp2pNode` set) while the database is still withheld (`database` unset). That
- * is the first-sync gate's state regardless of what `status` says — the hibernation
- * manager's idle timer can flip a long-gated joiner to `'idle'` without changing it.
- *
- * NOTE: `status` is therefore not the source of truth for "gated": a joiner that reaches
- * nobody for `idleTimeout` (5 min at the interactive hint) reads `'idle'`, then
- * `'hibernating'`, while still holding no database. Harmless today — every decision the
- * runtime makes goes through this predicate, {@link StrandInstanceManager.publishDatabase}
- * announces on the idle-flipped instance exactly as on a `'syncing'` one, and an app that
- * shows "waiting for the other member" on `'syncing'` merely sees "idle" instead after
- * five minutes. If an app ever needs the gated LABEL to survive the idle flip, make the
- * hibernation manager's idle transition preserve `'syncing'` rather than adding a flag.
- */
-export function isAwaitingFirstSync(instance: StrandInstance): boolean {
-  return instance.libp2pNode !== undefined && instance.database === undefined;
-}
-
-/**
- * The status a LIVE instance (runtime rebuilt or woken) should report: `'active'` once its
- * database is published, `'syncing'` while the first-sync gate still withholds it. Every
- * site that would otherwise write `'active'` after a build or a wake goes through this,
- * so a gated joiner never reads as writable.
- */
-export function liveStrandStatus(instance: StrandInstance): 'active' | 'syncing' {
-  return instance.database ? 'active' : 'syncing';
-}
-
-/**
- * Configuration for starting a strand instance
- */
-export interface StartStrandConfig {
-  strandRow: StrandRow;
-  /**
-   * sApp configuration provided by the hosting application. Absent for a **storage
-   * replica** — a strand this node stores and serves without the app installed
-   * (`CadreNodeConfig.hostUnclaimedStrands`): no schema signature check, no `App`
-   * schema, no `sAppInfo` on the instance, and the default latency hint. A replica is
-   * always a joiner; `StrandDatabase` refuses to found one. An app that later claims the
-   * strand upgrades the running replica in place ({@link StrandInstanceManager.attachSApp}).
-   */
-  sAppConfig?: SAppConfig;
-  storage?: StorageConfig;
-  network?: NetworkConfig;
-  profile: NodeProfile;
-  defaultLatencyHint: LatencyHint;
-  privateKey?: PrivateKey;
-  /** Cohort-derived discovery seed (multiaddr strings). Defaults to [] when omitted. */
-  bootstrapNodes?: string[];
-  /**
-   * Require a valid author signature on the sApp schema before bring-up.
-   * Defaults to true (fail closed) when omitted; set false only for dev/test
-   * with unsigned demo schemas. Mirrors {@link CadreNodeConfig.requireSignedSchemas}.
-   */
-  requireSignedSchemas?: boolean;
-  /**
-   * Whether this node founds the strand (vs. joins it). Forwarded to the
-   * StrandDatabase so the founder bootstrap (Header, founding Member/Manager)
-   * runs once at bring-up. Joiners leave this unset and write nothing. Callers
-   * resolve it BEFORE calling `startStrand` (`CadreNode.launchStrand` derives it
-   * from the row's `FounderOwnerKey` when no explicit flag is given — see
-   * {@link StrandConfig.founder}); a founder request that arrives while the
-   * instance is already tracked goes through {@link StrandInstanceManager.foundExistingStrand},
-   * never through a repeat `startStrand`.
-   */
-  founder?: boolean;
-  /**
-   * THIS party's own strand membership private key (base64 protobuf) for a closed
-   * strand — the identity the founder bootstrap derives `Member.Key` /
-   * `Manager.MemberKey` from. Resolved by the caller (`CadreNode.launchStrand`: an
-   * explicit attach-time key, else the control-layer `StrandPartyKey` row, healed by
-   * mint-on-launch for the founding machine) and retained with the launch config, so a
-   * hibernation wake rebuilds under the same identity. Deliberately NOT the strand
-   * row's shared `MemberPrivateKey` (see `StrandDatabaseConfig.partyMemberPrivateKey`).
-   * Absent for open strands, and for joiners with no persisted party key.
-   */
-  partyMemberPrivateKey?: string;
-  /**
-   * Number of nodes Optimystic is told this strand's replication cluster should
-   * have. Same rule as {@link CadreNodeConfig.strandClusterSize}, which CadreNode
-   * forwards here: every node on the strand should use the same value, and it is
-   * frozen when the strand's libp2p node is created. Defaults to
-   * `DEFAULT_STRAND_CLUSTER_SIZE` (4); values below `MIN_CLUSTER_SIZE` (2) are
-   * rejected by `resolveStrandClusterSize`. Strand-only — the
-   * control network's breadth is the fixed `CONTROL_REPLICATION_BREADTH`.
-   */
-  clusterSize?: number;
-  /**
-   * Machines that SERVE THIS STRAND (this node included), for the strand node's
-   * block-repair corroboration yardstick — see `resolveRepairYardstick`. It must be
-   * an authenticated per-strand count: the machines that actually run a node for this
-   * strand, not the machines that exist.
-   *
-   * **The party's enrolled-machine count is NOT this number and must never be passed
-   * here.** A strand launches only on machines whose embedding app registered its sApp
-   * config (`CadreNode.addStrand`), so a closed strand shared by two machines of a
-   * three-machine party is served by two. Passing three over-declares, and
-   * over-declaring is the unsafe direction: at a declaration of 3 or more Optimystic
-   * pins the repair corroboration floor at two corroborating peers, which a cohort
-   * that can only ever field one peer can never reach — so the strand can never repair
-   * a block (`cluster-fetch:no-quorum`, surfacing as reads failing with `Missing
-   * block`). That regression is why this field was renamed off "enrolledMachines":
-   * `bug-strand-yardstick-counts-party-machines`.
-   *
-   * **Nothing feeds it in production today.** No authenticated per-strand serving count
-   * exists yet, so `CadreNode` passes nothing and the strand node runs the frozen
-   * `STRAND_CLUSTER_POLICY` — declaring no yardstick, which leaves the known,
-   * upstream-tracked single-voter exposure
-   * (`backlog/debt-read-repair-single-voter-corroboration`). Building the count is
-   * `backlog/feat-strand-yardstick-from-serving-machines`; this field and its
-   * threading are the seam it plugs into.
-   *
-   * Volatile when a source does exist: re-resolve it on every resume beside the cohort
-   * seed, since machines join and leave a strand while it hibernates. Omitting it means
-   * "this node does not know", which declares nothing.
-   */
-  servingMachines?: number;
-  /**
-   * Tuning for the strand peer-join block catch-up ({@link PeerJoinBackfill}),
-   * forwarded from {@link CadreNodeConfig.strandBackfill}. When the strand's
-   * libp2p node connects to a peer this runtime has not yet caught up, every
-   * block in the strand's own raw store is pushed to it. Runs only on strands
-   * with per-strand storage; `{ enabled: false }` restores the pre-existing
-   * no-backfill behaviour.
-   */
-  backfill?: PeerJoinBackfillConfig;
-  /**
-   * Whether this strand's libp2p node is built with Optimystic's change notifications,
-   * forwarded from {@link CadreNodeConfig.strandReactivity} and resolved per strand by
-   * `strandCohortTopicOption`. Retained with the launch config, so a hibernation wake
-   * rebuilds the node with the same setting.
-   */
-  reactivity?: StrandReactivityConfig;
-  /**
-   * Tuning for the CLOSED-strand revoked-peer gate
-   * ({@link StrandRevocationEnforcer}), forwarded from
-   * {@link CadreNodeConfig.strandRevocationEnforcement}. Armed only for
-   * `strandRow.Type === 'c'` (an open strand has no membership rows to derive a
-   * deny set from); `{ enabled: false }` restores the pre-existing behaviour
-   * (a removed party's peers keep being served).
-   */
-  revocationEnforcement?: StrandRevocationEnforcementConfig;
-
-  /**
-   * The staged formation membership invitation for THIS strand — the seam the
-   * bring-up membership reconciler (`strand-membership-reconciler.ts`) reads to
-   * redeem the party's `Strand.Member` seat, and clears once the invitation is
-   * spent, burned, or dead. `CadreNode.launchStrand` wires it over its in-memory
-   * `pendingMembershipInvites` map; read lazily per pass, so a re-formation that
-   * replaces the entry between passes is picked up. Only consulted for closed
-   * strands launched with a {@link partyMemberPrivateKey}; harmless otherwise.
-   */
-  pendingMembershipInvite?: PendingMembershipInviteSource;
-
-  /**
-   * Tuning for the CLOSED-strand membership reconciler, forwarded from
-   * {@link CadreNodeConfig.strandMembershipReconciliation}. Armed only for
-   * closed strands launched with a {@link partyMemberPrivateKey};
-   * `{ enabled: false }` disarms the loop (test fixtures that hand-drive the
-   * membership writers). When it names no `pollIntervalMs` the reconciler
-   * mirrors {@link revocationEnforcement}'s cadence — as its IDLE cadence and
-   * the cap of its unfinished-join retry ladder, not as a flat retry interval.
-   */
-  membershipReconciliation?: StrandMembershipReconciliationConfig;
-
-  /**
-   * Called when THIS node's own peer id turns up in the strand's revoked set —
-   * this node's party was removed from the closed strand, or left it. Fires once
-   * per removal (re-armed if the party is re-admitted; a resume also rebuilds the
-   * enforcer and may legitimately re-fire), and only for closed strands with the
-   * gate armed. `CadreNode`
-   * wires it to its `strand:revoked` event; nothing is stopped or torn down on
-   * this node's behalf — what to do about it is the app's call.
-   */
-  onSelfRevoked?: (strandId: string) => void;
-
-  /**
-   * Called when the membership reconciler reports a blocked re-join — this node holds a
-   * staged membership invitation for the strand that its writes cannot redeem, which is
-   * what a REMOVED party handed a fresh invitation hits (see "Reporting a blocked
-   * re-join" in `strand-membership-reconciler.ts` for the two triggers, one of them a
-   * suspicion rather than a verdict). At most once per re-arm of the loop — a further
-   * invitation staged while the loop is still running does not reset the report — and
-   * only for closed strands with a party key. `CadreNode` wires it to its `strand:rejoin-blocked`
-   * event; nothing is stopped or torn down — a remaining manager admitting this party's
-   * key directly is the remedy, and the loop finishes the join by itself once that lands.
-   */
-  onRejoinBlocked?: (strandId: string) => void;
-
-  /**
-   * Tuning for the JOINING machine's first-sync write gate (`strand-first-sync-gate.ts`),
-   * forwarded from {@link CadreNodeConfig.strandFirstSync}: the Header probe cadence while
-   * a non-founder launch is `'syncing'`, and the default budget {@link whenWritable} waits
-   * before rejecting. Founders and machines that already hold the Header are never gated.
-   */
-  firstSync?: StrandFirstSyncConfig;
-
-  /**
-   * Called when a launch that came up `'syncing'` (see {@link firstSync}) becomes
-   * writable: the strand's `Strand.Header` arrived from a peer — or a founder request
-   * ran the bootstrap against the gated database — and `instance.database` is now set.
-   * Never called for a launch that was writable when `startStrand`/`resumeStrand`
-   * resolved. `CadreNode` wires it to its `strand:writable` event. Retained with the
-   * launch config, so a hibernation resume that comes up gated announces too.
-   */
-  onWritable?: (strandId: string) => void;
-
-  /**
-   * Re-announce this strand's delegate peer id (`delegatePeerId` — the strand
-   * node's own transport peer id) to ONE relay (`relayAddr`, a direct dial addr
-   * with a trailing `/p2p/<relayPeerId>`), unthrottled. Awaited before every
-   * relay-reservation RE-drive of that relay — never before the first attempt,
-   * which the launch/resume seed pass has already announced for.
-   *
-   * Why: a party control node running the relay server admits the strand node on
-   * an in-memory delegate grant that a relay restart drops, so a re-drive that did
-   * not re-announce first would be denied at the relay's gate. `CadreNode` wires
-   * its own control-mesh announce here; a throw or rejection is logged by the
-   * supervisor and the re-drive still runs. Retained with the launch config, so a
-   * hibernation wake's rebuilt supervisors carry it too.
-   */
-  announceDelegateToRelay?: (strandId: string, relayAddr: string, delegatePeerId: string) => Promise<void>;
-
-  /**
-   * The node-local strand network state (`strand-network-state.ts`): this strand's
-   * slice of it is handed to db-p2p as `persistence`, so the strand node saves its
-   * FRET routing table (with each peer's signed address record) on every connection
-   * and re-imports it when it is next built. Absent ⇒ nothing is saved and the node
-   * starts with an empty table. `CadreNode` passes its store. Retained with the launch
-   * config, so a hibernation wake rebuilds the node over the same state.
-   */
-  networkState?: StrandNetworkStateStore;
-}
-
-/**
- * Volatile inputs re-resolved when resuming a quiesced strand — the ones that can
- * have moved since it last ran, and that the rebuilt libp2p node freezes again. Each
- * one omitted keeps the value the retained launch config already holds, so a resume
- * that passes nothing rebuilds the strand exactly as it last ran.
- */
-export interface ResumeStrandOverrides {
-  /**
-   * Freshly-resolved cohort discovery seed (multiaddr strings). Grows as peers are
-   * learned since the strand first launched.
-   */
-  bootstrapNodes?: string[];
-  /**
-   * Freshly-read count of the machines serving this strand, for the repair yardstick
-   * (see {@link StartStrandConfig.servingMachines} — including why the party's
-   * enrolled-machine count is not it). Moves whenever a machine starts or stops
-   * serving the strand, which a hibernating strand does not otherwise notice. Nothing
-   * passes it today. Omitting it here does NOT mean "unknown" the way omitting it from
-   * {@link StartStrandConfig} does — it retains the last value; see `resumeStrand`.
-   */
-  servingMachines?: number;
-}
-
-/** A {@link StrandInstanceManager.resumeStrand} or {@link StrandInstanceManager.quiesceStrand} of one strand. */
-type RuntimeTransition =
-  | { kind: 'resume'; operation: Promise<StrandInstance> }
-  | { kind: 'quiesce'; operation: Promise<void> };
-
-/**
- * Get the isolated storage path for a specific strand.
- *
- * @deprecated This helper is Node-only and throws in React Native (it assumes a
- * filesystem layout). Use a storage provider factory function instead, which
- * receives the strandId and can create strand-specific storage paths using
- * platform-appropriate methods.
- *
- * @example
- * // Instead of using getStrandStoragePath, use a storage provider factory:
- * const storage = {
- *   provider: (strandId: string) => new FileRawStorage(`./data/strands/${strandId}`)
- * };
- */
-export function getStrandStoragePath(basePath: string, strandId: string): string {
-  // Check if we're in a Node.js environment
-  if (typeof process === 'undefined' || !process.versions?.node) {
-    throw new Error(
-      'getStrandStoragePath is not available in React Native. ' +
-      'Use a storage provider factory function instead.'
-    );
-  }
-
-  // Sanitize strandId for filesystem safety (UUIDs should be safe, but just in case)
-  const safeId = strandId.replace(/[^a-zA-Z0-9-]/g, '_');
-
-  // Build the path with plain string joins rather than the Node `path` module.
-  // A static `require('path')` forces RN bundlers (e.g. Metro) to *resolve* the
-  // module at bundle time even though this Node-only helper throws above before
-  // ever reaching here — joining by hand keeps the module free of any Node
-  // built-in reference, so RN bundles need no `path` shim.
-  const trimmedBase = basePath.replace(/[\\/]+$/, '');
-  return `${trimmedBase}/strands/${safeId}`;
-}
-
-/**
- * Resolve a storage provider for a specific strand.
- * If the provider is a factory function, call it with the strandId.
- *
- * Called only from {@link StrandInstanceManager.startStrand}, which owns the result
- * for the instance's lifetime — see `strandStorages`, and which has already asserted
- * the id is usable as a scope key. Do not move that assertion here: this function
- * returns early when no provider is configured, and the id becomes a libp2p protocol
- * prefix on that path too.
- *
- * @param provider - Storage provider (instance or factory)
- * @param strandId - The strand ID to create storage for
- * @returns The resolved IRawStorage instance, or undefined if no provider
- */
-function resolveStrandStorage(
-  provider: RawStorageProvider | undefined,
-  strandId: string
-): IRawStorage | undefined {
-  if (!provider) {
-    return undefined;
-  }
-  const storage = typeof provider === 'function' ? provider(strandId) : provider;
-  // Wrapped in the write-through raw-storage cache (quereus-plugin-sereus's cached-storage.ts).
-  // Called ONCE per strand launch — `startStrand` keeps the result for the instance's
-  // lifetime — so the wrap survives every runtime rebuild the instance goes through.
-  return wrapStorageWithCache(storage, strandId);
-}
-
-/**
- * Verify an sApp's schema signature (fail-closed unless `requireSignedSchemas` is
- * false) and project the config onto the instance's `sAppInfo`.
- */
-function verifiedSAppInfo(strandId: string, sAppConfig: SAppConfig, requireSignedSchemas: boolean | undefined): SAppInfo {
-  assertSchemaSignature(sAppConfig, { requireSignature: requireSignedSchemas ?? true });
-  log('Strand %s sApp schema signature verified (author: %s)', strandId, sAppConfig.id);
-  return {
-    id: sAppConfig.id,
-    version: sAppConfig.version,
-    schema: sAppConfig.schema,
-    signature: sAppConfig.signature
-  };
-}
-
-/** Log wording for what a launch runs: the sApp and its version, or a storage replica. */
-function describeSApp(sAppConfig: SAppConfig | undefined): string {
-  return sAppConfig ? `sApp: ${sAppConfig.id} v${sAppConfig.version}` : 'storage replica';
-}
-
-/**
- * Manages individual strand instances - creates and destroys isolated libp2p nodes
- * for each strand the cadre participates in.
- */
-export class StrandInstanceManager {
-  private instances: Map<string, StrandInstance> = new Map();
-  /**
-   * Retained launch config per strand, captured in `startStrand` and cleared in
-   * `stopStrand`. `resumeStrand` reuses it to rebuild a quiesced strand's runtime
-   * without the caller re-threading storage/network/profile/key/sApp config.
-   */
-  private launchConfigs: Map<string, StartStrandConfig> = new Map();
-  /**
-   * The per-strand peer-join block catch-up, keyed by strand id. Private (not on
-   * the public {@link StrandInstance}) because it is runtime plumbing with the
-   * same lifetime as the strand's libp2p node: created in `buildStrandRuntime`,
-   * stopped and dropped in `releaseRuntime` — so quiesce → resume rebuilds it
-   * with a fresh caught-up-peer memo, which is intended (a resumed node may have
-   * missed writes).
-   */
-  private backfills: Map<string, PeerJoinBackfill> = new Map();
-  /**
-   * The per-strand revoked-peer enforcer (closed strands only), keyed by strand
-   * id. Same lifecycle rationale as {@link backfills}: created in
-   * `buildStrandRuntime`, stopped and dropped in `releaseRuntime`, so quiesce →
-   * resume rebuilds it with a fresh (initially empty, fail-open) snapshot.
-   */
-  private revocationEnforcers: Map<string, StrandRevocationEnforcer> = new Map();
-  /**
-   * The per-strand membership reconciler (closed strands launched with a party
-   * key only), keyed by strand id. Same lifecycle rationale as {@link backfills}:
-   * created in `buildStrandRuntime`, stopped and dropped in `releaseRuntime`, so
-   * quiesce → resume rebuilds it and re-runs the idempotent join ladder (redeem a
-   * staged invitation, write this machine's own `MemberPeer` binding) from scratch.
-   */
-  private membershipReconcilers: Map<string, StrandMembershipReconciler> = new Map();
-  /**
-   * The per-strand first-sync gate (`strand-first-sync-gate.ts`), keyed by strand id:
-   * holds a launch's `StrandDatabase` while it is NOT yet published on the instance.
-   * Present from the moment the database object exists in `buildStrandRuntime` until
-   * {@link publishDatabase} hands it to the app — so it doubles as the rollback handle
-   * for a failed `initialize()` — and, for a joiner that holds no `Strand.Header` yet,
-   * for as long as the launch stays `'syncing'`. Same lifecycle as {@link backfills}:
-   * stopped (its database closed) and dropped in `releaseRuntime`, so a quiesce →
-   * resume rebuild re-probes over the same store. An entry and `instance.database`
-   * are mutually exclusive.
-   */
-  private firstSyncGates: Map<string, StrandFirstSyncGate> = new Map();
-  /**
-   * Callers blocked in {@link whenWritable}, keyed by strand id. Resolved by
-   * {@link publishDatabase}, rejected by {@link stopStrand}; each waiter also carries
-   * its own timeout. Deliberately NOT cleared by `releaseRuntime`: a waiter outlives a
-   * quiesce → resume cycle, since the rebuild is what may finally publish the database.
-   */
-  private writableWaiters: Map<string, Set<WritableWaiter>> = new Map();
-  /**
-   * The per-strand relay-reservation supervisors — ONE PER CONFIGURED RELAY, each
-   * over the node's own bare `/p2p-circuit` listener for that relay
-   * (`strand-network-config.ts`) — keyed by strand id. Same lifecycle rationale as
-   * {@link backfills}: started in `buildStrandRuntime` right after the libp2p node
-   * exists, stopped FIRST and dropped in `releaseRuntime` (before the database
-   * closes and the node stops, so no re-drive dials a node being torn down), so
-   * quiesce → resume rebuilds them over the new node. What makes a strand node's
-   * lost relay slot come back on its own — the control node has the same loop in
-   * `CadreNode.reserveRelays`.
-   */
-  private relaySupervisors: Map<string, RelayReservationSupervisor[]> = new Map();
-  /**
-   * The resolved (cache-wrapped) raw storage per strand id — the instance's OWN
-   * store, resolved once in `startStrand` and held until `stopStrand` disposes it.
-   * Private for the same reason `backfills` is: runtime plumbing, not part of the
-   * public {@link StrandInstance}.
-   *
-   * Deliberately NOT touched by `releaseRuntime`: that is what lets a quiesce →
-   * resume cycle rebuild the libp2p node over the SAME store, keeping its
-   * write-through cache warm (and, on an in-memory backend, keeping the strand's
-   * blocks at all). An entry exists iff `instances` does AND the launch config
-   * supplied a storage provider.
-   *
-   * NOTE: a hibernating strand therefore keeps its store — and its share of the
-   * process-wide cache pool — resident for as long as the instance is tracked. That
-   * is the point (a warm wake), and the pool evicts under pressure, so the cost is
-   * one map entry per hibernating strand today. If a device ever hibernates strands
-   * by the hundred, revisit: dropping the store at quiesce and paying for a cold
-   * wake becomes the better trade.
-   */
-  private strandStorages: Map<string, IRawStorage> = new Map();
-  /**
-   * The runtime build in flight per strand id (`buildStrandRuntime`, from `startStrand` or
-   * `resumeStrand`), present until it settles. The instance is tracked for the whole build,
-   * with its database not yet constructed or not yet initialized, so a caller that must act
-   * on the FINISHED runtime ({@link attachSApp}, {@link whenRuntimeBuilt}) waits on this instead. A failed build's
-   * rejection belongs to the call that started it.
-   */
-  private runtimeBuilds: Map<string, Promise<void>> = new Map();
-  /**
-   * The latest resume or quiesce requested per strand id, present until it settles. Each one
-   * starts only once the one requested before it has settled, so a wake that lands during a
-   * quiesce rebuilds after it, and a quiesce requested during a rebuild releases what that
-   * rebuild finished. The WHOLE operation, not its entry in {@link runtimeBuilds}: that settles
-   * before a resume's own catch records `'error'`, so a joiner awaiting it could read a stale
-   * status.
-   */
-  private transitions: Map<string, RuntimeTransition> = new Map();
-  /**
-   * The pending rebuild per strand id of a launch whose database could not be built
-   * because no seed peer had answered yet ({@link deferUnreachableJoin}). Cleared by
-   * {@link releaseRuntime}, so a stop, a quiesce or a resume cancels it.
-   */
-  private unreachableJoinRetries: Map<string, unknown> = new Map();
-  private stopping = false;
-
-  constructor() {
-    log('StrandInstanceManager created');
-  }
-
-  /**
-   * Get all current strand instances
-   */
-  getInstances(): Map<string, StrandInstance> {
-    return new Map(this.instances);
-  }
-
-  /**
-   * Get a specific strand instance
-   */
-  getInstance(strandId: string): StrandInstance | undefined {
-    return this.instances.get(strandId);
-  }
-
-  /**
-   * Check if a strand is currently running
-   */
-  hasStrand(strandId: string): boolean {
-    return this.instances.has(strandId);
-  }
-
-  /**
-   * Start a new strand instance.
-   *
-   * A failed launch leaves NOTHING tracked: the instance and its retained launch
-   * config are both dropped before the error is rethrown, so the strand id is
-   * free for a genuine retry. This matches the pre-registration failure path
-   * (a rejected schema signature, which throws before anything is recorded) —
-   * both failure modes of this call leave the same residue: none. Callers learn
-   * of the failure from the rejected promise (and, on the control-discovered
-   * path, from CadreNode's `strand:error` event), not from an error record left
-   * behind in `instances`.
-   *
-   * The strand's raw storage is resolved HERE, once, and held for the instance's
-   * lifetime (see `strandStorages`) — `buildStrandRuntime` only reads it, so a
-   * hibernation wake never re-enters the embedder's provider.
-   */
-  async startStrand(config: StartStrandConfig): Promise<StrandInstance> {
-    const { strandRow, sAppConfig } = config;
-    const strandId = strandRow.Id;
-
-    if (this.stopping) {
-      throw new Error('StrandInstanceManager is stopping');
-    }
-
-    if (this.instances.has(strandId)) {
-      // Callers resolve founder-ness BEFORE reaching here (CadreNode.launchStrand) and
-      // honor a founder request on a tracked instance via foundExistingStrand — so a
-      // founder flag arriving at this early return against a non-founder retained
-      // config is a dropped bootstrap, and must never again be silent.
-      if (config.founder === true && this.launchConfigs.get(strandId)?.founder !== true) {
-        log('startStrand: strand %s is already running but was NOT launched as founder — ' +
-          'this early return DROPS the founder request; use foundExistingStrand', strandId);
-      }
-      log('Strand %s already running', strandId);
-      return this.instances.get(strandId)!;
-    }
-
-    log('Starting strand instance: %s (%s)', strandId, describeSApp(sAppConfig));
-    const tTotal = performance.now();
-
-    // The id becomes two names below: the storage scope key the embedder's provider
-    // turns into a directory or database name, and `networkName` in buildStrandRuntime,
-    // from which the libp2p protocol prefix `/optimystic/strand-<id>` is built. A
-    // replicated strand row carries whatever id the founding node wrote, so neither is
-    // safe unchecked. Asserted HERE rather than in `resolveStrandStorage` because that
-    // function returns early when no provider is configured — a node with no storage
-    // still reaches buildStrandRuntime and still mints the protocol prefix.
-    assertStrandScopeKey(strandId);
-
-    const sAppInfo = sAppConfig ? verifiedSAppInfo(strandId, sAppConfig, config.requireSignedSchemas) : undefined;
-
-    // Resolve this strand's storage ONCE, before anything is recorded. If a factory
-    // function is provided, it is called with the strandId to create strand-specific
-    // storage (e.g. strand-isolated directories). A provider that throws therefore
-    // fails the launch alongside the schema-signature check, leaving nothing tracked.
-    const strandStorage = resolveStrandStorage(config.storage?.provider, strandId);
-    if (strandStorage) {
-      log('Strand %s using provided storage provider', strandId);
-    }
-
-    // Determine latency hint: sApp config > default
-    const latencyHint = sAppConfig?.latencyHint ?? config.defaultLatencyHint;
-
-    const instance: StrandInstance = {
-      strandId,
-      status: 'starting',
-      sAppInfo,
-      memberPrivateKey: strandRow.MemberPrivateKey ?? undefined,
-      connectedPeers: 0,
-      lastActivity: new Date(),
-      latencyHint
-    };
-
-    this.instances.set(strandId, instance);
-    this.launchConfigs.set(strandId, config);
-    if (strandStorage) {
-      this.strandStorages.set(strandId, strandStorage);
-    }
-
-    try {
-      await this.trackRuntimeBuild(strandId, this.buildStrandRuntime(instance, config));
-      timing('[startStrand:%s] total: %dms', strandId, Math.round(performance.now() - tTotal));
-      log('Strand %s started successfully (%s)', strandId, describeSApp(sAppConfig));
-      return instance;
-    } catch (error) {
-      // Status/error first — the (now discarded) record is still what `log` reports on.
-      instance.status = 'error';
-      instance.error = error instanceof Error ? error.message : String(error);
-      log('Failed to start strand %s: %s', strandId, instance.error);
-      // Drop the dead record so this strand id can be launched again. Keep the
-      // `launchConfigs` has an entry iff `instances` does invariant — resumeStrand
-      // reads both, and a config without an instance would strand the config.
-      this.instances.delete(strandId);
-      this.launchConfigs.delete(strandId);
-      // The store this launch resolved goes with it: nothing owns it any more, and a
-      // retained cache wrapper would be handed back (already retired) on a retry.
-      await this.disposeStrandStorage(strandId);
-      throw error;
-    }
-  }
-
-  /** Hold `build` in {@link runtimeBuilds} until it settles, and settle as it does. */
-  private async trackRuntimeBuild(strandId: string, build: Promise<void>): Promise<void> {
-    this.runtimeBuilds.set(strandId, build);
-    try {
-      await build;
-    } finally {
-      if (this.runtimeBuilds.get(strandId) === build) {
-        this.runtimeBuilds.delete(strandId);
-      }
-    }
-  }
-
-  /**
-   * Wait until no runtime build of `strandId` is in flight. Never rejects: a failed build
-   * is reported by the call that started it, and the caller re-reads the instance after.
-   *
-   * NOTE: after a failed launch the re-read sees the instance gone only because
-   * `startStrand`'s catch deletes it synchronously, and its continuation is queued before
-   * this waiter's. If that cleanup ever awaits before `instances.delete`, a waiter can act
-   * on the dying record; move the cleanup inside the tracked promise then.
-   */
-  private async settleRuntimeBuilds(strandId: string): Promise<void> {
-    for (let build = this.runtimeBuilds.get(strandId); build; build = this.runtimeBuilds.get(strandId)) {
-      await build.catch((error: unknown) => {
-        log('Strand %s: the runtime build waited on failed (reported by its launcher): %o', strandId, error);
-      });
-    }
-  }
-
-  /**
-   * Build (or rebuild) the libp2p node + StrandDatabase for an instance and
-   * attach them, transitioning it to `active`. Shared by `startStrand` (fresh
-   * launch) and `resumeStrand` (rehydrating a quiesced instance). Reads all
-   * volatile inputs (bootstrapNodes, servingMachines, network, profile,
-   * privateKey, sApp config) from `config`, so the caller controls the
-   * cohort-derived values. Storage is the
-   * one input it does NOT re-read from `config`: that belongs to the instance and
-   * comes from `strandStorages`.
-   */
-  private async buildStrandRuntime(instance: StrandInstance, config: StartStrandConfig): Promise<void> {
-    const strandId = instance.strandId;
-    const { sAppConfig } = config;
-
-    // Refuse a bad `linkRoundTripMs` before any strand bring-up, for the reason
-    // `CadreNode.start()` does: every budget derived from it here is behind a condition
-    // (backfill disabled, no relay addrs), so a zero or NaN declaration would otherwise
-    // surface later inside a path that logs and carries on. `link-budget.ts`.
-    resolveLinkRoundTripMs(config.network?.linkRoundTripMs);
-
-    // The store the instance OWNS (resolved once in `startStrand`), not a fresh
-    // resolution: a rebuild must reach the same backend through the same warm cache.
-    const strandStorage = this.strandStorages.get(strandId);
-
-    // db-p2p namespaces every one of the node's protocol ids by network name
-    // (`/optimystic/<networkName>/...`), so anything dialing this node's own
-    // services must derive its prefix from the SAME string the node was built
-    // with — hence one binding for both, not two literals that can drift.
-    const networkName = `strand-${strandId}`;
-    // Set when the database build failed only because no seed peer has answered yet; see
-    // {@link deferUnreachableJoin}.
-    let joinWaitsForSeed = false;
-    const protocolPrefix = `/optimystic/${networkName}`;
-
-    // Whether this node runs the circuit-relay server, and its init — the same
-    // resolution the control node takes from the same config (`relay-server.ts`).
-    const relayServer = resolveRelayServer(config.network, config.profile);
-    // The strand-node VIEW of the machine's one `NetworkConfig`, not the control
-    // node's resolution: fixed direct listen ports become ephemeral (two nodes cannot
-    // bind one port), the announce config is dropped (it names the control node's
-    // address), and every configured relay becomes one bare `/p2p-circuit` SEARCH
-    // listener plus a dial addr in `relayAddrs` for the per-relay supervisor started
-    // below — the same route the control node takes, because the configured
-    // `<relay>/p2p-circuit` shape loses its address on a relay restart, a hangup, or
-    // libp2p's own reservation refresh and recovers from none of them. See
-    // `strand-network-config.ts` for all three, and `relay-reservation.ts`.
-    //
-    // `relayAddrs` is destructured OFF: it is not a `createLibp2pNode` option, and
-    // spreading it would hand db-p2p a key it does not know.
-    const { relayAddrs: relayDialAddrs = [], ...addrOptions } = strandNodeAddrs(config.network);
-
-    // The CLOSED-strand revoked-peer gate (see strand-revocation-enforcer.ts).
-    // Constructed BEFORE the libp2p node because the node's options embed its
-    // predicates; its deny set starts empty (admit everything — fail-open, same
-    // posture as bring-up everywhere here) and is first populated after the
-    // strand database initializes below. Registered in the map immediately so
-    // the failure rollback (releaseRuntime) tears it down like every other
-    // runtime component. Open strands skip it entirely: their Member/MemberPeer
-    // tables are empty by schema (`OnlyClosed`), so there is nothing to derive
-    // a deny set from — arming would be pointless polling.
-    const revocationEnforcer =
-      config.strandRow.Type === 'c' && config.revocationEnforcement?.enabled !== false
-        ? new StrandRevocationEnforcer({
-            label: strandId,
-            readRows: () => {
-              const database = instance.database;
-              if (!database) {
-                throw new Error(`strand ${strandId} has no live database`);
-              }
-              return readStrandRevocationRows(database.getDatabase());
-            },
-            // Read per sweep, never captured: this closure is built BEFORE the
-            // libp2p node exists (its options embed the enforcer's predicates),
-            // and `releaseRuntime` clears the field again on quiesce. Undefined
-            // therefore means "no transport right now", which the enforcer
-            // treats as "nothing to tear down".
-            getNetwork: () => instance.libp2pNode,
-            onSelfRevoked: () => config.onSelfRevoked?.(strandId)
-          }, config.revocationEnforcement)
-        : undefined;
-    if (revocationEnforcer) {
-      this.revocationEnforcers.set(strandId, revocationEnforcer);
-    }
-
-    // Closed strands compose revoked-peer denial onto the caller's gater and arm
-    // the fail-closed per-stream gate; open strands keep the raw configured gater
-    // (their peers are legitimately cross-party and nothing is ever revoked).
-    const revocationGateOptions = revocationEnforcer
-      ? {
-          connectionGater: createRevocationConnectionGater(revocationEnforcer, config.network?.connectionGater),
-          authorizeInboundStream: (remotePeerId: string, protocol: string) =>
-            revocationEnforcer.authorizeStream(remotePeerId, protocol)
-        }
-      : config.network?.connectionGater
-        ? { connectionGater: config.network.connectionGater }
-        : {};
-
-    const reactivityOption = strandCohortTopicOption(config.reactivity, strandId);
-    if (reactivityOption.cohortTopic) {
-      log('strand %s: change notifications on', strandId);
-    }
-
-    try {
-      // Bound once: the breadth is also the ceiling on the repair yardstick below, and the
-      // two must be derived from the same resolution. Inside the `try` deliberately — a
-      // rejected clusterSize is a build failure that runs the same cleanup as any other.
-      const strandClusterSize = resolveStrandClusterSize(config.clusterSize);
-
-      let t0 = performance.now();
-      const node = await createLibp2pNode({
-        port: 0, // Random port
-        bootstrapNodes: config.bootstrapNodes ?? [],
-        networkName,
-        storage: strandStorage,
-        fretProfile: config.profile === 'storage' ? 'core' : 'edge',
-        relay: relayServer.enabled,
-        // NOTE: unlike the control node, a strand node has no unauthorized-reservation budget,
-        // so with the per-connection cap off any peer that reaches this node can hold one of
-        // `maxReservations` slots and forward uncapped through it. If strand relays are ever
-        // abused for bandwidth, give strand nodes a budget or a capped init of their own.
-        ...(relayServer.enabled && { relayServerInit: relayServer.init }),
-        clusterSize: strandClusterSize,
-        // Deliberately NOT CONTROL_CLUSTER_POLICY: a strand is application data with its own
-        // breadth reasoning, and the shape match with the control policy is a coincidence.
-        //
-        // The builder declares this node's block-repair corroboration yardstick from the
-        // count of machines SERVING this strand, capped at the breadth above (a block
-        // never lives on more machines than the cohort is wide). Given no count — the
-        // production path today, since no per-strand serving count exists yet — it
-        // returns the frozen STRAND_CLUSTER_POLICY itself, declaring nothing. Resolved
-        // HERE rather than at `startStrand`, so a wake from hibernation would pick up a
-        // serving set that changed while the strand slept.
-        // The read deadline is settled by the same helper the control node uses, because the
-        // two networks ride one link: the host's own, else derived from its declared
-        // `linkRoundTripMs`, else nothing, and the base policy's COHORT_READ_DEADLINE_MS stands.
-        clusterPolicy: strandClusterPolicy(strandClusterSize, {
-          servingMachines: config.servingMachines,
-          cohortQueryTimeoutMs: declaredCohortReadDeadlineMs(config.network)
-        }),
-        arachnode: {
-          enableRingZulu: config.profile === 'storage'
-        },
-        ...(config.privateKey && { privateKey: config.privateKey }),
-        // Built per runtime, not per launch: the adapter stops saving once the strand's
-        // state is forgotten, and a rebuilt runtime is what starts it saving again.
-        //
-        // NOTE: the saved state also restores which peers db-p2p saw serving the strand,
-        // and it counts them in the cohort until FRET finds them unreachable. A node built
-        // while every other member is offline therefore fails its first writes ("Failed to
-        // get super-majority") — for 4-6 s on loopback, the window a running node already
-        // sees when a member drops. If apps report failed sends after relaunching alone on
-        // a slow relayed link, retry the write in the app or raise it upstream (db-p2p's
-        // remembered-serving fallback); dropping `persistence` brings back the split
-        // restart.
-        ...(config.networkState && { persistence: strandNetworkStatePersistence(config.networkState, strandId) }),
-        ...(config.network?.transports && { transports: config.network.transports }),
-        ...(config.network?.noiseCrypto && { noiseCrypto: config.network.noiseCrypto }),
-        // Unconditional, and the same default the control node takes: every node of
-        // every party runs the monitor over its connection to a slow phone, so the
-        // widened ping deadline has to reach the strand nodes too (see
-        // DEFAULT_CONNECTION_MONITOR).
-        connectionMonitor: config.network?.connectionMonitor ?? DEFAULT_CONNECTION_MONITOR,
-        // The same declared link the control node states, and for the same reason: a strand
-        // node is also the listener for every other member's strand node, and Optimystic
-        // derives that listener's limit from this (`link-budget.ts`).
-        linkRoundTripMs: resolveLinkRoundTripMs(config.network?.linkRoundTripMs),
-        // The same dial limits as the control node: a strand node dials the same machines,
-        // through the same relays and gates.
-        ...optimysticDialLimits(config.network?.linkRoundTripMs),
-        // Listen entries plus the WebSocket transport switch they imply — a strand node
-        // announces nothing the operator configured (`strand-network-config.ts`), and
-        // spreads AFTER `transports` above because the switch is a no-op whenever the
-        // embedder supplied its own factories. The bare `/p2p-circuit` entries (one per
-        // relay) open no connection here; each registers a pending reservation that the
-        // matching supervisor below fills, which is what gives a NAT'd strand node a
-        // reachable relay slot — and works against a party-run relay because the launch
-        // path announced this strand's derived peerId to it first (delegate admission;
-        // see cadre-node.ts).
-        ...addrOptions,
-        // The raw configured gater (open strands), or the revocation-composed
-        // gater plus the fail-closed per-stream revoked-peer gate (closed
-        // strands) — resolved above, before the try.
-        ...revocationGateOptions,
-        // `{}` unless the operator enabled change notifications for this strand (`strand-reactivity.ts`).
-        ...reactivityOption
-      }) as Libp2pNodeWithRepo;
-      timing('[buildStrandRuntime:%s] createLibp2pNode: %dms', strandId, Math.round(performance.now() - t0));
-
-      instance.libp2pNode = node;
-
-      // One reservation supervisor PER RELAY, started now so the first attempts
-      // overlap the database bring-up below (the relay is never in a strand's
-      // Optimystic cohort — its protocol ids are namespaced `/optimystic/strand-<id>/…`
-      // — so an open relay connection cannot disturb that bring-up). Their first
-      // attempts are awaited before the strand goes `active`, below. Registered in the
-      // map immediately so the failure rollback (releaseRuntime) stops them like every
-      // other runtime component.
-      this.relaySupervisors.set(strandId, this.startRelaySupervisors(strandId, node, relayDialAddrs, config));
-
-      // Create and initialize the StrandDatabase.
-      //
-      // Held by the first-sync gate (registered before initialize) rather than attached
-      // to the instance: a failed init is then cleaned up by releaseRuntime below through
-      // the gate (close() is safe on a partially-initialized db), and the app only ever
-      // sees `instance.database` once this machine is allowed to write to the strand.
-      t0 = performance.now();
-      const strandDb = new StrandDatabase({
-        strandId,
-        sAppConfig,
-        libp2pNode: node,
-        coordinatedRepo: node.coordinatedRepo,
-        // Founder bootstrap inputs: the strand's type drives which membership rows
-        // are written, and the PARTY's own key (partyMemberPrivateKey, resolved by the
-        // caller from the control-layer StrandPartyKey row) derives the founding
-        // Member/Manager key. The row's shared MemberPrivateKey rides along as the
-        // read-gating secret only — it derives nobody's identity.
-        strandType: config.strandRow.Type,
-        memberPrivateKey: config.strandRow.MemberPrivateKey ?? undefined,
-        partyMemberPrivateKey: config.partyMemberPrivateKey,
-        founder: config.founder
-      });
-      const gate = new StrandFirstSyncGate({
-        label: strandId,
-        database: strandDb,
-        onHeaderHeld: () => this.publishDatabase(instance, strandDb)
-      }, config.firstSync);
-      this.firstSyncGates.set(strandId, gate);
-      try {
-        await strandDb.initialize();
-      } catch (error) {
-        joinWaitsForSeed = isCohortUnreachableRead(error);
-        throw error;
-      }
-      timing('[buildStrandRuntime:%s] strandDatabase.initialize: %dms', strandId, Math.round(performance.now() - t0));
-
-      // The first-sync write gate. A founder just wrote the Header in its bootstrap; any
-      // other machine may commit only once it holds the Header — which it can only have
-      // received from a peer (now, or on an earlier run over the same store) — and has read
-      // each App table once, so their collections are fetched too. Publishing the database
-      // is what makes the strand writable to the app; until then the instance stays
-      // `'syncing'` and the gate re-probes on its cadence. Everything armed below reads
-      // `instance.database` lazily, so the reconciler and enforcer wait with it — and
-      // `publishDatabase` kicks the reconciler when the wait ends, so a joiner's membership
-      // rows land with the strand rather than one retry later.
-      t0 = performance.now();
-      if (config.founder === true || await strandFirstSyncComplete(strandDb.getDatabase(), strandId)) {
-        this.publishDatabase(instance, strandDb);
-      } else {
-        gate.start();
-      }
-      timing('[buildStrandRuntime:%s] first-sync probe: %dms', strandId, Math.round(performance.now() - t0));
-
-      // The membership reconciler: finish this party's join on every machine —
-      // redeem a staged formation invitation if one is pending, then write the
-      // durable machine→party binding (Strand.MemberPeer). Armed here, the one
-      // seam where the strand's transport peer id and a live Database both exist
-      // for launch AND hibernation resume, for closed strands that carry the
-      // party's own membership key. start() kicks an immediate pass WITHOUT
-      // awaiting it — a joiner at this instant has not synced the founder's rows
-      // and may lack write quorum, so bring-up is never blocked; an unfinished join
-      // then retries on a short doubling ladder capped at the configured
-      // pollIntervalMs (which is also the flat cadence while nobody has admitted this
-      // party yet), and the loop stops once the member row and this machine's own
-      // binding are both in place. Armed AFTER the gate above, so a gated launch
-      // always has a reconciler registered by the time the gate can publish.
-      if (config.strandRow.Type === 'c' && config.partyMemberPrivateKey
-        && config.membershipReconciliation?.enabled !== false) {
-        const reconciler = new StrandMembershipReconciler({
-          label: strandId,
-          partyMemberPrivateKey: config.partyMemberPrivateKey,
-          // Read per pass, never captured — same lifecycle argument as the
-          // enforcer's getNetwork: quiesce drops both handles.
-          getDatabase: () => instance.database?.getDatabase(),
-          getOwnPeerId: () => instance.libp2pNode?.peerId.toString(),
-          pendingInvite: config.pendingMembershipInvite,
-          // The enforcer's CURRENT snapshot view of this node's own peer id — the
-          // reconciler stops rather than fight a self-revocation. Absent when the
-          // gate is disarmed (fail-open: the loop just runs).
-          isSelfRevoked: revocationEnforcer
-            ? () => {
-                const ownPeerId = instance.libp2pNode?.peerId.toString();
-                return ownPeerId !== undefined && revocationEnforcer.isRevoked(ownPeerId);
-              }
-            : undefined,
-          onRejoinBlocked: () => config.onRejoinBlocked?.(strandId)
-        }, {
-          pollIntervalMs: config.membershipReconciliation?.pollIntervalMs
-            ?? config.revocationEnforcement?.pollIntervalMs
-        });
-        this.membershipReconcilers.set(strandId, reconciler);
-        reconciler.start();
-      }
-
-      // Arm the revoked-peer deny-set poll now that the database it reads
-      // exists. start() kicks an immediate refresh WITHOUT awaiting it —
-      // bring-up and resume are never blocked on a membership read; until the
-      // first read lands the empty snapshot admits everyone (fail-open).
-      revocationEnforcer?.start();
-
-      // Peer-join block catch-up: push this strand's own blocks to each newly
-      // connected peer, so a machine that joined after blocks were committed
-      // still ends up physically holding them (without per-strand storage there
-      // is nothing to copy). Armed for EVERY stored strand: `PeerJoinBackfill`
-      // only does work when the strand's libp2p node reports a peer connection,
-      // so on a device that is genuinely alone it is inert, and arming it at
-      // launch is what closes the "founded alone, never replicates" hole — a
-      // peer that joins later gets the founder's blocks without any relaunch.
-      // No `authorizePeer` gate, deliberately — see the module comment in
-      // peer-join-backfill.ts for the strand-side argument (and why the control
-      // network, which DOES gate, is different).
-      //
-      // NOTE: cost is one PeerJoinBackfill object + one `peer:identify` listener
-      // per running strand — linear in strand count, negligible at the handful a
-      // device or host runs today. If a node ever hosts strands by the hundred,
-      // move to one shared listener that dispatches by strand id.
-      if (strandStorage && config.backfill?.enabled !== false) {
-        if (node.keyNetwork) {
-          const backfill = new PeerJoinBackfill({
-            label: strandId,
-            libp2p: node,
-            peerNetwork: node.keyNetwork,
-            storage: strandStorage,
-            // The same prefix the receiver registered its block-transfer handler
-            // under — derived from networkName above, never re-spelled here.
-            protocolPrefix
-          }, {
-            // Dial and response deadlines counted in link round trips, not fixed milliseconds:
-            // a relayed dial costs a fixed number of exchanges, so a fixed budget has a link
-            // speed above which this catch-up can never reach the peer at all. See
-            // `link-budget.ts`. Spread BEFORE the host's own config so an explicit
-            // `strandBackfill.dialTimeoutMs` still wins.
-            ...peerJoinPushBudget(config.network?.linkRoundTripMs),
-            ...config.backfill
-          });
-          backfill.start();
-          this.backfills.set(strandId, backfill);
-        } else {
-          log('Strand %s: libp2p node exposes no keyNetwork; peer-join block catch-up is inert', strandId);
-        }
-      }
-
-      // The relay supervisors' FIRST attempts, all of them, before `active`: the happy
-      // path still publishes its circuit addr before `addStrand` resolves (the
-      // strand-addr RPC answers read it, and the same-party circuit scenario asserts
-      // it). Fail-SOFT — a first attempt that lands nothing does not fail the launch.
-      // The strand's database is up, and the supervisor keeps trying on its backoff;
-      // failing here would only trade that for `StrandWatcher`'s full-rebuild retry.
-      //
-      // NOTE: a relay that is down costs this launch one full drive — the reservation budget
-      // counted from the declared link round trip, 18 s at its default (`link-budget.ts`) — and
-      // `StrandWatcher` launches strands one at a time, so N strands cost N of those in
-      // bring-up during a relay outage, and MORE on a host that declared a slower link. If that
-      // ever matters, stop awaiting here (the circuit addr then lands after `active`) rather
-      // than shortening the drive.
-      t0 = performance.now();
-      await this.awaitFirstRelayAttempts(strandId);
-      timing('[buildStrandRuntime:%s] relay first attempts: %dms', strandId, Math.round(performance.now() - t0));
-
-      // No `lastActivity` stamp: bringing a runtime up is not activity. A check-in marks
-      // activity before its resume and must see only what landed while the strand was up.
-      instance.status = liveStrandStatus(instance);
-      if (instance.status === 'syncing') {
-        log('Strand %s launched as a joiner with no Strand.Header held yet — writes are withheld until ' +
-          'a member of the strand is reached', strandId);
-      }
-    } catch (error) {
-      // Roll back any partially-attached runtime so the instance is left with
-      // NEITHER handle. Otherwise the `libp2pNode || database` "already live"
-      // guard in resumeStrand/handleStrandWake would treat a half-built strand
-      // as healthy — leaking the libp2p node and never retrying the rebuild.
-      await this.releaseRuntime(instance).catch((cleanupErr) => {
-        log('buildStrandRuntime cleanup for strand %s also failed: %o', strandId, cleanupErr);
-      });
-      if (joinWaitsForSeed) {
-        this.deferUnreachableJoin(instance, error);
-        return;
-      }
-      throw error;
-    }
-  }
-
-  /**
-   * Keep a launch `'syncing'` when its database could not be built because no seed peer has
-   * answered yet, and rebuild the runtime every {@link UNREACHABLE_JOIN_RETRY_MS}.
-   *
-   * Optimystic (since 1.11.0) will not rule a block never-created while any of a node's
-   * `bootstrapNodes` is unheard, so a joiner alone with an unreachable seed cannot build even
-   * the schema catalog: `strandDb.initialize()` fails with `cohort-unreachable`. A founder
-   * given a seed is refused the same way (normally it has none: no sibling runs the strand yet). Refusing is
-   * right (building it alone would fork the catalog), but failing the launch would drop the
-   * strand until the watcher relaunched it, and `addStrand` would reject with a raw Optimystic
-   * error. Deferring gives a joiner the same contract the first-sync gate gives one that did
-   * build: the strand stays tracked, {@link whenWritable} waits and rejects with the retryable
-   * `StrandAwaitingFirstSyncError`, and the launch proceeds once a seed peer answers. A failed
-   * composition stops the injected libp2p node, so each retry rebuilds the whole runtime.
-   */
-  private deferUnreachableJoin(instance: StrandInstance, error: unknown): void {
-    const strandId = instance.strandId;
-    instance.status = 'syncing';
-    log('Strand %s: no seed peer has answered yet, so this joiner cannot build its database (%s); ' +
-      'retrying in %dms', strandId, error instanceof Error ? error.message : String(error), UNREACHABLE_JOIN_RETRY_MS);
-    const timer = setTimeout(() => {
-      this.unreachableJoinRetries.delete(strandId);
-      void this.retryUnreachableJoin(instance);
-    }, UNREACHABLE_JOIN_RETRY_MS);
-    (timer as { unref?: () => void }).unref?.();
-    this.unreachableJoinRetries.set(strandId, timer);
-  }
-
-  /** One retry scheduled by {@link deferUnreachableJoin}, unless the strand was stopped or rebuilt since. */
-  private async retryUnreachableJoin(instance: StrandInstance): Promise<void> {
-    const strandId = instance.strandId;
-    const config = this.launchConfigs.get(strandId);
-    if (this.stopping || this.instances.get(strandId) !== instance || !config
-      || instance.libp2pNode || this.runtimeBuilds.has(strandId) || this.transitions.has(strandId)) {
-      return;
-    }
-    try {
-      await this.trackRuntimeBuild(strandId, this.buildStrandRuntime(instance, config));
-    } catch (error) {
-      instance.status = 'error';
-      instance.error = error instanceof Error ? error.message : String(error);
-      log('Strand %s: rebuilding a joiner that was waiting for a seed peer failed: %s', strandId, instance.error);
-    }
-  }
-
-  /**
-   * Release an instance's strand-network runtime: stop the relay supervisors and
-   * the other background loops, close the StrandDatabase, then stop the libp2p
-   * node (construction order in reverse), clearing both fields and zeroing
-   * connectedPeers. Tolerant of partially-built state — either handle may be
-   * absent — so it doubles as rollback for a failed `buildStrandRuntime`. Shared
-   * by `quiesceStrand`, `stopStrand`, and that rollback path.
-   *
-   * Leaves `strandStorages` untouched by design — the store outlives the runtime it
-   * was built into, which is what makes a resume warm. Only `stopStrand` disposes it.
-   */
-  private async releaseRuntime(instance: StrandInstance): Promise<void> {
-    // Relay supervisors FIRST of all — before anything below is torn down — so no
-    // re-drive is scheduled against a node being stopped. `stop()` is synchronous
-    // and never awaits a drive, so a relay that is down cannot delay this teardown.
-    // `stop()` also cancels the drive already in flight, so an attempt started
-    // moments before this does not keep dialing and polling the node the lines
-    // below are stopping; its result is discarded either way.
-    const unreachableJoinRetry = this.unreachableJoinRetries.get(instance.strandId);
-    if (unreachableJoinRetry !== undefined) {
-      clearTimeout(unreachableJoinRetry as ReturnType<typeof setTimeout>);
-      this.unreachableJoinRetries.delete(instance.strandId);
-    }
-    const relaySupervisors = this.relaySupervisors.get(instance.strandId);
-    if (relaySupervisors) {
-      relaySupervisors.forEach((supervisor) => supervisor.stop());
-      this.relaySupervisors.delete(instance.strandId);
-    }
-    // Backfill next — before the database closes and the libp2p node stops — so
-    // no NEW catch-up push is issued against a torn-down transport. A push already
-    // in flight is not awaited; it fails into the module's own per-chunk catch.
-    const backfill = this.backfills.get(instance.strandId);
-    if (backfill) {
-      backfill.stop();
-      this.backfills.delete(instance.strandId);
-    }
-    // The revoked-peer enforcer goes with the runtime it gated: a resume
-    // rebuilds it with a fresh (initially empty, fail-open) snapshot.
-    const revocationEnforcer = this.revocationEnforcers.get(instance.strandId);
-    if (revocationEnforcer) {
-      revocationEnforcer.stop();
-      this.revocationEnforcers.delete(instance.strandId);
-    }
-    // The membership reconciler likewise: a resume rebuilds it and re-runs the
-    // idempotent ladder, which is what makes resume-after-partial-join heal.
-    const membershipReconciler = this.membershipReconcilers.get(instance.strandId);
-    if (membershipReconciler) {
-      membershipReconciler.stop();
-      this.membershipReconcilers.delete(instance.strandId);
-    }
-    // A still-gated database (a joiner that never received the Header, or a launch whose
-    // initialize() failed) is closed through its gate; a published one through the
-    // instance. Never both — publishing moves the handle from the gate to the instance.
-    const gate = this.firstSyncGates.get(instance.strandId);
-    if (gate) {
-      gate.stop();
-      this.firstSyncGates.delete(instance.strandId);
-      await gate.database.close();
-    }
-    if (instance.database) {
-      await instance.database.close();
-      instance.database = undefined;
-    }
-    if (instance.libp2pNode) {
-      await instance.libp2pNode.stop();
-      instance.libp2pNode = undefined;
-    }
-    instance.connectedPeers = 0;
-  }
-
-  /**
-   * One {@link superviseRelayReservation} per relay dial addr, each over exactly
-   * that relay so "held" is judged per relay (`circuitMultiaddrsVia`) and losing
-   * one relay re-drives only that one. Default retry timings — the control node's (2 s
-   * doubling to 60 s between failed attempts, a 5 s liveness check while held). Each DRIVE's
-   * own deadline is counted in link round trips at this host's declared
-   * `network.linkRoundTripMs` (`link-budget.ts`), so the same declaration that lengthens the
-   * control node's drive lengthens these.
-   *
-   * Each supervisor's `beforeRedrive` is the caller's
-   * {@link StartStrandConfig.announceDelegateToRelay} for THIS relay and THIS
-   * node's peer id, so a re-drive after a relay restart is preceded by a fresh
-   * delegate grant on the relay about to be dialed. Returns `[]` when no relay is
-   * configured — the strand then simply has no circuit listener to fill.
-   */
-  private startRelaySupervisors(
-    strandId: string,
-    node: Libp2pNodeWithRepo,
-    relayDialAddrs: readonly string[],
-    config: StartStrandConfig
-  ): RelayReservationSupervisor[] {
-    if (relayDialAddrs.length === 0) {
-      return [];
-    }
-    const announce = config.announceDelegateToRelay;
-    const delegatePeerId = node.peerId.toString();
-    const timeoutMs = relayReservationBudgetMs(config.network?.linkRoundTripMs);
-    return relayDialAddrs.map((relayAddr) => superviseRelayReservation(node, [relayAddr], {
-      timeoutMs,
-      ...(announce && { beforeRedrive: () => announce(strandId, relayAddr, delegatePeerId) })
-    }));
-  }
-
-  /**
-   * Wait for every relay supervisor's first attempt to settle, concurrently, and log
-   * the ones that landed nothing. Never throws — a strand whose relay is unreachable
-   * at launch still comes up, with its supervisors retrying in the background.
-   */
-  private async awaitFirstRelayAttempts(strandId: string): Promise<void> {
-    const supervisors = this.relaySupervisors.get(strandId) ?? [];
-    if (supervisors.length === 0) {
-      return;
-    }
-    await Promise.all(supervisors.map((supervisor) => supervisor.firstAttempt));
-    supervisors.forEach((supervisor) => {
-      if (supervisor.lastError !== null) {
-        log('Strand %s: relay reservation not held after the first attempt (retrying in the background): %s',
-          strandId, supervisor.lastError);
-      }
-    });
-  }
-
-  /**
-   * Drop the strand's owned store and release this strand's claim on its cache. The
-   * wrapper counts holders, so the cache is emptied and unregistered from the shared
-   * pool only if no other scope still holds it. Called only where the strand's whole
-   * lifetime ends (`stopStrand`, and the failed-launch rollback) — never on a quiesce.
-   *
-   * A dispose failure is logged, not thrown: the store is already unreferenced here,
-   * and failing the stop over a cache-bookkeeping error would leave the caller unable
-   * to tear the strand down. `disposeStorageCache` no-ops for an unwrapped store
-   * (e.g. `MemoryRawStorage`), so no instanceof test is needed.
-   */
-  private async disposeStrandStorage(strandId: string): Promise<void> {
-    const storage = this.strandStorages.get(strandId);
-    if (!storage) {
-      return;
-    }
-    this.strandStorages.delete(strandId);
-    try {
-      await disposeStorageCache(storage);
-    } catch (error) {
-      log('Failed to dispose storage cache for strand %s: %o', strandId, error);
-    }
-  }
-
-  /**
-   * Refresh a strand's revoked-peer deny set NOW — and, with it, run the
-   * teardown sweep that hangs up every connected revoked peer.
-   *
-   * The gate is otherwise poll-driven (default
-   * `DEFAULT_REVOCATION_POLL_INTERVAL_MS`), because neither a replicated
-   * revocation nor a local `revokeMember`/`leaveStrand` raises anything this
-   * runtime can hook. A caller that just wrote a revocation against the
-   * strand's database should follow it with this call, which makes the cut
-   * immediate instead of waiting out the interval.
-   *
-   * Quiet no-op — matching `quiesceStrand`'s posture — when the strand is not
-   * tracked, is quiesced, is open, or has the gate disabled: in every one of
-   * those cases there is no enforcer to refresh. Never rejects (the enforcer's
-   * refresh contains its own failures), and resolves only once the sweep has
-   * finished.
-   */
-  async refreshRevocationEnforcement(strandId: string): Promise<void> {
-    const enforcer = this.revocationEnforcers.get(strandId);
-    if (!enforcer) {
-      log('refreshRevocationEnforcement: strand %s has no armed revocation enforcer', strandId);
-      return;
-    }
-    await enforcer.refresh();
-  }
-
-  /**
-   * A fresh membership invitation was staged for `strandId` (`CadreNode`'s
-   * `adoptFormationMembershipInvite`): re-arm the strand's membership reconciler so the
-   * invitation is attempted now rather than never — the loop finished during the first
-   * join and would otherwise stay stopped until a relaunch. Quiet no-op when no reconciler
-   * is armed: a first formation stages before the strand is added, and bring-up then arms
-   * a loop that finds the invitation by itself. Returns at once; the pass runs on the
-   * loop's own serialized chain and never rejects.
-   */
-  notifyMembershipInviteStaged(strandId: string): void {
-    const reconciler = this.membershipReconcilers.get(strandId);
-    if (!reconciler) {
-      log('notifyMembershipInviteStaged: strand %s has no armed membership reconciler — bring-up will find the invitation', strandId);
-      return;
-    }
-    void reconciler.rearm();
-  }
-
-  /**
-   * Best-effort removal of THIS machine's own `Strand.MemberPeer` binding — the
-   * self arm of `removeMemberPeer`, signed with the retained launch config's party
-   * key. Called by `CadreNode.unpublishStrand` BEFORE the local stop, while the
-   * strand database and transport are still live: after the unpublish commits, the
-   * party's `StrandPartyKey` control row is gone, so no restart could ever sign
-   * this removal again — the in-memory retained key is the last chance.
-   *
-   * Never throws. Quiet no-op for a strand that is untracked, open, or launched
-   * without a party key; a quiesced (hibernating) strand, or a strand whose write
-   * quorum is already unreachable, leaves the stale binding behind with a log —
-   * it grants nothing today (admission is not allowlist-gated) and only
-   * mis-credits diversity. The strand's own membership reconciler is stopped —
-   * and its in-flight pass awaited — FIRST, so a pass that was already past its
-   * own stopped check cannot re-register the binding this is about to clear.
-   *
-   * NOTE: deliberately NOT handled here (or anywhere yet): clearing the bindings
-   * of a machine removed from the CADRE at the control layer, or of a party's
-   * OTHER machines when the party unpublishes — the party still holds the key on
-   * every surviving machine and can clear from any of them, and the stale rows
-   * grant nothing while admission is deny-list only. Revisit when
-   * `feat-strand-member-allowlist-admission` lands and a stale binding starts
-   * granting admission rather than merely mis-crediting diversity.
-   */
-  async clearOwnMemberPeerBinding(strandId: string): Promise<void> {
-    const instance = this.instances.get(strandId);
-    const config = this.launchConfigs.get(strandId);
-    if (!instance || !config || config.strandRow.Type !== 'c' || !config.partyMemberPrivateKey) {
-      return;
-    }
-    const reconciler = this.membershipReconcilers.get(strandId);
-    if (reconciler) {
-      reconciler.stop();
-      await reconciler.settle();
-    }
-    const db = instance.database?.getDatabase();
-    const peerId = instance.libp2pNode?.peerId.toString();
-    if (!db || !peerId) {
-      log('clearOwnMemberPeerBinding: strand %s has no live runtime — own MemberPeer binding left in place', strandId);
-      return;
-    }
-    try {
-      // Background write on the app's database: never join a transaction the app has open.
-      await removeMemberPeer(db, {
-        memberKeyPair: strandMemberKeyPair(config.partyMemberPrivateKey),
-        peerId
-      }, { joinOpenTransaction: false });
-      log('clearOwnMemberPeerBinding: strand %s — own binding for peer %s cleared', strandId, peerId);
-    } catch (error) {
-      log('clearOwnMemberPeerBinding: strand %s — best-effort clear failed (strand may already be unreachable): %o',
-        strandId, error);
-    }
-  }
-
-  /**
-   * Quiesce a strand: release its strand-network resources (stop the libp2p node,
-   * close the StrandDatabase) while RETAINING the instance record — identity,
-   * sAppInfo, keys, latency hint, metadata — and its launch config so it can be
-   * resumed later, then mark it `hibernating`. Mechanically this is `stopStrand` minus
-   * the instance/config deletion. A tracked strand with no runtime left (a failed
-   * rebuild already released it) is only marked; an untracked one is left alone.
-   *
-   * Ordered with {@link resumeStrand} (see {@link transitions}): it starts once the resume
-   * or launch build before it has settled, so it releases a finished runtime, and the
-   * status write is its last step, so a resume requested meanwhile writes after it. A call
-   * made while another quiesce is the latest request joins it. A failed release leaves the
-   * status as it was and rejects.
-   */
-  async quiesceStrand(strandId: string): Promise<void> {
-    const previous = this.transitions.get(strandId);
-    if (previous?.kind === 'quiesce') {
-      log('quiesceStrand: strand %s — joining the quiesce already requested', strandId);
-      return previous.operation;
-    }
-    const operation = this.runQuiesce(strandId, previous);
-    this.holdTransition(strandId, { kind: 'quiesce', operation });
-    return operation;
-  }
-
-  /** Body of {@link quiesceStrand}. */
-  private async runQuiesce(strandId: string, previous: RuntimeTransition | undefined): Promise<void> {
-    await this.settleTransition(strandId, previous);
-    await this.settleRuntimeBuilds(strandId);
-    const instance = this.instances.get(strandId);
-    if (!instance) {
-      log('quiesceStrand: strand %s not found', strandId);
-      return;
-    }
-    if (instance.libp2pNode || instance.database) {
-      log('Quiescing strand instance: %s', strandId);
-      await this.releaseRuntime(instance);
-      log('Strand %s quiesced (resources released, instance retained)', strandId);
-    } else {
-      log('quiesceStrand: strand %s already quiesced', strandId);
-    }
-    instance.status = 'hibernating';
-  }
-
-  /**
-   * Whether the latest runtime transition requested for `strandId` is a quiesce that has
-   * not finished. Its handles may still be attached and its status still live, but it is on
-   * its way to `hibernating`, so a caller that needs it up must {@link resumeStrand} it — the
-   * resume runs after the quiesce — rather than treat it as live.
-   */
-  isQuiescing(strandId: string): boolean {
-    return this.transitions.get(strandId)?.kind === 'quiesce';
-  }
-
-  /**
-   * Resume a previously-quiesced strand: rebuild its libp2p node + StrandDatabase
-   * from the retained launch config and re-attach them, transitioning it back to
-   * `active`. `overrides` re-applies volatile inputs that may have changed since
-   * launch (the cohort `bootstrapNodes` seed and the strand's `servingMachines`
-   * count) and updates the retained config so a later resume reuses the latest
-   * values. Returns the live instance unchanged if it is already running.
-   *
-   * Overlapping calls share one rebuild: a call made while a resume is the latest request
-   * for the strand — a wake landing during a check-in — joins it and settles exactly as it
-   * does, `'error'` status included. The joiner's own `overrides` are ignored; the first
-   * resume's seed wins, and both callers resolved the cohort seed moments apart. A call made
-   * while a quiesce is the latest request runs after that quiesce and rebuilds (see
-   * {@link transitions}), and one made while `startStrand` is still building waits for that
-   * build rather than starting a second.
-   */
-  async resumeStrand(strandId: string, overrides?: ResumeStrandOverrides): Promise<StrandInstance> {
-    if (this.stopping) {
-      throw new Error('StrandInstanceManager is stopping');
-    }
-    // No `await` between this read and `holdTransition`, or a second caller slips through
-    // and builds.
-    const previous = this.transitions.get(strandId);
-    if (previous?.kind === 'resume') {
-      log('resumeStrand: strand %s — joining the resume already in flight', strandId);
-      return previous.operation;
-    }
-    const operation = this.runResume(strandId, previous, overrides);
-    this.holdTransition(strandId, { kind: 'resume', operation });
-    return operation;
-  }
-
-  /** Record `transition` as the latest for `strandId` until it settles. */
-  private holdTransition(strandId: string, transition: RuntimeTransition): void {
-    this.transitions.set(strandId, transition);
-    const release = (): void => {
-      if (this.transitions.get(strandId) === transition) {
-        this.transitions.delete(strandId);
-      }
-    };
-    // The rejection itself reaches the caller through the operation returned to it.
-    void transition.operation.then(release, release);
-  }
-
-  /** Wait for the transition requested before this one. Never rejects: its failure belongs to its caller. */
-  private async settleTransition(strandId: string, previous: RuntimeTransition | undefined): Promise<void> {
-    if (!previous) {
-      return;
-    }
-    await (previous.operation as Promise<unknown>).catch((error: unknown) => {
-      log('Strand %s: the %s queued ahead failed (reported by its caller): %o', strandId, previous.kind, error);
-    });
-  }
-
-  /** Body of {@link resumeStrand}; at most one runs per strand at a time. */
-  private async runResume(
-    strandId: string,
-    previous: RuntimeTransition | undefined,
-    overrides?: ResumeStrandOverrides
-  ): Promise<StrandInstance> {
-    // Before any "already live" test: a quiesce ahead still has the handles attached, and
-    // mid-build (a launch still building) the libp2p node is attached while the database is
-    // not. A failed launch has already dropped the instance, and the check below reports it
-    // untracked.
-    await this.settleTransition(strandId, previous);
-    await this.settleRuntimeBuilds(strandId);
-    const instance = this.instances.get(strandId);
-    if (!instance) {
-      throw new Error(`Cannot resume strand ${strandId}: not tracked`);
-    }
-    const launchConfig = this.launchConfigs.get(strandId);
-    if (!launchConfig) {
-      throw new Error(`Cannot resume strand ${strandId}: no retained launch config`);
-    }
-    if (instance.libp2pNode || instance.database) {
-      log('resumeStrand: strand %s already live', strandId);
-      return instance;
-    }
-
-    log('Resuming strand instance: %s', strandId);
-    const tTotal = performance.now();
-
-    // Re-apply volatile inputs and persist them so a subsequent resume reuses them.
-    // Each `??` matters: a resume that passes no override must keep the retained value,
-    // and a resume that passes one must leave it retained for the next resume — otherwise
-    // a later no-override wake silently reverts to whatever launch time saw.
-    const resumeConfig: StartStrandConfig = {
-      ...launchConfig,
-      bootstrapNodes: overrides?.bootstrapNodes ?? launchConfig.bootstrapNodes,
-      // NOTE: `??` retains, so an override can raise or lower the count but cannot CLEAR it
-      // back to "this node no longer knows" — the direction that would declare nothing. Moot
-      // while nothing feeds `servingMachines` at all; if a source lands
-      // (`backlog/feat-strand-yardstick-from-serving-machines`) that can legitimately lose the
-      // count — a strand whose member rows became unreadable — this merge must gain an explicit
-      // clear rather than silently declaring a stale number over a serving set it can no longer
-      // see. Same applies to `bootstrapNodes` above, where a stale seed is harmless.
-      servingMachines: overrides?.servingMachines ?? launchConfig.servingMachines
-    };
-    this.launchConfigs.set(strandId, resumeConfig);
-
-    instance.status = 'starting';
-    try {
-      await this.trackRuntimeBuild(strandId, this.buildStrandRuntime(instance, resumeConfig));
-      timing('[resumeStrand:%s] total: %dms', strandId, Math.round(performance.now() - tTotal));
-      log('Strand %s resumed successfully', strandId);
-      return instance;
-    } catch (error) {
-      instance.status = 'error';
-      instance.error = error instanceof Error ? error.message : String(error);
-      log('Failed to resume strand %s: %s', strandId, instance.error);
-      throw error;
-    }
-  }
-
-  /**
-   * Honor a founder request against an ALREADY-TRACKED strand — the seam that
-   * closes the "whoever launches first decides whether the bootstrap runs" gap:
-   * an instance first launched as a joiner (an app's own attach, or a watcher
-   * poll winning the launch race) used to swallow a later founder request
-   * silently, leaving the strand active with no `Strand.Header`.
-   *
-   * Flips the RETAINED launch config's `founder` to true, so every later
-   * quiesce → resume rebuild founds as well (the bootstrap is insert-if-absent —
-   * {@link StrandDatabase.ensureFounderBootstrap} — so re-running it per rebuild
-   * writes nothing twice), and runs the bootstrap against the live database now.
-   *
-   * @returns how the request resolved:
-   * - `'already-founder'` — the retained config already founds; nothing to do.
-   * - `'bootstrapped'` — config flipped and the live database ran the bootstrap.
-   * - `'needs-resume'` — config flipped, but the instance is quiesced (no live
-   *   database) or being quiesced (its database is closing), so the bootstrap could
-   *   not run here: the CALLER must wake the
-   *   strand (`CadreNode.wakeStrand`, which owns the hibernation bookkeeping this
-   *   manager does not) so the rebuild — which now founds — runs it.
-   * @param resolvePartyKey - Asked for the party's own membership key when (and only
-   *   when) a CLOSED strand's retained config carries none — the instance was launched
-   *   as a joiner before its `StrandPartyKey` row existed or replicated. The resolved
-   *   key is retained alongside the founder flip so the bootstrap (now, or on the
-   *   caller's wake for `'needs-resume'`) can seat the founding Member/Manager.
-   * @throws when the strand is not tracked — this seam exists only for the
-   *   tracked-instance launch path; an untracked id is a caller bug.
-   * @throws whatever the live bootstrap throws (e.g. `PreSplitStrandIdentityError`),
-   *   after {@link withdrawFounderRequest} — so a retry re-attempts the founding rather
-   *   than resolving `'already-founder'` over an instance that never founded. A caller
-   *   whose `'needs-resume'` wake fails owes the same withdrawal.
-   */
-  async foundExistingStrand(
-    strandId: string,
-    resolvePartyKey?: () => Promise<string | undefined>
-  ): Promise<'already-founder' | 'bootstrapped' | 'needs-resume'> {
-    const instance = this.instances.get(strandId);
-    const config = this.launchConfigs.get(strandId);
-    if (!instance || !config) {
-      throw new Error(`Cannot found strand ${strandId}: not tracked`);
-    }
-    if (config.founder === true) {
-      return 'already-founder';
-    }
-    const partyMemberPrivateKey =
-      config.partyMemberPrivateKey
-        ?? (config.strandRow.Type === 'c' ? await resolvePartyKey?.() : undefined);
-    // A fresh object rather than mutating in place: startStrand retains the CALLER'S
-    // config object, which is not ours to rewrite.
-    this.launchConfigs.set(strandId, { ...config, founder: true, partyMemberPrivateKey });
-    if (this.isQuiescing(strandId) || (!instance.database && !this.firstSyncGates.has(strandId))) {
-      return 'needs-resume';
-    }
-    try {
-      await this.ensureFounderBootstrap(strandId);
-    } catch (error) {
-      this.withdrawFounderRequest(strandId);
-      throw error;
-    }
-    return 'bootstrapped';
-  }
-
-  /**
-   * Give a tracked storage replica ({@link StartStrandConfig.sAppConfig} absent) the app's
-   * schema in place — the seam for an app on this machine claiming a strand this node was
-   * already hosting. No runtime rebuild: the libp2p node, its peer id and connections, the
-   * store and its warm cache all stay, and the new `App` tables read the blocks the replica
-   * already holds. Callers attach BEFORE {@link foundExistingStrand}: the founder bootstrap
-   * writes the sApp into `Strand.Header` and refuses a database that has none.
-   *
-   * A runtime build in flight (the replica's launch, or a hibernation wake) is waited out
-   * first, so the attach acts on the database that build produced. Then, in order: the
-   * schema signature is checked before anything changes; the retained launch config takes
-   * the sApp, so every later rebuild applies it through `composeStrand`; the live database —
-   * published, or still held by the first-sync gate — gets the schema; and only then does
-   * the instance record `sAppInfo` and the sApp's latency hint. A quiesced instance gets the
-   * config alone, and its next resume applies the schema.
-   *
-   * @returns `'already-attached'` when the instance already runs an sApp — a DIFFERENT one
-   *   is logged and left as it is, as a second claim of a claimed strand always has been —
-   *   else `'attached'`.
-   * @throws when the strand is not tracked (a caller bug, or a launch that failed while
-   *   this waited on it); when the schema signature is refused, having changed nothing; and
-   *   whatever the live apply throws (e.g. a quiesce closed the database mid-apply) — the
-   *   retained config then still carries the sApp and `sAppInfo` stays unset, so the next
-   *   claim retries the apply.
-   *
-   * NOTE: two claims of one replica racing each other both pass the `sAppInfo` check and
-   * apply concurrently; the loser of the declarative diff can reject, and its `addStrand`
-   * retry then resolves `'already-attached'`. Needs an app calling `addStrand` twice at once
-   * (or beside a watcher retry); if it is ever seen, chain attaches per strand id.
-   */
-  async attachSApp(
-    strandId: string,
-    sAppConfig: SAppConfig,
-    options: { requireSignedSchemas?: boolean } = {}
-  ): Promise<'attached' | 'already-attached'> {
-    await this.settleRuntimeBuilds(strandId);
-    const instance = this.instances.get(strandId);
-    const config = this.launchConfigs.get(strandId);
-    if (!instance || !config) {
-      throw new Error(`Cannot attach an sApp to strand ${strandId}: not tracked`);
-    }
-    if (instance.sAppInfo) {
-      if (instance.sAppInfo.id !== sAppConfig.id) {
-        log('attachSApp: strand %s already runs sApp %s — the claim for sApp %s is ignored',
-          strandId, instance.sAppInfo.id, sAppConfig.id);
-      }
-      return 'already-attached';
-    }
-    const sAppInfo = verifiedSAppInfo(strandId, sAppConfig, options.requireSignedSchemas);
-    // A fresh object rather than mutating in place: startStrand retains the CALLER'S
-    // config object, which is not ours to rewrite.
-    this.launchConfigs.set(strandId, { ...config, sAppConfig });
-    const database = instance.database ?? this.firstSyncGates.get(strandId)?.database;
-    if (database) {
-      await database.attachAppSchema(sAppConfig);
-    } else {
-      log('attachSApp: strand %s is quiesced — its next resume applies the sApp schema', strandId);
-    }
-    instance.sAppInfo = sAppInfo;
-    if (sAppConfig.latencyHint) {
-      // HibernationManager reads the hint whenever it arms a timer, so the app's hint governs
-      // from the next one: a timer already armed runs once at the old duration, and a replica
-      // launched under a realtime default was never tracked, so it stays up until relaunched.
-      instance.latencyHint = sAppConfig.latencyHint;
-    }
-    log('Strand %s upgraded from storage replica to %s', strandId, describeSApp(sAppConfig));
-    return 'attached';
-  }
-
-  /**
-   * Whether `strandId` is launched but still waiting for its first sync — a joiner whose
-   * runtime is up while its database is withheld (status `'syncing'`, or `'idle'` after
-   * the hibernation manager's idle timer fired on it), or one waiting to rebuild because no
-   * seed peer had answered ({@link deferUnreachableJoin}). `false` for a writable, quiesced,
-   * or untracked strand. The predicate {@link whenWritable}'s callers gate on.
-   */
-  isAwaitingFirstSync(strandId: string): boolean {
-    const instance = this.instances.get(strandId);
-    return instance !== undefined
-      && (isAwaitingFirstSync(instance) || this.unreachableJoinRetries.has(strandId));
-  }
-
-  /**
-   * Resolve once the strand's database is published to the app — immediately for a
-   * writable strand — or reject with {@link StrandAwaitingFirstSyncError} after
-   * `timeoutMs` (default: the retained launch config's `firstSync.timeoutMs`, else
-   * {@link DEFAULT_STRAND_FIRST_SYNC_TIMEOUT_MS}). The rejection is RETRYABLE: nothing is
-   * torn down, the gate keeps probing, and a later call waits afresh. A quiesced
-   * (hibernating) strand is not woken here — the wait spans a resume, so a caller that
-   * wants one wakes it (`CadreNode.wakeStrand`) and waits. Rejects immediately for an
-   * untracked strand, and whenever the strand is stopped while a wait is pending.
-   */
-  whenWritable(strandId: string, options?: { timeoutMs?: number }): Promise<StrandInstance> {
-    const instance = this.instances.get(strandId);
-    if (!instance) {
-      return Promise.reject(new Error(`Cannot wait for strand ${strandId} to become writable: not tracked`));
-    }
-    if (instance.database) {
-      return Promise.resolve(instance);
-    }
-    const timeoutMs = options?.timeoutMs
-      ?? this.launchConfigs.get(strandId)?.firstSync?.timeoutMs
-      ?? DEFAULT_STRAND_FIRST_SYNC_TIMEOUT_MS;
-    const startedAt = Date.now();
-    return new Promise<StrandInstance>((resolve, reject) => {
-      const waiters = this.writableWaiters.get(strandId) ?? new Set<WritableWaiter>();
-      this.writableWaiters.set(strandId, waiters);
-      const settle = (): void => {
-        clearTimeout(timer);
-        waiters.delete(waiter);
-        if (waiters.size === 0) {
-          this.writableWaiters.delete(strandId);
-        }
-      };
-      const waiter: WritableWaiter = {
-        resolve: () => { settle(); resolve(instance); },
-        reject: (error) => { settle(); reject(error); }
-      };
-      const timer = setTimeout(
-        () => waiter.reject(new StrandAwaitingFirstSyncError(strandId, Date.now() - startedAt)),
-        timeoutMs);
-      (timer as { unref?: () => void }).unref?.();
-      waiters.add(waiter);
-    });
-  }
-
-  /**
-   * The tracked instance once no runtime build of it is in flight, or `undefined` when none
-   * is tracked by then — never launched, or the launch waited on failed and dropped its
-   * record. For a caller about to hand a tracked instance to an app: until its build settles
-   * the instance is `'starting'` with no database, and a failed launch leaves no instance.
-   * Never rejects; a failed build is reported by the call that started it.
-   */
-  async whenRuntimeBuilt(strandId: string): Promise<StrandInstance | undefined> {
-    await this.settleRuntimeBuilds(strandId);
-    return this.instances.get(strandId);
-  }
-
-  /**
-   * Hand a launch's database to the app: set `instance.database`, retire the gate that
-   * held it, release every {@link whenWritable} waiter, and — when the launch had already
-   * been reported gated — flip it `'active'` and announce it through the retained config's
-   * `onWritable`. "Reported gated" is any status but `'starting'`: `'syncing'`, or the
-   * `'idle'` the hibernation manager's idle timer flips a long-gated joiner to (see
-   * {@link isAwaitingFirstSync}) — the Header arriving after that flip must announce too,
-   * or an app hanging on `strand:writable` never learns the strand opened. A publish that
-   * lands while the instance is still `'starting'` (the Header arrived during the rest of
-   * bring-up) announces nothing: `startStrand` / `resumeStrand` are about to report the
-   * strand `'active'`, and `strand:writable` is defined as the follow-up to a gated
-   * launch, never a duplicate of `strand:started`.
-   *
-   * It is also the one seam where a WITHHELD database becomes available, so it kicks the
-   * membership reconciler: on a gated launch that loop's own immediate pass ran while
-   * `instance.database` was still unset, found no database, and would otherwise not try
-   * again until its retry timer fired. A launch that was never gated publishes here BEFORE
-   * the reconciler is constructed, so there is nothing registered to kick and the loop's own
-   * immediate pass already sees the database — which is why the kick needs no `wasGated`
-   * test of its own.
-   */
-  private publishDatabase(instance: StrandInstance, database: StrandDatabase): void {
-    const strandId = instance.strandId;
-    const gate = this.firstSyncGates.get(strandId);
-    if (gate) {
-      gate.open();
-      this.firstSyncGates.delete(strandId);
-    }
-    instance.database = database;
-    const wasGated = instance.status !== 'starting';
-    if (wasGated) {
-      // A Header delivered by a peer is strand activity; a publish during bring-up is not
-      // (see the build's end in `buildStrandRuntime`).
-      instance.lastActivity = new Date();
-      instance.status = 'active';
-      log('Strand %s is now writable: Strand.Header held', strandId);
-    }
-    const waiters = this.writableWaiters.get(strandId);
-    if (waiters) {
-      this.writableWaiters.delete(strandId);
-      waiters.forEach((waiter) => waiter.resolve());
-    }
-    if (wasGated) {
-      this.launchConfigs.get(strandId)?.onWritable?.(strandId);
-    }
-    // Not awaited: publishing must stay synchronous, and the pass chains on the loop's own
-    // tail, so it cannot overlap a pass already in flight.
-    void this.membershipReconcilers.get(strandId)?.reconcile();
-  }
-
-  /** Reject every {@link whenWritable} waiter for a strand that is going away. */
-  private rejectWritableWaiters(strandId: string, reason: string): void {
-    const waiters = this.writableWaiters.get(strandId);
-    if (!waiters) {
-      return;
-    }
-    this.writableWaiters.delete(strandId);
-    waiters.forEach((waiter) => waiter.reject(new Error(`Strand ${strandId} ${reason} before becoming writable`)));
-  }
-
-  /**
-   * Undo {@link foundExistingStrand}'s founder flip after the founding it promised
-   * failed, so the retained config again says what this instance actually runs as (a
-   * joiner) and the next founder request re-attempts the bootstrap. The resolved party
-   * key stays retained — it is this party's identity for the strand either way. No-op
-   * when the strand is untracked or its config does not found.
-   */
-  withdrawFounderRequest(strandId: string): void {
-    const config = this.launchConfigs.get(strandId);
-    if (config?.founder === true) {
-      this.launchConfigs.set(strandId, { ...config, founder: false });
-    }
-  }
-
-  /**
-   * Run the (idempotent) founder bootstrap against a tracked strand's LIVE
-   * database, independently of what the retained launch config says.
-   *
-   * Separate from {@link foundExistingStrand} because a caller that resolved
-   * `'needs-resume'` and woke the strand must not assume the wake's own rebuild
-   * founded it: `HibernationManager` COALESCES wakes, so a wake already in flight
-   * when the config flipped had already read the PRE-flip config and rebuilt as a
-   * joiner. Re-running the bootstrap costs one insert-if-absent probe per table
-   * and is the only thing that makes "founding resolves once the Header is
-   * written" true on that path.
-   *
-   * A strand still `'syncing'` (its database withheld behind the first-sync gate) is
-   * founded through the gate: the bootstrap writes the Header this machine was waiting
-   * to receive, so a successful run publishes the database and the strand goes
-   * `'active'` — the one legitimate way a joiner launch becomes writable without a peer.
-   *
-   * @throws when the strand is not tracked, or is still quiesced (no live
-   *   database) — both mean the bootstrap did NOT run, which a founder request
-   *   must never swallow.
-   */
-  async ensureFounderBootstrap(strandId: string): Promise<void> {
-    const instance = this.instances.get(strandId);
-    if (!instance) {
-      throw new Error(`Cannot run the founder bootstrap for strand ${strandId}: not tracked`);
-    }
-    const gate = this.firstSyncGates.get(strandId);
-    const database = instance.database ?? gate?.database;
-    if (!database) {
-      throw new Error(
-        `Cannot run the founder bootstrap for strand ${strandId}: it is quiesced, so there ` +
-        'is no live database to write to — resume it first.'
-      );
-    }
-    // Forward the RETAINED config's party key: a foundExistingStrand that flipped a
-    // joiner launch to founder may have resolved a key the live database's captured
-    // config (built at the original launch) never saw.
-    await database.ensureFounderBootstrap(this.launchConfigs.get(strandId)?.partyMemberPrivateKey);
-    if (!instance.database) {
-      this.publishDatabase(instance, database);
-    }
-  }
-
-  /**
-   * Stop a strand instance
-   */
-  async stopStrand(strandId: string): Promise<void> {
-    const instance = this.instances.get(strandId);
-    if (!instance) {
-      log('Strand %s not found', strandId);
-      return;
-    }
-
-    log('Stopping strand instance: %s', strandId);
-    instance.status = 'stopping';
-    this.rejectWritableWaiters(strandId, 'was stopped');
-
-    try {
-      await this.releaseRuntime(instance);
-      instance.status = 'stopped';
-      this.instances.delete(strandId);
-      this.launchConfigs.delete(strandId);
-      // Storage is released only here — NOT in releaseRuntime, which a quiesce shares.
-      await this.disposeStrandStorage(strandId);
-      log('Strand %s stopped successfully', strandId);
-    } catch (error) {
-      instance.status = 'error';
-      instance.error = error instanceof Error ? error.message : String(error);
-      log('Error stopping strand %s: %s', strandId, instance.error);
-      throw error;
-    }
-  }
-
-  /**
-   * Stop all strand instances
-   */
-  async stopAll(): Promise<void> {
-    this.stopping = true;
-    log('Stopping all strand instances (%d)', this.instances.size);
-    
-    const stopPromises = Array.from(this.instances.keys()).map(id => 
-      this.stopStrand(id).catch(err => {
-        log('Error stopping strand %s during shutdown: %s', id, err);
-      })
-    );
-    
-    await Promise.all(stopPromises);
-    this.stopping = false;
-    log('All strand instances stopped');
-  }
-}
-
+import debug from 'debug';
+import type { Libp2p, PrivateKey } from '@libp2p/interface';
+import { createLibp2pNode, type IRawStorage } from '@optimystic/db-p2p';
+import { wrapStorageWithCache, disposeStorageCache } from '@serfab/quereus-plugin-sereus';
+import { StrandDatabase } from './strand-database.js';
+import type { StrandCommitDurability, StrandCommitHeld } from './strand-commit-durability.js';
+import { PeerJoinBackfill, type PeerJoinBackfillConfig } from './peer-join-backfill.js';
+import { strandNetworkStatePersistence, type StrandNetworkStateStore } from './strand-network-state.js';
+import {
+  StrandRevocationEnforcer,
+  createRevocationConnectionGater,
+  readStrandRevocationRows,
+  type StrandRevocationEnforcementConfig
+} from './strand-revocation-enforcer.js';
+import {
+  StrandMembershipReconciler,
+  type PendingMembershipInviteSource,
+  type StrandMembershipReconciliationConfig
+} from './strand-membership-reconciler.js';
+import { removeMemberPeer } from './strand-membership-writer.js';
+import { strandMemberKeyPair } from './strand-member-key.js';
+import { assertSchemaSignature } from './schema-verification.js';
+import { assertStrandScopeKey } from './storage-scope.js';
+import {
+  StrandFirstSyncGate,
+  StrandAwaitingFirstSyncError,
+  strandFirstSyncComplete,
+  DEFAULT_STRAND_FIRST_SYNC_TIMEOUT_MS,
+  type StrandFirstSyncConfig
+} from './strand-first-sync-gate.js';
+import type {
+  StrandInstance,
+  StrandRow,
+  StorageConfig,
+  NetworkConfig,
+  LatencyHint,
+  NodeProfile,
+  SAppConfig,
+  SAppInfo,
+  RawStorageProvider,
+  Libp2pNodeWithRepo
+} from './types.js';
+import { DEFAULT_CONNECTION_MONITOR, resolveStrandClusterSize, strandClusterPolicy } from './types.js';
+import { strandNodeAddrs } from './strand-network-config.js';
+import { nodeTransportsOption } from './node-transports.js';
+import { resolveRelayServer } from './relay-server.js';
+import { strandCohortTopicOption, type StrandReactivityConfig } from './strand-reactivity.js';
+import { superviseRelayReservation, type RelayReservationSupervisor } from './relay-reservation.js';
+import { strandRedialTimings, superviseStrandPeerRedial, type StrandPeerRedial } from './strand-peer-redial.js';
+import { groupAddrsByPeerId } from './peer-addr-book.js';
+import { isCohortUnreachableRead } from './control-read-retry.js';
+import { declaredCohortReadDeadlineMs, optimysticDialLimits, peerJoinPushBudget, relayReservationBudgetMs, resolveLinkRoundTripMs } from './link-budget.js';
+
+const log = debug('sereus:cadre:strand-manager');
+
+/**
+ * How often a joining launch that could not build its database, because no seed peer had
+ * answered, retries the build on its running libp2p node
+ * (`StrandInstanceManager.deferUnreachableJoin`). Each attempt composes a fresh database and
+ * re-reads the schema catalog over the network, so it is spaced well above the first-sync
+ * gate's probe cadence.
+ *
+ * NOTE: the 5 s dates from when each attempt also rebuilt the libp2p node. A joiner whose seed
+ * peer answered just after a failed attempt waits up to this long to become writable; if that
+ * wait shows up in a measured join, try a shorter first retry.
+ */
+export const UNREACHABLE_JOIN_RETRY_MS = 5_000;
+
+/**
+ * How one database build of a runtime ended ({@link StrandInstanceManager.buildStrandDatabase}):
+ * published or held by the first-sync gate; failed only because no seed peer has answered yet;
+ * or overtaken by a release of the runtime it was building on.
+ */
+type StrandDatabaseBuild = 'built' | 'awaiting-seed' | 'released';
+
+const timing = debug('sereus:cadre:timing');
+
+/** One caller blocked in {@link StrandInstanceManager.whenWritable}. */
+interface WritableWaiter {
+  resolve: () => void;
+  reject: (error: Error) => void;
+}
+
+/**
+ * Whether a tracked instance is launched but waiting for its first sync: the runtime is
+ * up (`libp2pNode` set) while the database is still withheld (`database` unset). That
+ * is the first-sync gate's state regardless of what `status` says — the hibernation
+ * manager's idle timer can flip a long-gated joiner to `'idle'` without changing it.
+ *
+ * NOTE: `status` is therefore not the source of truth for "gated": a joiner that reaches
+ * nobody for `idleTimeout` (5 min at the interactive hint) reads `'idle'`, then
+ * `'hibernating'`, while still holding no database. Harmless today — every decision the
+ * runtime makes goes through this predicate, {@link StrandInstanceManager.publishDatabase}
+ * announces on the idle-flipped instance exactly as on a `'syncing'` one, and an app that
+ * shows "waiting for the other member" on `'syncing'` merely sees "idle" instead after
+ * five minutes. If an app ever needs the gated LABEL to survive the idle flip, make the
+ * hibernation manager's idle transition preserve `'syncing'` rather than adding a flag.
+ */
+export function isAwaitingFirstSync(instance: StrandInstance): boolean {
+  return instance.libp2pNode !== undefined && instance.database === undefined;
+}
+
+/**
+ * The status a LIVE instance (runtime rebuilt or woken) should report: `'active'` once its
+ * database is published, `'syncing'` while the first-sync gate still withholds it. Every
+ * site that would otherwise write `'active'` after a build or a wake goes through this,
+ * so a gated joiner never reads as writable.
+ */
+export function liveStrandStatus(instance: StrandInstance): 'active' | 'syncing' {
+  return instance.database ? 'active' : 'syncing';
+}
+
+/**
+ * Configuration for starting a strand instance
+ */
+export interface StartStrandConfig {
+  strandRow: StrandRow;
+  /**
+   * sApp configuration provided by the hosting application. Absent for a **storage
+   * replica** — a strand this node stores and serves without the app installed
+   * (`CadreNodeConfig.hostUnclaimedStrands`): no schema signature check, no `App`
+   * schema, no `sAppInfo` on the instance, and the default latency hint. A replica is
+   * always a joiner; `StrandDatabase` refuses to found one. An app that later claims the
+   * strand upgrades the running replica in place ({@link StrandInstanceManager.attachSApp}).
+   */
+  sAppConfig?: SAppConfig;
+  storage?: StorageConfig;
+  network?: NetworkConfig;
+  profile: NodeProfile;
+  defaultLatencyHint: LatencyHint;
+  privateKey?: PrivateKey;
+  /** Cohort-derived discovery seed (multiaddr strings). Defaults to [] when omitted. */
+  bootstrapNodes?: string[];
+  /**
+   * Require a valid author signature on the sApp schema before bring-up.
+   * Defaults to true (fail closed) when omitted; set false only for dev/test
+   * with unsigned demo schemas. Mirrors {@link CadreNodeConfig.requireSignedSchemas}.
+   */
+  requireSignedSchemas?: boolean;
+  /**
+   * Whether this node founds the strand (vs. joins it). Forwarded to the
+   * StrandDatabase so the founder bootstrap (Header, founding Member/Manager)
+   * runs once at bring-up. Joiners leave this unset and write nothing. Callers
+   * resolve it BEFORE calling `startStrand` (`CadreNode.launchStrand` derives it
+   * from the row's `FounderOwnerKey` when no explicit flag is given — see
+   * {@link StrandConfig.founder}); a founder request that arrives while the
+   * instance is already tracked goes through {@link StrandInstanceManager.foundExistingStrand},
+   * never through a repeat `startStrand`.
+   */
+  founder?: boolean;
+  /**
+   * THIS party's own strand membership private key (base64 protobuf) for a closed
+   * strand — the identity the founder bootstrap derives `Member.Key` /
+   * `Manager.MemberKey` from. Resolved by the caller (`CadreNode.launchStrand`: an
+   * explicit attach-time key, else the control-layer `StrandPartyKey` row, healed by
+   * mint-on-launch for the founding machine) and retained with the launch config, so a
+   * hibernation wake rebuilds under the same identity. Deliberately NOT the strand
+   * row's shared `MemberPrivateKey` (see `StrandDatabaseConfig.partyMemberPrivateKey`).
+   * Absent for open strands, and for joiners with no persisted party key.
+   */
+  partyMemberPrivateKey?: string;
+  /**
+   * Number of nodes Optimystic is told this strand's replication cluster should
+   * have. Same rule as {@link CadreNodeConfig.strandClusterSize}, which CadreNode
+   * forwards here: every node on the strand should use the same value, and it is
+   * frozen when the strand's libp2p node is created. Defaults to
+   * `DEFAULT_STRAND_CLUSTER_SIZE` (4); values below `MIN_CLUSTER_SIZE` (2) are
+   * rejected by `resolveStrandClusterSize`. Strand-only — the
+   * control network's breadth is the fixed `CONTROL_REPLICATION_BREADTH`.
+   */
+  clusterSize?: number;
+  /**
+   * Machines that SERVE THIS STRAND (this node included), for the strand node's
+   * block-repair corroboration yardstick — see `resolveRepairYardstick`. It must be
+   * an authenticated per-strand count: the machines that actually run a node for this
+   * strand, not the machines that exist.
+   *
+   * **The party's enrolled-machine count is NOT this number and must never be passed
+   * here.** A strand launches only on machines whose embedding app registered its sApp
+   * config (`CadreNode.addStrand`), so a closed strand shared by two machines of a
+   * three-machine party is served by two. Passing three over-declares, and
+   * over-declaring is the unsafe direction: at a declaration of 3 or more Optimystic
+   * pins the repair corroboration floor at two corroborating peers, which a cohort
+   * that can only ever field one peer can never reach — so the strand can never repair
+   * a block (`cluster-fetch:no-quorum`, surfacing as reads failing with `Missing
+   * block`). That regression is why this field was renamed off "enrolledMachines":
+   * `bug-strand-yardstick-counts-party-machines`.
+   *
+   * **Nothing feeds it in production today.** No authenticated per-strand serving count
+   * exists yet, so `CadreNode` passes nothing and the strand node runs the frozen
+   * `STRAND_CLUSTER_POLICY` — declaring no yardstick, which leaves the known,
+   * upstream-tracked single-voter exposure
+   * (`backlog/debt-read-repair-single-voter-corroboration`). Building the count is
+   * `backlog/feat-strand-yardstick-from-serving-machines`; this field and its
+   * threading are the seam it plugs into.
+   *
+   * Volatile when a source does exist: re-resolve it on every resume beside the cohort
+   * seed, since machines join and leave a strand while it hibernates. Omitting it means
+   * "this node does not know", which declares nothing.
+   */
+  servingMachines?: number;
+  /**
+   * Tuning for the strand peer-join block catch-up ({@link PeerJoinBackfill}),
+   * forwarded from {@link CadreNodeConfig.strandBackfill}. When the strand's
+   * libp2p node connects to a peer this runtime has not yet caught up, every
+   * block in the strand's own raw store is pushed to it. Runs only on strands
+   * with per-strand storage; `{ enabled: false }` restores the pre-existing
+   * no-backfill behaviour.
+   */
+  backfill?: PeerJoinBackfillConfig;
+  /**
+   * Whether this strand's libp2p node is built with Optimystic's change notifications,
+   * forwarded from {@link CadreNodeConfig.strandReactivity} and resolved per strand by
+   * `strandCohortTopicOption`. Retained with the launch config, so a hibernation wake
+   * rebuilds the node with the same setting.
+   */
+  reactivity?: StrandReactivityConfig;
+  /**
+   * Tuning for the CLOSED-strand revoked-peer gate
+   * ({@link StrandRevocationEnforcer}), forwarded from
+   * {@link CadreNodeConfig.strandRevocationEnforcement}. Armed only for
+   * `strandRow.Type === 'c'` (an open strand has no membership rows to derive a
+   * deny set from); `{ enabled: false }` restores the pre-existing behaviour
+   * (a removed party's peers keep being served).
+   */
+  revocationEnforcement?: StrandRevocationEnforcementConfig;
+
+  /**
+   * The staged formation membership invitation for THIS strand — the seam the
+   * bring-up membership reconciler (`strand-membership-reconciler.ts`) reads to
+   * redeem the party's `Strand.Member` seat, and clears once the invitation is
+   * spent, burned, or dead. `CadreNode.launchStrand` wires it over its in-memory
+   * `pendingMembershipInvites` map; read lazily per pass, so a re-formation that
+   * replaces the entry between passes is picked up. Only consulted for closed
+   * strands launched with a {@link partyMemberPrivateKey}; harmless otherwise.
+   */
+  pendingMembershipInvite?: PendingMembershipInviteSource;
+
+  /**
+   * Tuning for the CLOSED-strand membership reconciler, forwarded from
+   * {@link CadreNodeConfig.strandMembershipReconciliation}. Armed only for
+   * closed strands launched with a {@link partyMemberPrivateKey};
+   * `{ enabled: false }` disarms the loop (test fixtures that hand-drive the
+   * membership writers). When it names no `pollIntervalMs` the reconciler
+   * mirrors {@link revocationEnforcement}'s cadence — as its IDLE cadence and
+   * the cap of its unfinished-join retry ladder, not as a flat retry interval.
+   */
+  membershipReconciliation?: StrandMembershipReconciliationConfig;
+
+  /**
+   * Called when THIS node's own peer id turns up in the strand's revoked set —
+   * this node's party was removed from the closed strand, or left it. Fires once
+   * per removal (re-armed if the party is re-admitted; a resume also rebuilds the
+   * enforcer and may legitimately re-fire), and only for closed strands with the
+   * gate armed. `CadreNode`
+   * wires it to its `strand:revoked` event; nothing is stopped or torn down on
+   * this node's behalf — what to do about it is the app's call.
+   */
+  onSelfRevoked?: (strandId: string) => void;
+
+  /**
+   * Called when the membership reconciler reports a blocked re-join — this node holds a
+   * staged membership invitation for the strand that its writes cannot redeem, which is
+   * what a REMOVED party handed a fresh invitation hits (see "Reporting a blocked
+   * re-join" in `strand-membership-reconciler.ts` for the two triggers, one of them a
+   * suspicion rather than a verdict). At most once per re-arm of the loop — a further
+   * invitation staged while the loop is still running does not reset the report — and
+   * only for closed strands with a party key. `CadreNode` wires it to its `strand:rejoin-blocked`
+   * event; nothing is stopped or torn down — a remaining manager admitting this party's
+   * key directly is the remedy, and the loop finishes the join by itself once that lands.
+   */
+  onRejoinBlocked?: (strandId: string) => void;
+
+  /**
+   * Tuning for the JOINING machine's first-sync write gate (`strand-first-sync-gate.ts`),
+   * forwarded from {@link CadreNodeConfig.strandFirstSync}: the Header probe cadence while
+   * a non-founder launch is `'syncing'`, and the default budget {@link whenWritable} waits
+   * before rejecting. Founders and machines that already hold the Header are never gated.
+   */
+  firstSync?: StrandFirstSyncConfig;
+
+  /**
+   * Called when a launch that came up `'syncing'` (see {@link firstSync}) becomes
+   * writable: the strand's `Strand.Header` arrived from a peer — or a founder request
+   * ran the bootstrap against the gated database — and `instance.database` is now set.
+   * Never called for a launch that was writable when `startStrand`/`resumeStrand`
+   * resolved. `CadreNode` wires it to its `strand:writable` event. Retained with the
+   * launch config, so a hibernation resume that comes up gated announces too.
+   */
+  onWritable?: (strandId: string) => void;
+
+  /**
+   * Called synchronously inside every commit on the strand's database that wrote something,
+   * with who holds it — the app's commits and cadre-core's own membership writes alike.
+   * `CadreNode` wires it to its `strand:commit-durability` event. Retained with the launch
+   * config and passed to every database built for the launch, so a hibernation wake's rebuilt
+   * database reports too. A throw is logged and does not affect the commit.
+   */
+  onCommitDurability?: (commit: StrandCommitDurability) => void;
+
+  /**
+   * Called when a commit reported `followed` below full replication becomes held by every member
+   * of its cohort, while the database that made it is still open. `CadreNode` wires it to its
+   * `strand:commit-held` event. Retained and isolated like {@link onCommitDurability}.
+   */
+  onCommitHeld?: (event: StrandCommitHeld) => void;
+
+  /**
+   * Re-announce this strand's delegate peer id (`delegatePeerId` — the strand
+   * node's own transport peer id) to ONE relay (`relayAddr`, a direct dial addr
+   * with a trailing `/p2p/<relayPeerId>`), unthrottled. Awaited before every
+   * relay-reservation RE-drive of that relay — never before the first attempt,
+   * which the launch/resume seed pass has already announced for.
+   *
+   * Why: a party control node running the relay server admits the strand node on
+   * an in-memory delegate grant that a relay restart drops, so a re-drive that did
+   * not re-announce first would be denied at the relay's gate. `CadreNode` wires
+   * its own control-mesh announce here; a throw or rejection is logged by the
+   * supervisor and the re-drive still runs. Retained with the launch config, so a
+   * hibernation wake's rebuilt supervisors carry it too.
+   */
+  announceDelegateToRelay?: (strandId: string, relayAddr: string, delegatePeerId: string) => Promise<void>;
+
+  /**
+   * The node-local strand network state (`strand-network-state.ts`): this strand's
+   * slice of it is handed to db-p2p as `persistence`, so the strand node saves its
+   * FRET routing table (with each peer's signed address record) on every connection
+   * and re-imports it when it is next built. Absent ⇒ nothing is saved and the node
+   * starts with an empty table. `CadreNode` passes its store. Retained with the launch
+   * config, so a hibernation wake rebuilds the node over the same state.
+   */
+  networkState?: StrandNetworkStateStore;
+}
+
+/**
+ * Volatile inputs re-resolved when resuming a quiesced strand — the ones that can
+ * have moved since it last ran, and that the rebuilt libp2p node freezes again. Each
+ * one omitted keeps the value the retained launch config already holds, so a resume
+ * that passes nothing rebuilds the strand exactly as it last ran.
+ */
+export interface ResumeStrandOverrides {
+  /**
+   * Freshly-resolved cohort discovery seed (multiaddr strings). Grows as peers are
+   * learned since the strand first launched.
+   */
+  bootstrapNodes?: string[];
+  /**
+   * Freshly-read count of the machines serving this strand, for the repair yardstick
+   * (see {@link StartStrandConfig.servingMachines} — including why the party's
+   * enrolled-machine count is not it). Moves whenever a machine starts or stops
+   * serving the strand, which a hibernating strand does not otherwise notice. Nothing
+   * passes it today. Omitting it here does NOT mean "unknown" the way omitting it from
+   * {@link StartStrandConfig} does — it retains the last value; see `resumeStrand`.
+   */
+  servingMachines?: number;
+}
+
+/** A {@link StrandInstanceManager.resumeStrand} or {@link StrandInstanceManager.quiesceStrand} of one strand. */
+type RuntimeTransition =
+  | { kind: 'resume'; operation: Promise<StrandInstance> }
+  | { kind: 'quiesce'; operation: Promise<void> };
+
+/**
+ * Get the isolated storage path for a specific strand.
+ *
+ * @deprecated This helper is Node-only and throws in React Native (it assumes a
+ * filesystem layout). Use a storage provider factory function instead, which
+ * receives the strandId and can create strand-specific storage paths using
+ * platform-appropriate methods.
+ *
+ * @example
+ * // Instead of using getStrandStoragePath, use a storage provider factory:
+ * const storage = {
+ *   provider: (strandId: string) => new FileRawStorage(`./data/strands/${strandId}`)
+ * };
+ */
+export function getStrandStoragePath(basePath: string, strandId: string): string {
+  // Check if we're in a Node.js environment
+  if (typeof process === 'undefined' || !process.versions?.node) {
+    throw new Error(
+      'getStrandStoragePath is not available in React Native. ' +
+      'Use a storage provider factory function instead.'
+    );
+  }
+
+  // Sanitize strandId for filesystem safety (UUIDs should be safe, but just in case)
+  const safeId = strandId.replace(/[^a-zA-Z0-9-]/g, '_');
+
+  // Build the path with plain string joins rather than the Node `path` module.
+  // A static `require('path')` forces RN bundlers (e.g. Metro) to *resolve* the
+  // module at bundle time even though this Node-only helper throws above before
+  // ever reaching here — joining by hand keeps the module free of any Node
+  // built-in reference, so RN bundles need no `path` shim.
+  const trimmedBase = basePath.replace(/[\\/]+$/, '');
+  return `${trimmedBase}/strands/${safeId}`;
+}
+
+/**
+ * Resolve a storage provider for a specific strand.
+ * If the provider is a factory function, call it with the strandId.
+ *
+ * Called only from {@link StrandInstanceManager.startStrand}, which owns the result
+ * for the instance's lifetime — see `strandStorages`, and which has already asserted
+ * the id is usable as a scope key. Do not move that assertion here: this function
+ * returns early when no provider is configured, and the id becomes a libp2p protocol
+ * prefix on that path too.
+ *
+ * @param provider - Storage provider (instance or factory)
+ * @param strandId - The strand ID to create storage for
+ * @returns The resolved IRawStorage instance, or undefined if no provider
+ */
+function resolveStrandStorage(
+  provider: RawStorageProvider | undefined,
+  strandId: string
+): IRawStorage | undefined {
+  if (!provider) {
+    return undefined;
+  }
+  const storage = typeof provider === 'function' ? provider(strandId) : provider;
+  // Wrapped in the write-through raw-storage cache (quereus-plugin-sereus's cached-storage.ts).
+  // Called ONCE per strand launch — `startStrand` keeps the result for the instance's
+  // lifetime — so the wrap survives every runtime rebuild the instance goes through.
+  return wrapStorageWithCache(storage, strandId);
+}
+
+/**
+ * Verify an sApp's schema signature (fail-closed unless `requireSignedSchemas` is
+ * false) and project the config onto the instance's `sAppInfo`.
+ */
+function verifiedSAppInfo(strandId: string, sAppConfig: SAppConfig, requireSignedSchemas: boolean | undefined): SAppInfo {
+  assertSchemaSignature(sAppConfig, { requireSignature: requireSignedSchemas ?? true });
+  log('Strand %s sApp schema signature verified (author: %s)', strandId, sAppConfig.id);
+  return {
+    id: sAppConfig.id,
+    version: sAppConfig.version,
+    schema: sAppConfig.schema,
+    signature: sAppConfig.signature
+  };
+}
+
+/** Log wording for what a launch runs: the sApp and its version, or a storage replica. */
+function describeSApp(sAppConfig: SAppConfig | undefined): string {
+  return sAppConfig ? `sApp: ${sAppConfig.id} v${sAppConfig.version}` : 'storage replica';
+}
+
+/**
+ * Manages individual strand instances - creates and destroys isolated libp2p nodes
+ * for each strand the cadre participates in.
+ */
+export class StrandInstanceManager {
+  private instances: Map<string, StrandInstance> = new Map();
+  /**
+   * Retained launch config per strand, captured in `startStrand` and cleared in
+   * `stopStrand`. `resumeStrand` reuses it to rebuild a quiesced strand's runtime
+   * without the caller re-threading storage/network/profile/key/sApp config.
+   */
+  private launchConfigs: Map<string, StartStrandConfig> = new Map();
+  /**
+   * The per-strand peer-join block catch-up, keyed by strand id. Private (not on
+   * the public {@link StrandInstance}) because it is runtime plumbing with the
+   * same lifetime as the strand's libp2p node: created in `buildStrandRuntime`,
+   * stopped and dropped in `releaseRuntime` — so quiesce → resume rebuilds it
+   * with a fresh caught-up-peer memo, which is intended (a resumed node may have
+   * missed writes).
+   */
+  private backfills: Map<string, PeerJoinBackfill> = new Map();
+  /**
+   * The per-strand revoked-peer enforcer (closed strands only), keyed by strand
+   * id. Same lifecycle rationale as {@link backfills}: created in
+   * `buildStrandRuntime`, stopped and dropped in `releaseRuntime`, so quiesce →
+   * resume rebuilds it with a fresh (initially empty, fail-open) snapshot.
+   */
+  private revocationEnforcers: Map<string, StrandRevocationEnforcer> = new Map();
+  /**
+   * The per-strand membership reconciler (closed strands launched with a party
+   * key only), keyed by strand id. Same lifecycle rationale as {@link backfills}:
+   * created in `buildStrandRuntime`, stopped and dropped in `releaseRuntime`, so
+   * quiesce → resume rebuilds it and re-runs the idempotent join ladder (redeem a
+   * staged invitation, write this machine's own `MemberPeer` binding) from scratch.
+   */
+  private membershipReconcilers: Map<string, StrandMembershipReconciler> = new Map();
+  /**
+   * The per-strand first-sync gate (`strand-first-sync-gate.ts`), keyed by strand id:
+   * holds a launch's `StrandDatabase` while it is NOT yet published on the instance.
+   * Present from the moment the database object exists in `buildStrandRuntime` until
+   * {@link publishDatabase} hands it to the app — so it doubles as the rollback handle
+   * for a failed `initialize()` — and, for a joiner that holds no `Strand.Header` yet,
+   * for as long as the launch stays `'syncing'`. Same lifecycle as {@link backfills}:
+   * stopped (its database closed) and dropped in `releaseRuntime`, so a quiesce →
+   * resume rebuild re-probes over the same store. An entry and `instance.database`
+   * are mutually exclusive.
+   */
+  private firstSyncGates: Map<string, StrandFirstSyncGate> = new Map();
+  /**
+   * Callers blocked in {@link whenWritable}, keyed by strand id. Resolved by
+   * {@link publishDatabase}, rejected by {@link stopStrand}; each waiter also carries
+   * its own timeout. Deliberately NOT cleared by `releaseRuntime`: a waiter outlives a
+   * quiesce → resume cycle, since the rebuild is what may finally publish the database.
+   */
+  private writableWaiters: Map<string, Set<WritableWaiter>> = new Map();
+  /**
+   * The per-strand relay-reservation supervisors — ONE PER CONFIGURED RELAY, each
+   * over the node's own bare `/p2p-circuit` listener for that relay
+   * (`strand-network-config.ts`) — keyed by strand id. Same lifecycle rationale as
+   * {@link backfills}: started in `buildStrandRuntime` right after the libp2p node
+   * exists, stopped FIRST and dropped in `releaseRuntime` (before the database
+   * closes and the node stops, so no re-drive dials a node being torn down), so
+   * quiesce → resume rebuilds them over the new node. What makes a strand node's
+   * lost relay slot come back on its own — the control node has the same loop in
+   * `CadreNode.reserveRelays`.
+   */
+  private relaySupervisors: Map<string, RelayReservationSupervisor[]> = new Map();
+  /**
+   * The per-strand redial of the strand's other members (`strand-peer-redial.ts`), keyed by
+   * strand id. Same lifecycle as {@link relaySupervisors}: started in `buildStrandRuntime` once
+   * the libp2p node exists, stopped first in `releaseRuntime`, so a hibernating strand dials
+   * nobody and a resume redials over the new node.
+   */
+  private peerRedials: Map<string, StrandPeerRedial> = new Map();
+  /**
+   * The resolved (cache-wrapped) raw storage per strand id — the instance's OWN
+   * store, resolved once in `startStrand` and held until `stopStrand` disposes it.
+   * Private for the same reason `backfills` is: runtime plumbing, not part of the
+   * public {@link StrandInstance}.
+   *
+   * Deliberately NOT touched by `releaseRuntime`: that is what lets a quiesce →
+   * resume cycle rebuild the libp2p node over the SAME store, keeping its
+   * write-through cache warm (and, on an in-memory backend, keeping the strand's
+   * blocks at all). An entry exists iff `instances` does AND the launch config
+   * supplied a storage provider.
+   *
+   * NOTE: a hibernating strand therefore keeps its store — and its share of the
+   * process-wide cache pool — resident for as long as the instance is tracked. That
+   * is the point (a warm wake), and the pool evicts under pressure, so the cost is
+   * one map entry per hibernating strand today. If a device ever hibernates strands
+   * by the hundred, revisit: dropping the store at quiesce and paying for a cold
+   * wake becomes the better trade.
+   */
+  private strandStorages: Map<string, IRawStorage> = new Map();
+  /**
+   * The runtime build in flight per strand id (`buildStrandRuntime`, from `startStrand` or
+   * `resumeStrand`, or a deferred join's database retry), present until it settles. The
+   * instance is tracked for the whole build, with its database not yet constructed or not yet
+   * initialized, so a caller that must act on the FINISHED runtime ({@link attachSApp},
+   * {@link whenRuntimeBuilt}) waits on this instead. A failed build's rejection belongs to the
+   * call that started it.
+   */
+  private runtimeBuilds: Map<string, Promise<unknown>> = new Map();
+  /**
+   * The latest resume or quiesce requested per strand id, present until it settles. Each one
+   * starts only once the one requested before it has settled, so a wake that lands during a
+   * quiesce rebuilds after it, and a quiesce requested during a rebuild releases what that
+   * rebuild finished. The WHOLE operation, not its entry in {@link runtimeBuilds}: that settles
+   * before a resume's own catch records `'error'`, so a joiner awaiting it could read a stale
+   * status.
+   */
+  private transitions: Map<string, RuntimeTransition> = new Map();
+  /**
+   * The pending database retry per strand id of a launch whose database could not be built
+   * because no seed peer had answered yet ({@link deferUnreachableJoin}). Cleared by
+   * {@link releaseRuntime}, so a stop or a quiesce cancels it.
+   */
+  private unreachableJoinRetries: Map<string, unknown> = new Map();
+  private stopping = false;
+
+  constructor() {
+    log('StrandInstanceManager created');
+  }
+
+  /**
+   * Get all current strand instances
+   */
+  getInstances(): Map<string, StrandInstance> {
+    return new Map(this.instances);
+  }
+
+  /**
+   * Get a specific strand instance
+   */
+  getInstance(strandId: string): StrandInstance | undefined {
+    return this.instances.get(strandId);
+  }
+
+  /**
+   * Check if a strand is currently running
+   */
+  hasStrand(strandId: string): boolean {
+    return this.instances.has(strandId);
+  }
+
+  /**
+   * Start a new strand instance.
+   *
+   * A failed launch leaves NOTHING tracked: the instance and its retained launch
+   * config are both dropped before the error is rethrown, so the strand id is
+   * free for a genuine retry. This matches the pre-registration failure path
+   * (a rejected schema signature, which throws before anything is recorded) —
+   * both failure modes of this call leave the same residue: none. Callers learn
+   * of the failure from the rejected promise (and, on the control-discovered
+   * path, from CadreNode's `strand:error` event), not from an error record left
+   * behind in `instances`.
+   *
+   * The strand's raw storage is resolved HERE, once, and held for the instance's
+   * lifetime (see `strandStorages`) — `buildStrandRuntime` only reads it, so a
+   * hibernation wake never re-enters the embedder's provider.
+   */
+  async startStrand(config: StartStrandConfig): Promise<StrandInstance> {
+    const { strandRow, sAppConfig } = config;
+    const strandId = strandRow.Id;
+
+    if (this.stopping) {
+      throw new Error('StrandInstanceManager is stopping');
+    }
+
+    if (this.instances.has(strandId)) {
+      // Callers resolve founder-ness BEFORE reaching here (CadreNode.launchStrand) and
+      // honor a founder request on a tracked instance via foundExistingStrand — so a
+      // founder flag arriving at this early return against a non-founder retained
+      // config is a dropped bootstrap, and must never again be silent.
+      if (config.founder === true && this.launchConfigs.get(strandId)?.founder !== true) {
+        log('startStrand: strand %s is already running but was NOT launched as founder — ' +
+          'this early return DROPS the founder request; use foundExistingStrand', strandId);
+      }
+      log('Strand %s already running', strandId);
+      return this.instances.get(strandId)!;
+    }
+
+    log('Starting strand instance: %s (%s)', strandId, describeSApp(sAppConfig));
+    const tTotal = performance.now();
+
+    // The id becomes two names below: the storage scope key the embedder's provider
+    // turns into a directory or database name, and `networkName` in buildStrandRuntime,
+    // from which the libp2p protocol prefix `/optimystic/strand-<id>` is built. A
+    // replicated strand row carries whatever id the founding node wrote, so neither is
+    // safe unchecked. Asserted HERE rather than in `resolveStrandStorage` because that
+    // function returns early when no provider is configured — a node with no storage
+    // still reaches buildStrandRuntime and still mints the protocol prefix.
+    assertStrandScopeKey(strandId);
+
+    const sAppInfo = sAppConfig ? verifiedSAppInfo(strandId, sAppConfig, config.requireSignedSchemas) : undefined;
+
+    // Resolve this strand's storage ONCE, before anything is recorded. If a factory
+    // function is provided, it is called with the strandId to create strand-specific
+    // storage (e.g. strand-isolated directories). A provider that throws therefore
+    // fails the launch alongside the schema-signature check, leaving nothing tracked.
+    const strandStorage = resolveStrandStorage(config.storage?.provider, strandId);
+    if (strandStorage) {
+      log('Strand %s using provided storage provider', strandId);
+    }
+
+    // Determine latency hint: sApp config > default
+    const latencyHint = sAppConfig?.latencyHint ?? config.defaultLatencyHint;
+
+    const instance: StrandInstance = {
+      strandId,
+      status: 'starting',
+      sAppInfo,
+      memberPrivateKey: strandRow.MemberPrivateKey ?? undefined,
+      connectedPeers: 0,
+      lastActivity: new Date(),
+      latencyHint
+    };
+
+    this.instances.set(strandId, instance);
+    this.launchConfigs.set(strandId, config);
+    if (strandStorage) {
+      this.strandStorages.set(strandId, strandStorage);
+    }
+
+    try {
+      await this.trackRuntimeBuild(strandId, this.buildStrandRuntime(instance, config));
+      timing('[startStrand:%s] total: %dms', strandId, Math.round(performance.now() - tTotal));
+      log('Strand %s started successfully (%s)', strandId, describeSApp(sAppConfig));
+      return instance;
+    } catch (error) {
+      // Status/error first — the (now discarded) record is still what `log` reports on.
+      instance.status = 'error';
+      instance.error = error instanceof Error ? error.message : String(error);
+      log('Failed to start strand %s: %s', strandId, instance.error);
+      // Drop the dead record so this strand id can be launched again. Keep the
+      // `launchConfigs` has an entry iff `instances` does invariant — resumeStrand
+      // reads both, and a config without an instance would strand the config.
+      this.instances.delete(strandId);
+      this.launchConfigs.delete(strandId);
+      // The store this launch resolved goes with it: nothing owns it any more, and a
+      // retained cache wrapper would be handed back (already retired) on a retry.
+      await this.disposeStrandStorage(strandId);
+      throw error;
+    }
+  }
+
+  /** Hold `build` in {@link runtimeBuilds} until it settles, and settle as it does. */
+  private async trackRuntimeBuild<T>(strandId: string, build: Promise<T>): Promise<T> {
+    this.runtimeBuilds.set(strandId, build);
+    try {
+      return await build;
+    } finally {
+      if (this.runtimeBuilds.get(strandId) === build) {
+        this.runtimeBuilds.delete(strandId);
+      }
+    }
+  }
+
+  /**
+   * Wait until no runtime build of `strandId` is in flight. Never rejects: a failed build
+   * is reported by the call that started it, and the caller re-reads the instance after.
+   *
+   * NOTE: after a failed launch the re-read sees the instance gone only because
+   * `startStrand`'s catch deletes it synchronously, and its continuation is queued before
+   * this waiter's. If that cleanup ever awaits before `instances.delete`, a waiter can act
+   * on the dying record; move the cleanup inside the tracked promise then.
+   */
+  private async settleRuntimeBuilds(strandId: string): Promise<void> {
+    for (let build = this.runtimeBuilds.get(strandId); build; build = this.runtimeBuilds.get(strandId)) {
+      await build.catch((error: unknown) => {
+        log('Strand %s: the runtime build waited on failed (reported by its launcher): %o', strandId, error);
+      });
+    }
+  }
+
+  /**
+   * Build (or rebuild) the libp2p node + StrandDatabase for an instance and
+   * attach them, transitioning it to `active`. Shared by `startStrand` (fresh
+   * launch) and `resumeStrand` (rehydrating a quiesced instance). Reads all
+   * volatile inputs (bootstrapNodes, servingMachines, network, profile,
+   * privateKey, sApp config) from `config`, so the caller controls the
+   * cohort-derived values. Storage is the
+   * one input it does NOT re-read from `config`: that belongs to the instance and
+   * comes from `strandStorages`.
+   *
+   * A database that cannot be built because no seed peer has answered yet does not fail the
+   * build: the runtime comes up `'syncing'` without it and the database phase alone is
+   * retried on the same node ({@link deferUnreachableJoin}).
+   */
+  private async buildStrandRuntime(instance: StrandInstance, config: StartStrandConfig): Promise<void> {
+    const strandId = instance.strandId;
+
+    // Refuse a bad `linkRoundTripMs` before any strand bring-up, for the reason
+    // `CadreNode.start()` does: every budget derived from it here is behind a condition
+    // (backfill disabled, no relay addrs), so a zero or NaN declaration would otherwise
+    // surface later inside a path that logs and carries on. `link-budget.ts`.
+    resolveLinkRoundTripMs(config.network?.linkRoundTripMs);
+
+    // The store the instance OWNS (resolved once in `startStrand`), not a fresh
+    // resolution: a rebuild must reach the same backend through the same warm cache.
+    const strandStorage = this.strandStorages.get(strandId);
+
+    // db-p2p namespaces every one of the node's protocol ids by network name
+    // (`/optimystic/<networkName>/...`), so anything dialing this node's own
+    // services must derive its prefix from the SAME string the node was built
+    // with — hence one binding for both, not two literals that can drift.
+    const networkName = `strand-${strandId}`;
+    const protocolPrefix = `/optimystic/${networkName}`;
+
+    // Whether this node runs the circuit-relay server, and its init — the same
+    // resolution the control node takes from the same config (`relay-server.ts`).
+    const relayServer = resolveRelayServer(config.network, config.profile);
+    // The strand-node VIEW of the machine's one `NetworkConfig`, not the control
+    // node's resolution: fixed direct listen ports become ephemeral (two nodes cannot
+    // bind one port), the announce config is dropped (it names the control node's
+    // address), and every configured relay becomes one bare `/p2p-circuit` SEARCH
+    // listener plus a dial addr in `relayAddrs` for the per-relay supervisor started
+    // below — the same route the control node takes, because the configured
+    // `<relay>/p2p-circuit` shape loses its address on a relay restart, a hangup, or
+    // libp2p's own reservation refresh and recovers from none of them. See
+    // `strand-network-config.ts` for all three, and `relay-reservation.ts`.
+    //
+    // `relayAddrs` is destructured OFF: it is not a `createLibp2pNode` option, and
+    // spreading it would hand db-p2p a key it does not know.
+    const { relayAddrs: relayDialAddrs = [], ...addrOptions } = strandNodeAddrs(config.network);
+
+    // The CLOSED-strand revoked-peer gate (see strand-revocation-enforcer.ts).
+    // Constructed BEFORE the libp2p node because the node's options embed its
+    // predicates; its deny set starts empty (admit everything — fail-open, same
+    // posture as bring-up everywhere here) and is first populated after the
+    // strand database initializes below. Registered in the map immediately so
+    // the failure rollback (releaseRuntime) tears it down like every other
+    // runtime component. Open strands skip it entirely: their Member/MemberPeer
+    // tables are empty by schema (`OnlyClosed`), so there is nothing to derive
+    // a deny set from — arming would be pointless polling.
+    const revocationEnforcer =
+      config.strandRow.Type === 'c' && config.revocationEnforcement?.enabled !== false
+        ? new StrandRevocationEnforcer({
+            label: strandId,
+            readRows: () => {
+              const database = instance.database;
+              if (!database) {
+                throw new Error(`strand ${strandId} has no live database`);
+              }
+              return readStrandRevocationRows(database.getDatabase());
+            },
+            // Read per sweep, never captured: this closure is built BEFORE the
+            // libp2p node exists (its options embed the enforcer's predicates),
+            // and `releaseRuntime` clears the field again on quiesce. Undefined
+            // therefore means "no transport right now", which the enforcer
+            // treats as "nothing to tear down".
+            getNetwork: () => instance.libp2pNode,
+            onSelfRevoked: () => config.onSelfRevoked?.(strandId)
+          }, config.revocationEnforcement)
+        : undefined;
+    if (revocationEnforcer) {
+      this.revocationEnforcers.set(strandId, revocationEnforcer);
+    }
+
+    // Closed strands compose revoked-peer denial onto the caller's gater and arm
+    // the fail-closed per-stream gate; open strands keep the raw configured gater
+    // (their peers are legitimately cross-party and nothing is ever revoked).
+    const revocationGateOptions = revocationEnforcer
+      ? {
+          connectionGater: createRevocationConnectionGater(revocationEnforcer, config.network?.connectionGater),
+          authorizeInboundStream: (remotePeerId: string, protocol: string) =>
+            revocationEnforcer.authorizeStream(remotePeerId, protocol)
+        }
+      : config.network?.connectionGater
+        ? { connectionGater: config.network.connectionGater }
+        : {};
+
+    const reactivityOption = strandCohortTopicOption(config.reactivity, strandId);
+    if (reactivityOption.cohortTopic) {
+      log('strand %s: change notifications on', strandId);
+    }
+
+    try {
+      // Bound once: the breadth is also the ceiling on the repair yardstick below, and the
+      // two must be derived from the same resolution. Inside the `try` deliberately — a
+      // rejected clusterSize is a build failure that runs the same cleanup as any other.
+      const strandClusterSize = resolveStrandClusterSize(config.clusterSize);
+
+      let t0 = performance.now();
+      const node = await createLibp2pNode({
+        port: 0, // Random port
+        bootstrapNodes: config.bootstrapNodes ?? [],
+        networkName,
+        storage: strandStorage,
+        fretProfile: config.profile === 'storage' ? 'core' : 'edge',
+        relay: relayServer.enabled,
+        // NOTE: unlike the control node, a strand node has no unauthorized-reservation budget,
+        // so with the per-connection cap off any peer that reaches this node can hold one of
+        // `maxReservations` slots and forward uncapped through it. If strand relays are ever
+        // abused for bandwidth, give strand nodes a budget or a capped init of their own.
+        ...(relayServer.enabled && { relayServerInit: relayServer.init }),
+        clusterSize: strandClusterSize,
+        // Deliberately NOT CONTROL_CLUSTER_POLICY: a strand is application data with its own
+        // breadth reasoning, and the shape match with the control policy is a coincidence.
+        //
+        // The builder declares this node's block-repair corroboration yardstick from the
+        // count of machines SERVING this strand, capped at the breadth above (a block
+        // never lives on more machines than the cohort is wide). Given no count — the
+        // production path today, since no per-strand serving count exists yet — it
+        // returns the frozen STRAND_CLUSTER_POLICY itself, declaring nothing. Resolved
+        // HERE rather than at `startStrand`, so a wake from hibernation would pick up a
+        // serving set that changed while the strand slept.
+        // The read deadline is settled by the same helper the control node uses, because the
+        // two networks ride one link: the host's own, else derived from its declared
+        // `linkRoundTripMs`, else nothing, and the base policy's COHORT_READ_DEADLINE_MS stands.
+        clusterPolicy: strandClusterPolicy(strandClusterSize, {
+          servingMachines: config.servingMachines,
+          cohortQueryTimeoutMs: declaredCohortReadDeadlineMs(config.network)
+        }),
+        arachnode: {
+          enableRingZulu: config.profile === 'storage'
+        },
+        ...(config.privateKey && { privateKey: config.privateKey }),
+        // Built per runtime, not per launch: the adapter stops saving once the strand's
+        // state is forgotten, and a rebuilt runtime is what starts it saving again.
+        //
+        // NOTE: the saved state also restores which peers db-p2p saw serving the strand,
+        // and it counts them in the cohort until FRET finds them unreachable. A node built
+        // while every other member is offline therefore fails its first writes ("Failed to
+        // get super-majority") — for 4-6 s on loopback, the window a running node already
+        // sees when a member drops. If apps report failed sends after relaunching alone on
+        // a slow relayed link, retry the write in the app or raise it upstream (db-p2p's
+        // remembered-serving fallback); dropping `persistence` brings back the split
+        // restart.
+        ...(config.networkState && { persistence: strandNetworkStatePersistence(config.networkState, strandId) }),
+        // With relays configured, built with libp2p's relay discovery off (given a transport
+        // list to wrap — `node-transports.ts`), so the supervisors below are the only thing
+        // that fills the bare `/p2p-circuit` slots: otherwise a cohort seed running the relay
+        // server takes one as soon as it connects (`relay-reservation.ts`, "WHY LIBP2P'S
+        // RELAY DISCOVERY IS SWITCHED OFF").
+        ...nodeTransportsOption(config.network, addrOptions, relayDialAddrs.length > 0),
+        ...(config.network?.noiseCrypto && { noiseCrypto: config.network.noiseCrypto }),
+        // Unconditional, and the same default the control node takes: every node of
+        // every party runs the monitor over its connection to a slow phone, so the
+        // widened ping deadline has to reach the strand nodes too (see
+        // DEFAULT_CONNECTION_MONITOR).
+        connectionMonitor: config.network?.connectionMonitor ?? DEFAULT_CONNECTION_MONITOR,
+        // The same declared link the control node states, and for the same reason: a strand
+        // node is also the listener for every other member's strand node, and Optimystic
+        // derives that listener's limit from this (`link-budget.ts`).
+        linkRoundTripMs: resolveLinkRoundTripMs(config.network?.linkRoundTripMs),
+        // The same dial limits as the control node: a strand node dials the same machines,
+        // through the same relays and gates.
+        ...optimysticDialLimits(config.network?.linkRoundTripMs),
+        // Listen entries plus the WebSocket transport switch they imply — a strand node
+        // announces nothing the operator configured (`strand-network-config.ts`), and
+        // spreads AFTER `transports` above because the switch is a no-op whenever a
+        // transport list was passed. The bare `/p2p-circuit` entries (one per relay) open
+        // no connection here; each registers a pending reservation that the matching
+        // supervisor below fills. That is what gives a NAT'd strand node a reachable relay
+        // slot — and works against a party-run relay because the launch path announced
+        // this strand's derived peerId to it first (delegate admission; see cadre-node.ts).
+        ...addrOptions,
+        // The raw configured gater (open strands), or the revocation-composed
+        // gater plus the fail-closed per-stream revoked-peer gate (closed
+        // strands) — resolved above, before the try.
+        ...revocationGateOptions,
+        // `{}` unless the operator enabled change notifications for this strand (`strand-reactivity.ts`).
+        ...reactivityOption
+      }) as Libp2pNodeWithRepo;
+      timing('[buildStrandRuntime:%s] createLibp2pNode: %dms', strandId, Math.round(performance.now() - t0));
+
+      instance.libp2pNode = node;
+
+      // One reservation supervisor PER RELAY, started now so the first attempts
+      // overlap the database bring-up below (the relay is never in a strand's
+      // Optimystic cohort — its protocol ids are namespaced `/optimystic/strand-<id>/…`
+      // — so an open relay connection cannot disturb that bring-up). Their first
+      // attempts are awaited before the strand goes `active`, below. Registered in the
+      // map immediately so the failure rollback (releaseRuntime) stops them like every
+      // other runtime component.
+      this.relaySupervisors.set(strandId, this.startRelaySupervisors(strandId, node, relayDialAddrs, config));
+      // Redials the strand's other members whenever none is connected, on the budget the
+      // declared link gives a relayed dial — what reconnects two restarted phones on a slow
+      // link, where FRET's own redial cannot (`strand-peer-redial.ts`). The relays are
+      // excluded by peer id: the node is connected to them throughout, and they are no member.
+      this.peerRedials.set(strandId, superviseStrandPeerRedial({
+        label: strandId,
+        node,
+        seed: config.bootstrapNodes ?? [],
+        excluded: new Set(groupAddrsByPeerId([...relayDialAddrs]).keys()),
+        ...(revocationEnforcer && { isRevoked: (peerId: string) => revocationEnforcer.isRevoked(peerId) })
+      }, strandRedialTimings(config.network?.linkRoundTripMs)));
+
+      const databaseBuild = await this.buildStrandDatabase(instance, config, node);
+      if (!this.isCurrentRuntime(instance, node)) {
+        // A stop released this runtime during the database phase (`stopStrand` does not
+        // wait for builds); arming the loops below would restart them on a stopped node.
+        log('Strand %s was stopped while its runtime was being built', strandId);
+        return;
+      }
+
+      // The membership reconciler: finish this party's join on every machine —
+      // redeem a staged formation invitation if one is pending, then write the
+      // durable machine→party binding (Strand.MemberPeer). Armed here, the one
+      // seam where the strand's transport peer id exists and its database has been
+      // attempted, for launch AND hibernation resume, for closed strands that carry the
+      // party's own membership key. start() kicks an immediate pass WITHOUT
+      // awaiting it — a joiner at this instant has not synced the founder's rows
+      // and may lack write quorum, so bring-up is never blocked; an unfinished join
+      // then retries on a short doubling ladder capped at the configured
+      // pollIntervalMs (which is also the flat cadence while nobody has admitted this
+      // party yet), and the loop stops once the member row and this machine's own
+      // binding are both in place. Armed AFTER the database phase above, so a gated or
+      // deferred launch always has a reconciler registered by the time it can publish.
+      if (config.strandRow.Type === 'c' && config.partyMemberPrivateKey
+        && config.membershipReconciliation?.enabled !== false) {
+        const reconciler = new StrandMembershipReconciler({
+          label: strandId,
+          partyMemberPrivateKey: config.partyMemberPrivateKey,
+          // Read per pass, never captured — same lifecycle argument as the
+          // enforcer's getNetwork: quiesce drops both handles.
+          getDatabase: () => instance.database?.getDatabase(),
+          getOwnPeerId: () => instance.libp2pNode?.peerId.toString(),
+          pendingInvite: config.pendingMembershipInvite,
+          // The enforcer's CURRENT snapshot view of this node's own peer id — the
+          // reconciler stops rather than fight a self-revocation. Absent when the
+          // gate is disarmed (fail-open: the loop just runs).
+          isSelfRevoked: revocationEnforcer
+            ? () => {
+                const ownPeerId = instance.libp2pNode?.peerId.toString();
+                return ownPeerId !== undefined && revocationEnforcer.isRevoked(ownPeerId);
+              }
+            : undefined,
+          onRejoinBlocked: () => config.onRejoinBlocked?.(strandId)
+        }, {
+          pollIntervalMs: config.membershipReconciliation?.pollIntervalMs
+            ?? config.revocationEnforcement?.pollIntervalMs
+        });
+        this.membershipReconcilers.set(strandId, reconciler);
+        reconciler.start();
+      }
+
+      // Arm the revoked-peer deny-set poll now that the database it reads
+      // exists. start() kicks an immediate refresh WITHOUT awaiting it —
+      // bring-up and resume are never blocked on a membership read; until the
+      // first read lands the empty snapshot admits everyone (fail-open).
+      revocationEnforcer?.start();
+
+      // Peer-join block catch-up: push this strand's own blocks to each newly
+      // connected peer, so a machine that joined after blocks were committed
+      // still ends up physically holding them (without per-strand storage there
+      // is nothing to copy). Armed for EVERY stored strand: `PeerJoinBackfill`
+      // only does work when the strand's libp2p node reports a peer connection,
+      // so on a device that is genuinely alone it is inert, and arming it at
+      // launch is what closes the "founded alone, never replicates" hole — a
+      // peer that joins later gets the founder's blocks without any relaunch.
+      // No `authorizePeer` gate, deliberately — see the module comment in
+      // peer-join-backfill.ts for the strand-side argument (and why the control
+      // network, which DOES gate, is different).
+      //
+      // NOTE: cost is one PeerJoinBackfill object + one `peer:identify` listener
+      // per running strand — linear in strand count, negligible at the handful a
+      // device or host runs today. If a node ever hosts strands by the hundred,
+      // move to one shared listener that dispatches by strand id.
+      if (strandStorage && config.backfill?.enabled !== false) {
+        if (node.keyNetwork) {
+          const backfill = new PeerJoinBackfill({
+            label: strandId,
+            libp2p: node,
+            peerNetwork: node.keyNetwork,
+            storage: strandStorage,
+            // The same prefix the receiver registered its block-transfer handler
+            // under — derived from networkName above, never re-spelled here.
+            protocolPrefix
+          }, {
+            // Dial and response deadlines counted in link round trips, not fixed milliseconds:
+            // a relayed dial costs a fixed number of exchanges, so a fixed budget has a link
+            // speed above which this catch-up can never reach the peer at all. See
+            // `link-budget.ts`. Spread BEFORE the host's own config so an explicit
+            // `strandBackfill.dialTimeoutMs` still wins.
+            ...peerJoinPushBudget(config.network?.linkRoundTripMs),
+            ...config.backfill
+          });
+          backfill.start();
+          this.backfills.set(strandId, backfill);
+        } else {
+          log('Strand %s: libp2p node exposes no keyNetwork; peer-join block catch-up is inert', strandId);
+        }
+      }
+
+      // The relay supervisors' FIRST attempts, all of them, before `active`: the happy
+      // path still publishes its circuit addr before `addStrand` resolves (the
+      // strand-addr RPC answers read it, and the same-party circuit scenario asserts
+      // it). Fail-SOFT — a first attempt that lands nothing does not fail the launch.
+      // The strand's database is up, and the supervisor keeps trying on its backoff;
+      // failing here would only trade that for `StrandWatcher`'s full-rebuild retry.
+      //
+      // NOTE: a relay that is down costs this launch one full drive — the reservation budget
+      // counted from the declared link round trip, 18 s at its default (`link-budget.ts`) — and
+      // `StrandWatcher` launches strands one at a time, so N strands cost N of those in
+      // bring-up during a relay outage, and MORE on a host that declared a slower link. If that
+      // ever matters, stop awaiting here (the circuit addr then lands after `active`) rather
+      // than shortening the drive.
+      t0 = performance.now();
+      await this.awaitFirstRelayAttempts(strandId);
+      timing('[buildStrandRuntime:%s] relay first attempts: %dms', strandId, Math.round(performance.now() - t0));
+
+      // No `lastActivity` stamp: bringing a runtime up is not activity. A check-in marks
+      // activity before its resume and must see only what landed while the strand was up.
+      instance.status = liveStrandStatus(instance);
+      if (databaseBuild === 'awaiting-seed') {
+        // Armed only now that the build is over: a retry firing mid-build would find the
+        // build in flight.
+        this.deferUnreachableJoin(instance);
+      } else if (instance.status === 'syncing') {
+        log('Strand %s launched as a joiner with no Strand.Header held yet — writes are withheld until ' +
+          'a member of the strand is reached', strandId);
+      }
+    } catch (error) {
+      // Roll back any partially-attached runtime so the instance is left with
+      // NEITHER handle. Otherwise the `libp2pNode || database` "already live"
+      // guard in resumeStrand/handleStrandWake would treat a half-built strand
+      // as healthy — leaking the libp2p node and never retrying the rebuild.
+      await this.releaseRuntime(instance).catch((cleanupErr) => {
+        log('buildStrandRuntime cleanup for strand %s also failed: %o', strandId, cleanupErr);
+      });
+      throw error;
+    }
+  }
+
+  /**
+   * The database phase of a runtime build, on the libp2p node already attached: create the
+   * StrandDatabase, initialize it, then publish it or arm its first-sync gate. Run by
+   * {@link buildStrandRuntime} and again by each {@link retryUnreachableJoin}, with a new
+   * StrandDatabase per attempt.
+   *
+   * Returns `'awaiting-seed'` when initialize failed only because no seed peer has answered
+   * (`cohort-unreachable`): that attempt's database is closed and the node keeps running. Any
+   * other failure throws, with the attempt's database left with its gate for
+   * {@link releaseRuntime} to close.
+   */
+  private async buildStrandDatabase(
+    instance: StrandInstance,
+    config: StartStrandConfig,
+    node: Libp2pNodeWithRepo
+  ): Promise<StrandDatabaseBuild> {
+    const strandId = instance.strandId;
+    // Held by the first-sync gate (registered before initialize) rather than attached
+    // to the instance: a failed init is then cleaned up through the gate (close() is safe on
+    // a partially-initialized db), and the app only ever sees `instance.database` once this
+    // machine is allowed to write to the strand.
+    let t0 = performance.now();
+    const strandDb = new StrandDatabase({
+      strandId,
+      sAppConfig: config.sAppConfig,
+      libp2pNode: node,
+      coordinatedRepo: node.coordinatedRepo,
+      // Founder bootstrap inputs: the strand's type drives which membership rows
+      // are written, and the PARTY's own key (partyMemberPrivateKey, resolved by the
+      // caller from the control-layer StrandPartyKey row) derives the founding
+      // Member/Manager key. The row's shared MemberPrivateKey rides along as the
+      // read-gating secret only — it derives nobody's identity.
+      strandType: config.strandRow.Type,
+      memberPrivateKey: config.strandRow.MemberPrivateKey ?? undefined,
+      partyMemberPrivateKey: config.partyMemberPrivateKey,
+      founder: config.founder,
+      onCommitDurability: config.onCommitDurability,
+      onCommitHeld: config.onCommitHeld
+    });
+    const gate = new StrandFirstSyncGate({
+      label: strandId,
+      database: strandDb,
+      onHeaderHeld: () => this.publishDatabase(instance, strandDb)
+    }, config.firstSync);
+    this.firstSyncGates.set(strandId, gate);
+    try {
+      await strandDb.initialize();
+    } catch (error) {
+      if (!isCohortUnreachableRead(error)) {
+        throw error;
+      }
+      await this.discardDatabaseAttempt(strandId, gate);
+      log('Strand %s: no seed peer has answered yet, so this joiner cannot build its database (%s); ' +
+        'its runtime stays up and the build is retried every %dms',
+        strandId, error instanceof Error ? error.message : String(error), UNREACHABLE_JOIN_RETRY_MS);
+      return 'awaiting-seed';
+    }
+    timing('[buildStrandRuntime:%s] strandDatabase.initialize: %dms', strandId, Math.round(performance.now() - t0));
+    if (this.firstSyncGates.get(strandId) !== gate) {
+      // A stop released the runtime while initialize ran, closing this database through its
+      // gate part-way; close what initialize finished after that.
+      await strandDb.close();
+      return 'released';
+    }
+
+    // The first-sync write gate. A founder just wrote the Header in its bootstrap; any
+    // other machine may commit only once it holds the Header — which it can only have
+    // received from a peer (now, or on an earlier run over the same store) — and has read
+    // each App table once, so their collections are fetched too. Publishing the database
+    // is what makes the strand writable to the app; until then the instance stays
+    // `'syncing'` and the gate re-probes on its cadence. Everything armed after the database
+    // phase reads `instance.database` lazily, so the reconciler and enforcer wait with it —
+    // and `publishDatabase` kicks the reconciler when the wait ends, so a joiner's membership
+    // rows land with the strand rather than one retry later.
+    t0 = performance.now();
+    const synced = config.founder === true || await strandFirstSyncComplete(strandDb.getDatabase(), strandId);
+    timing('[buildStrandRuntime:%s] first-sync probe: %dms', strandId, Math.round(performance.now() - t0));
+    if (this.firstSyncGates.get(strandId) !== gate) {
+      // A stop released the runtime during the probe and closed this database through its gate.
+      return 'released';
+    }
+    if (synced) {
+      this.publishDatabase(instance, strandDb);
+    } else {
+      gate.start();
+    }
+    return 'built';
+  }
+
+  /** Close a database attempt that failed to initialize, and drop its gate if it is still the registered one. */
+  private async discardDatabaseAttempt(strandId: string, gate: StrandFirstSyncGate): Promise<void> {
+    gate.stop();
+    if (this.firstSyncGates.get(strandId) === gate) {
+      this.firstSyncGates.delete(strandId);
+    }
+    await gate.database.close();
+  }
+
+  /**
+   * Keep a launch `'syncing'` when its database could not be built because no seed peer has
+   * answered yet, and retry the database build every {@link UNREACHABLE_JOIN_RETRY_MS} on the
+   * runtime already up.
+   *
+   * Optimystic (since 1.11.0) will not rule a block never-created while any of a node's
+   * `bootstrapNodes` is unheard, so a joiner alone with an unreachable seed cannot build even
+   * the schema catalog: `strandDb.initialize()` fails with `cohort-unreachable`. A founder
+   * given a seed is refused the same way (normally it has none: no sibling runs the strand yet). Refusing is
+   * right (building it alone would fork the catalog), but failing the launch would drop the
+   * strand until the watcher relaunched it, and `addStrand` would reject with a raw Optimystic
+   * error. Deferring gives a joiner the same contract the first-sync gate gives one that did
+   * build: the strand stays tracked, {@link whenWritable} waits and rejects with the retryable
+   * `StrandAwaitingFirstSyncError`, and the launch proceeds once a seed peer answers.
+   *
+   * The libp2p node, its connections and the background loops stay up meanwhile. A seed peer
+   * can answer just after the schema read gave up on it — on a slow link it connects and its
+   * peer-join backfill delivers the whole store — and the retry then builds the database, and
+   * runs the first-sync probe, while that peer is still connected.
+   */
+  private deferUnreachableJoin(instance: StrandInstance): void {
+    const strandId = instance.strandId;
+    const timer = setTimeout(() => {
+      this.unreachableJoinRetries.delete(strandId);
+      void this.retryUnreachableJoin(instance);
+    }, UNREACHABLE_JOIN_RETRY_MS);
+    (timer as { unref?: () => void }).unref?.();
+    this.unreachableJoinRetries.set(strandId, timer);
+  }
+
+  /**
+   * One retry scheduled by {@link deferUnreachableJoin}: the database phase again, on the
+   * runtime still attached. Skipped when the strand was stopped or released since — a quiesce
+   * releases it and the resume after rebuilds the whole runtime — or already has a database.
+   * A failure other than an unanswered seed releases the runtime and marks the strand
+   * `'error'`, as a failed resume does.
+   */
+  private async retryUnreachableJoin(instance: StrandInstance): Promise<void> {
+    const strandId = instance.strandId;
+    const config = this.launchConfigs.get(strandId);
+    // Every strand node is built by `buildStrandRuntime`, as a node with its coordinated repo.
+    const node = instance.libp2pNode as Libp2pNodeWithRepo | undefined;
+    if (!config || !node || !this.isCurrentRuntime(instance, node) || instance.database
+      || this.firstSyncGates.has(strandId) || this.isQuiescing(strandId)) {
+      return;
+    }
+    let build: StrandDatabaseBuild;
+    try {
+      // Tracked like any runtime build, so a wake, a quiesce or an sApp attach waits for it
+      // rather than acting on a database that is still initializing.
+      build = await this.trackRuntimeBuild(strandId, this.buildStrandDatabase(instance, config, node));
+    } catch (error) {
+      if (!this.isCurrentRuntime(instance, node)) {
+        log('Strand %s: a database retry failed after its runtime was released: %o', strandId, error);
+        return;
+      }
+      await this.releaseRuntime(instance).catch((cleanupErr) => {
+        log('Strand %s: releasing the runtime after a failed database retry also failed: %o', strandId, cleanupErr);
+      });
+      instance.status = 'error';
+      instance.error = error instanceof Error ? error.message : String(error);
+      log('Strand %s: building the database of a joiner that was waiting for a seed peer failed: %s', strandId, instance.error);
+      return;
+    }
+    if (build === 'awaiting-seed' && this.isCurrentRuntime(instance, node) && !this.isQuiescing(strandId)) {
+      this.deferUnreachableJoin(instance);
+    }
+  }
+
+  /** Whether `instance` is still tracked, with `node` still its runtime's libp2p node. */
+  private isCurrentRuntime(instance: StrandInstance, node: Libp2p): boolean {
+    return !this.stopping && this.instances.get(instance.strandId) === instance && instance.libp2pNode === node;
+  }
+
+  /**
+   * Release an instance's strand-network runtime: stop the relay supervisors and
+   * the other background loops, close the StrandDatabase, then stop the libp2p
+   * node (construction order in reverse), clearing both fields and zeroing
+   * connectedPeers. Tolerant of partially-built state — either handle may be
+   * absent — so it doubles as rollback for a failed `buildStrandRuntime`. Shared
+   * by `quiesceStrand`, `stopStrand`, and that rollback path.
+   *
+   * Leaves `strandStorages` untouched by design — the store outlives the runtime it
+   * was built into, which is what makes a resume warm. Only `stopStrand` disposes it.
+   */
+  private async releaseRuntime(instance: StrandInstance): Promise<void> {
+    // Relay supervisors FIRST of all — before anything below is torn down — so no
+    // re-drive is scheduled against a node being stopped. `stop()` is synchronous
+    // and never awaits a drive, so a relay that is down cannot delay this teardown.
+    // `stop()` also cancels the drive already in flight, so an attempt started
+    // moments before this does not keep dialing and polling the node the lines
+    // below are stopping; its result is discarded either way.
+    const unreachableJoinRetry = this.unreachableJoinRetries.get(instance.strandId);
+    if (unreachableJoinRetry !== undefined) {
+      clearTimeout(unreachableJoinRetry as ReturnType<typeof setTimeout>);
+      this.unreachableJoinRetries.delete(instance.strandId);
+    }
+    const relaySupervisors = this.relaySupervisors.get(instance.strandId);
+    if (relaySupervisors) {
+      relaySupervisors.forEach((supervisor) => supervisor.stop());
+      this.relaySupervisors.delete(instance.strandId);
+    }
+    this.peerRedials.get(instance.strandId)?.stop();
+    this.peerRedials.delete(instance.strandId);
+    // Backfill next — before the database closes and the libp2p node stops — so
+    // no NEW catch-up push is issued against a torn-down transport. A push already
+    // in flight is not awaited; it fails into the module's own per-chunk catch.
+    const backfill = this.backfills.get(instance.strandId);
+    if (backfill) {
+      backfill.stop();
+      this.backfills.delete(instance.strandId);
+    }
+    // The revoked-peer enforcer goes with the runtime it gated: a resume
+    // rebuilds it with a fresh (initially empty, fail-open) snapshot.
+    const revocationEnforcer = this.revocationEnforcers.get(instance.strandId);
+    if (revocationEnforcer) {
+      revocationEnforcer.stop();
+      this.revocationEnforcers.delete(instance.strandId);
+    }
+    // The membership reconciler likewise: a resume rebuilds it and re-runs the
+    // idempotent ladder, which is what makes resume-after-partial-join heal.
+    const membershipReconciler = this.membershipReconcilers.get(instance.strandId);
+    if (membershipReconciler) {
+      membershipReconciler.stop();
+      this.membershipReconcilers.delete(instance.strandId);
+    }
+    // A still-gated database (a joiner that never received the Header, or a launch whose
+    // initialize() failed) is closed through its gate; a published one through the
+    // instance. Never both — publishing moves the handle from the gate to the instance.
+    const gate = this.firstSyncGates.get(instance.strandId);
+    if (gate) {
+      gate.stop();
+      this.firstSyncGates.delete(instance.strandId);
+      await gate.database.close();
+    }
+    if (instance.database) {
+      await instance.database.close();
+      instance.database = undefined;
+    }
+    if (instance.libp2pNode) {
+      await instance.libp2pNode.stop();
+      instance.libp2pNode = undefined;
+    }
+    instance.connectedPeers = 0;
+  }
+
+  /**
+   * One {@link superviseRelayReservation} per relay dial addr, each over exactly
+   * that relay so "held" is judged per relay (`circuitMultiaddrsVia`) and losing
+   * one relay re-drives only that one. They are the only thing that fills the node's
+   * search slots once the node is built with libp2p's relay discovery off
+   * (`nodeTransportsOption`). Default retry timings — the control node's (2 s
+   * doubling to 60 s between failed attempts, a 5 s liveness check while held). Each DRIVE's
+   * own deadline is counted in link round trips at this host's declared
+   * `network.linkRoundTripMs` (`link-budget.ts`), so the same declaration that lengthens the
+   * control node's drive lengthens these.
+   *
+   * Each supervisor's `beforeRedrive` is the caller's
+   * {@link StartStrandConfig.announceDelegateToRelay} for THIS relay and THIS
+   * node's peer id, so a re-drive after a relay restart is preceded by a fresh
+   * delegate grant on the relay about to be dialed. Returns `[]` when no relay is
+   * configured — the strand then simply has no circuit listener to fill.
+   */
+  private startRelaySupervisors(
+    strandId: string,
+    node: Libp2pNodeWithRepo,
+    relayDialAddrs: readonly string[],
+    config: StartStrandConfig
+  ): RelayReservationSupervisor[] {
+    if (relayDialAddrs.length === 0) {
+      return [];
+    }
+    const announce = config.announceDelegateToRelay;
+    const delegatePeerId = node.peerId.toString();
+    const timeoutMs = relayReservationBudgetMs(config.network?.linkRoundTripMs);
+    return relayDialAddrs.map((relayAddr) => superviseRelayReservation(node, [relayAddr], {
+      timeoutMs,
+      ...(announce && { beforeRedrive: () => announce(strandId, relayAddr, delegatePeerId) })
+    }));
+  }
+
+  /**
+   * Wait for every relay supervisor's first attempt to settle, concurrently, and log
+   * the ones that landed nothing. Never throws — a strand whose relay is unreachable
+   * at launch still comes up, with its supervisors retrying in the background.
+   */
+  private async awaitFirstRelayAttempts(strandId: string): Promise<void> {
+    const supervisors = this.relaySupervisors.get(strandId) ?? [];
+    if (supervisors.length === 0) {
+      return;
+    }
+    await Promise.all(supervisors.map((supervisor) => supervisor.firstAttempt));
+    supervisors.forEach((supervisor) => {
+      if (supervisor.lastError !== null) {
+        log('Strand %s: relay reservation not held after the first attempt (retrying in the background): %s',
+          strandId, supervisor.lastError);
+      }
+    });
+  }
+
+  /**
+   * Drop the strand's owned store and release this strand's claim on its cache. The
+   * wrapper counts holders, so the cache is emptied and unregistered from the shared
+   * pool only if no other scope still holds it. Called only where the strand's whole
+   * lifetime ends (`stopStrand`, and the failed-launch rollback) — never on a quiesce.
+   *
+   * A dispose failure is logged, not thrown: the store is already unreferenced here,
+   * and failing the stop over a cache-bookkeeping error would leave the caller unable
+   * to tear the strand down. `disposeStorageCache` no-ops for an unwrapped store
+   * (e.g. `MemoryRawStorage`), so no instanceof test is needed.
+   */
+  private async disposeStrandStorage(strandId: string): Promise<void> {
+    const storage = this.strandStorages.get(strandId);
+    if (!storage) {
+      return;
+    }
+    this.strandStorages.delete(strandId);
+    try {
+      await disposeStorageCache(storage);
+    } catch (error) {
+      log('Failed to dispose storage cache for strand %s: %o', strandId, error);
+    }
+  }
+
+  /**
+   * Refresh a strand's revoked-peer deny set NOW — and, with it, run the
+   * teardown sweep that hangs up every connected revoked peer.
+   *
+   * The gate is otherwise poll-driven (default
+   * `DEFAULT_REVOCATION_POLL_INTERVAL_MS`), because neither a replicated
+   * revocation nor a local `revokeMember`/`leaveStrand` raises anything this
+   * runtime can hook. A caller that just wrote a revocation against the
+   * strand's database should follow it with this call, which makes the cut
+   * immediate instead of waiting out the interval.
+   *
+   * Quiet no-op — matching `quiesceStrand`'s posture — when the strand is not
+   * tracked, is quiesced, is open, or has the gate disabled: in every one of
+   * those cases there is no enforcer to refresh. Never rejects (the enforcer's
+   * refresh contains its own failures), and resolves only once the sweep has
+   * finished.
+   */
+  async refreshRevocationEnforcement(strandId: string): Promise<void> {
+    const enforcer = this.revocationEnforcers.get(strandId);
+    if (!enforcer) {
+      log('refreshRevocationEnforcement: strand %s has no armed revocation enforcer', strandId);
+      return;
+    }
+    await enforcer.refresh();
+  }
+
+  /**
+   * A fresh membership invitation was staged for `strandId` (`CadreNode`'s
+   * `adoptFormationMembershipInvite`): re-arm the strand's membership reconciler so the
+   * invitation is attempted now rather than never — the loop finished during the first
+   * join and would otherwise stay stopped until a relaunch. Quiet no-op when no reconciler
+   * is armed: a first formation stages before the strand is added, and bring-up then arms
+   * a loop that finds the invitation by itself. Returns at once; the pass runs on the
+   * loop's own serialized chain and never rejects.
+   */
+  notifyMembershipInviteStaged(strandId: string): void {
+    const reconciler = this.membershipReconcilers.get(strandId);
+    if (!reconciler) {
+      log('notifyMembershipInviteStaged: strand %s has no armed membership reconciler — bring-up will find the invitation', strandId);
+      return;
+    }
+    void reconciler.rearm();
+  }
+
+  /**
+   * Best-effort removal of THIS machine's own `Strand.MemberPeer` binding — the
+   * self arm of `removeMemberPeer`, signed with the retained launch config's party
+   * key. Called by `CadreNode.unpublishStrand` BEFORE the local stop, while the
+   * strand database and transport are still live: after the unpublish commits, the
+   * party's `StrandPartyKey` control row is gone, so no restart could ever sign
+   * this removal again — the in-memory retained key is the last chance.
+   *
+   * Never throws. Quiet no-op for a strand that is untracked, open, or launched
+   * without a party key; a quiesced (hibernating) strand, or a strand whose write
+   * quorum is already unreachable, leaves the stale binding behind with a log —
+   * it grants nothing today (admission is not allowlist-gated) and only
+   * mis-credits diversity. The strand's own membership reconciler is stopped —
+   * and its in-flight pass awaited — FIRST, so a pass that was already past its
+   * own stopped check cannot re-register the binding this is about to clear.
+   *
+   * NOTE: deliberately NOT handled here (or anywhere yet): clearing the bindings
+   * of a machine removed from the CADRE at the control layer, or of a party's
+   * OTHER machines when the party unpublishes — the party still holds the key on
+   * every surviving machine and can clear from any of them, and the stale rows
+   * grant nothing while admission is deny-list only. Revisit when
+   * `feat-strand-member-allowlist-admission` lands and a stale binding starts
+   * granting admission rather than merely mis-crediting diversity.
+   */
+  async clearOwnMemberPeerBinding(strandId: string): Promise<void> {
+    const instance = this.instances.get(strandId);
+    const config = this.launchConfigs.get(strandId);
+    if (!instance || !config || config.strandRow.Type !== 'c' || !config.partyMemberPrivateKey) {
+      return;
+    }
+    const reconciler = this.membershipReconcilers.get(strandId);
+    if (reconciler) {
+      reconciler.stop();
+      await reconciler.settle();
+    }
+    const db = instance.database?.getDatabase();
+    const peerId = instance.libp2pNode?.peerId.toString();
+    if (!db || !peerId) {
+      log('clearOwnMemberPeerBinding: strand %s has no live runtime — own MemberPeer binding left in place', strandId);
+      return;
+    }
+    try {
+      // Background write on the app's database: never join a transaction the app has open.
+      await removeMemberPeer(db, {
+        memberKeyPair: strandMemberKeyPair(config.partyMemberPrivateKey),
+        peerId
+      }, { joinOpenTransaction: false });
+      log('clearOwnMemberPeerBinding: strand %s — own binding for peer %s cleared', strandId, peerId);
+    } catch (error) {
+      log('clearOwnMemberPeerBinding: strand %s — best-effort clear failed (strand may already be unreachable): %o',
+        strandId, error);
+    }
+  }
+
+  /**
+   * Quiesce a strand: release its strand-network resources (stop the libp2p node,
+   * close the StrandDatabase) while RETAINING the instance record — identity,
+   * sAppInfo, keys, latency hint, metadata — and its launch config so it can be
+   * resumed later, then mark it `hibernating`. Mechanically this is `stopStrand` minus
+   * the instance/config deletion. A tracked strand with no runtime left (a failed
+   * rebuild already released it) is only marked; an untracked one is left alone.
+   *
+   * Ordered with {@link resumeStrand} (see {@link transitions}): it starts once the resume
+   * or launch build before it has settled, so it releases a finished runtime, and the
+   * status write is its last step, so a resume requested meanwhile writes after it. A call
+   * made while another quiesce is the latest request joins it. A failed release leaves the
+   * status as it was and rejects.
+   */
+  async quiesceStrand(strandId: string): Promise<void> {
+    const previous = this.transitions.get(strandId);
+    if (previous?.kind === 'quiesce') {
+      log('quiesceStrand: strand %s — joining the quiesce already requested', strandId);
+      return previous.operation;
+    }
+    const operation = this.runQuiesce(strandId, previous);
+    this.holdTransition(strandId, { kind: 'quiesce', operation });
+    return operation;
+  }
+
+  /** Body of {@link quiesceStrand}. */
+  private async runQuiesce(strandId: string, previous: RuntimeTransition | undefined): Promise<void> {
+    await this.settleTransition(strandId, previous);
+    await this.settleRuntimeBuilds(strandId);
+    const instance = this.instances.get(strandId);
+    if (!instance) {
+      log('quiesceStrand: strand %s not found', strandId);
+      return;
+    }
+    if (instance.libp2pNode || instance.database) {
+      log('Quiescing strand instance: %s', strandId);
+      await this.releaseRuntime(instance);
+      log('Strand %s quiesced (resources released, instance retained)', strandId);
+    } else {
+      log('quiesceStrand: strand %s already quiesced', strandId);
+    }
+    instance.status = 'hibernating';
+  }
+
+  /**
+   * Whether the latest runtime transition requested for `strandId` is a quiesce that has
+   * not finished. Its handles may still be attached and its status still live, but it is on
+   * its way to `hibernating`, so a caller that needs it up must {@link resumeStrand} it — the
+   * resume runs after the quiesce — rather than treat it as live.
+   */
+  isQuiescing(strandId: string): boolean {
+    return this.transitions.get(strandId)?.kind === 'quiesce';
+  }
+
+  /**
+   * Resume a previously-quiesced strand: rebuild its libp2p node + StrandDatabase
+   * from the retained launch config and re-attach them, transitioning it back to
+   * `active`. `overrides` re-applies volatile inputs that may have changed since
+   * launch (the cohort `bootstrapNodes` seed and the strand's `servingMachines`
+   * count) and updates the retained config so a later resume reuses the latest
+   * values. Returns the live instance unchanged if it is already running.
+   *
+   * Overlapping calls share one rebuild: a call made while a resume is the latest request
+   * for the strand — a wake landing during a check-in — joins it and settles exactly as it
+   * does, `'error'` status included. The joiner's own `overrides` are ignored; the first
+   * resume's seed wins, and both callers resolved the cohort seed moments apart. A call made
+   * while a quiesce is the latest request runs after that quiesce and rebuilds (see
+   * {@link transitions}), and one made while `startStrand` is still building waits for that
+   * build rather than starting a second.
+   */
+  async resumeStrand(strandId: string, overrides?: ResumeStrandOverrides): Promise<StrandInstance> {
+    if (this.stopping) {
+      throw new Error('StrandInstanceManager is stopping');
+    }
+    // No `await` between this read and `holdTransition`, or a second caller slips through
+    // and builds.
+    const previous = this.transitions.get(strandId);
+    if (previous?.kind === 'resume') {
+      log('resumeStrand: strand %s — joining the resume already in flight', strandId);
+      return previous.operation;
+    }
+    const operation = this.runResume(strandId, previous, overrides);
+    this.holdTransition(strandId, { kind: 'resume', operation });
+    return operation;
+  }
+
+  /** Record `transition` as the latest for `strandId` until it settles. */
+  private holdTransition(strandId: string, transition: RuntimeTransition): void {
+    this.transitions.set(strandId, transition);
+    const release = (): void => {
+      if (this.transitions.get(strandId) === transition) {
+        this.transitions.delete(strandId);
+      }
+    };
+    // The rejection itself reaches the caller through the operation returned to it.
+    void transition.operation.then(release, release);
+  }
+
+  /** Wait for the transition requested before this one. Never rejects: its failure belongs to its caller. */
+  private async settleTransition(strandId: string, previous: RuntimeTransition | undefined): Promise<void> {
+    if (!previous) {
+      return;
+    }
+    await (previous.operation as Promise<unknown>).catch((error: unknown) => {
+      log('Strand %s: the %s queued ahead failed (reported by its caller): %o', strandId, previous.kind, error);
+    });
+  }
+
+  /** Body of {@link resumeStrand}; at most one runs per strand at a time. */
+  private async runResume(
+    strandId: string,
+    previous: RuntimeTransition | undefined,
+    overrides?: ResumeStrandOverrides
+  ): Promise<StrandInstance> {
+    // Before any "already live" test: a quiesce ahead still has the handles attached, and
+    // mid-build (a launch still building) the libp2p node is attached while the database is
+    // not. A failed launch has already dropped the instance, and the check below reports it
+    // untracked.
+    await this.settleTransition(strandId, previous);
+    await this.settleRuntimeBuilds(strandId);
+    const instance = this.instances.get(strandId);
+    if (!instance) {
+      throw new Error(`Cannot resume strand ${strandId}: not tracked`);
+    }
+    const launchConfig = this.launchConfigs.get(strandId);
+    if (!launchConfig) {
+      throw new Error(`Cannot resume strand ${strandId}: no retained launch config`);
+    }
+    if (instance.libp2pNode || instance.database) {
+      log('resumeStrand: strand %s already live', strandId);
+      return instance;
+    }
+
+    log('Resuming strand instance: %s', strandId);
+    const tTotal = performance.now();
+
+    // Re-apply volatile inputs and persist them so a subsequent resume reuses them.
+    // Each `??` matters: a resume that passes no override must keep the retained value,
+    // and a resume that passes one must leave it retained for the next resume — otherwise
+    // a later no-override wake silently reverts to whatever launch time saw.
+    const resumeConfig: StartStrandConfig = {
+      ...launchConfig,
+      bootstrapNodes: overrides?.bootstrapNodes ?? launchConfig.bootstrapNodes,
+      // NOTE: `??` retains, so an override can raise or lower the count but cannot CLEAR it
+      // back to "this node no longer knows" — the direction that would declare nothing. Moot
+      // while nothing feeds `servingMachines` at all; if a source lands
+      // (`backlog/feat-strand-yardstick-from-serving-machines`) that can legitimately lose the
+      // count — a strand whose member rows became unreadable — this merge must gain an explicit
+      // clear rather than silently declaring a stale number over a serving set it can no longer
+      // see. Same applies to `bootstrapNodes` above, where a stale seed is harmless.
+      servingMachines: overrides?.servingMachines ?? launchConfig.servingMachines
+    };
+    this.launchConfigs.set(strandId, resumeConfig);
+
+    instance.status = 'starting';
+    try {
+      await this.trackRuntimeBuild(strandId, this.buildStrandRuntime(instance, resumeConfig));
+      timing('[resumeStrand:%s] total: %dms', strandId, Math.round(performance.now() - tTotal));
+      log('Strand %s resumed successfully', strandId);
+      return instance;
+    } catch (error) {
+      instance.status = 'error';
+      instance.error = error instanceof Error ? error.message : String(error);
+      log('Failed to resume strand %s: %s', strandId, instance.error);
+      throw error;
+    }
+  }
+
+  /**
+   * Honor a founder request against an ALREADY-TRACKED strand — the seam that
+   * closes the "whoever launches first decides whether the bootstrap runs" gap:
+   * an instance first launched as a joiner (an app's own attach, or a watcher
+   * poll winning the launch race) used to swallow a later founder request
+   * silently, leaving the strand active with no `Strand.Header`.
+   *
+   * Flips the RETAINED launch config's `founder` to true, so every later
+   * quiesce → resume rebuild founds as well (the bootstrap is insert-if-absent —
+   * {@link StrandDatabase.ensureFounderBootstrap} — so re-running it per rebuild
+   * writes nothing twice), and runs the bootstrap against the live database now.
+   *
+   * @returns how the request resolved:
+   * - `'already-founder'` — the retained config already founds; nothing to do.
+   * - `'bootstrapped'` — config flipped and the live database ran the bootstrap.
+   * - `'needs-resume'` — config flipped, but the instance is quiesced (no live
+   *   database) or being quiesced (its database is closing), so the bootstrap could
+   *   not run here: the CALLER must wake the
+   *   strand (`CadreNode.wakeStrand`, which owns the hibernation bookkeeping this
+   *   manager does not) so the rebuild — which now founds — runs it.
+   * @param resolvePartyKey - Asked for the party's own membership key when (and only
+   *   when) a CLOSED strand's retained config carries none — the instance was launched
+   *   as a joiner before its `StrandPartyKey` row existed or replicated. The resolved
+   *   key is retained alongside the founder flip so the bootstrap (now, or on the
+   *   caller's wake for `'needs-resume'`) can seat the founding Member/Manager.
+   * @throws when the strand is not tracked — this seam exists only for the
+   *   tracked-instance launch path; an untracked id is a caller bug.
+   * @throws whatever the live bootstrap throws (e.g. `PreSplitStrandIdentityError`),
+   *   after {@link withdrawFounderRequest} — so a retry re-attempts the founding rather
+   *   than resolving `'already-founder'` over an instance that never founded. A caller
+   *   whose `'needs-resume'` wake fails owes the same withdrawal.
+   */
+  async foundExistingStrand(
+    strandId: string,
+    resolvePartyKey?: () => Promise<string | undefined>
+  ): Promise<'already-founder' | 'bootstrapped' | 'needs-resume'> {
+    const instance = this.instances.get(strandId);
+    const config = this.launchConfigs.get(strandId);
+    if (!instance || !config) {
+      throw new Error(`Cannot found strand ${strandId}: not tracked`);
+    }
+    if (config.founder === true) {
+      return 'already-founder';
+    }
+    const partyMemberPrivateKey =
+      config.partyMemberPrivateKey
+        ?? (config.strandRow.Type === 'c' ? await resolvePartyKey?.() : undefined);
+    // A fresh object rather than mutating in place: startStrand retains the CALLER'S
+    // config object, which is not ours to rewrite.
+    this.launchConfigs.set(strandId, { ...config, founder: true, partyMemberPrivateKey });
+    if (this.isQuiescing(strandId) || (!instance.database && !this.firstSyncGates.has(strandId))) {
+      return 'needs-resume';
+    }
+    try {
+      await this.ensureFounderBootstrap(strandId);
+    } catch (error) {
+      this.withdrawFounderRequest(strandId);
+      throw error;
+    }
+    return 'bootstrapped';
+  }
+
+  /**
+   * Give a tracked storage replica ({@link StartStrandConfig.sAppConfig} absent) the app's
+   * schema in place — the seam for an app on this machine claiming a strand this node was
+   * already hosting. No runtime rebuild: the libp2p node, its peer id and connections, the
+   * store and its warm cache all stay, and the new `App` tables read the blocks the replica
+   * already holds. Callers attach BEFORE {@link foundExistingStrand}: the founder bootstrap
+   * writes the sApp into `Strand.Header` and refuses a database that has none.
+   *
+   * A runtime build in flight (the replica's launch, or a hibernation wake) is waited out
+   * first, so the attach acts on the database that build produced. Then, in order: the
+   * schema signature is checked before anything changes; the retained launch config takes
+   * the sApp, so every later rebuild applies it through `composeStrand`; the live database —
+   * published, or still held by the first-sync gate — gets the schema; and only then does
+   * the instance record `sAppInfo` and the sApp's latency hint. An instance with no live
+   * database gets the config alone: a quiesced one applies the schema at its next resume, a
+   * launch waiting for a seed peer at its next database retry.
+   *
+   * @returns `'already-attached'` when the instance already runs an sApp — a DIFFERENT one
+   *   is logged and left as it is, as a second claim of a claimed strand always has been —
+   *   else `'attached'`.
+   * @throws when the strand is not tracked (a caller bug, or a launch that failed while
+   *   this waited on it); when the schema signature is refused, having changed nothing; and
+   *   whatever the live apply throws (e.g. a quiesce closed the database mid-apply) — the
+   *   retained config then still carries the sApp and `sAppInfo` stays unset, so the next
+   *   claim retries the apply.
+   *
+   * NOTE: two claims of one replica racing each other both pass the `sAppInfo` check and
+   * apply concurrently; the loser of the declarative diff can reject, and its `addStrand`
+   * retry then resolves `'already-attached'`. Needs an app calling `addStrand` twice at once
+   * (or beside a watcher retry); if it is ever seen, chain attaches per strand id.
+   */
+  async attachSApp(
+    strandId: string,
+    sAppConfig: SAppConfig,
+    options: { requireSignedSchemas?: boolean } = {}
+  ): Promise<'attached' | 'already-attached'> {
+    await this.settleRuntimeBuilds(strandId);
+    const instance = this.instances.get(strandId);
+    const config = this.launchConfigs.get(strandId);
+    if (!instance || !config) {
+      throw new Error(`Cannot attach an sApp to strand ${strandId}: not tracked`);
+    }
+    if (instance.sAppInfo) {
+      if (instance.sAppInfo.id !== sAppConfig.id) {
+        log('attachSApp: strand %s already runs sApp %s — the claim for sApp %s is ignored',
+          strandId, instance.sAppInfo.id, sAppConfig.id);
+      }
+      return 'already-attached';
+    }
+    const sAppInfo = verifiedSAppInfo(strandId, sAppConfig, options.requireSignedSchemas);
+    // A fresh object rather than mutating in place: startStrand retains the CALLER'S
+    // config object, which is not ours to rewrite.
+    this.launchConfigs.set(strandId, { ...config, sAppConfig });
+    const database = instance.database ?? this.firstSyncGates.get(strandId)?.database;
+    if (database) {
+      await database.attachAppSchema(sAppConfig);
+    } else {
+      log('attachSApp: strand %s has no live database — its next build applies the sApp schema', strandId);
+    }
+    instance.sAppInfo = sAppInfo;
+    if (sAppConfig.latencyHint) {
+      // HibernationManager reads the hint whenever it arms a timer, so the app's hint governs
+      // from the next one: a timer already armed runs once at the old duration, and a replica
+      // launched under a realtime default was never tracked, so it stays up until relaunched.
+      instance.latencyHint = sAppConfig.latencyHint;
+    }
+    log('Strand %s upgraded from storage replica to %s', strandId, describeSApp(sAppConfig));
+    return 'attached';
+  }
+
+  /**
+   * Whether `strandId` is launched but still waiting for its first sync — a joiner whose
+   * runtime is up while its database is withheld (status `'syncing'`, or `'idle'` after
+   * the hibernation manager's idle timer fired on it), including one whose database build
+   * waits for a seed peer to answer ({@link deferUnreachableJoin}). `false` for a writable,
+   * quiesced, or untracked strand. The predicate {@link whenWritable}'s callers gate on.
+   */
+  isAwaitingFirstSync(strandId: string): boolean {
+    const instance = this.instances.get(strandId);
+    return instance !== undefined && isAwaitingFirstSync(instance);
+  }
+
+  /**
+   * Resolve once the strand's database is published to the app — immediately for a
+   * writable strand — or reject with {@link StrandAwaitingFirstSyncError} after
+   * `timeoutMs` (default: the retained launch config's `firstSync.timeoutMs`, else
+   * {@link DEFAULT_STRAND_FIRST_SYNC_TIMEOUT_MS}). The rejection is RETRYABLE: nothing is
+   * torn down, the gate keeps probing, and a later call waits afresh. A quiesced
+   * (hibernating) strand is not woken here — the wait spans a resume, so a caller that
+   * wants one wakes it (`CadreNode.wakeStrand`) and waits. Rejects immediately for an
+   * untracked strand, and whenever the strand is stopped while a wait is pending.
+   */
+  whenWritable(strandId: string, options?: { timeoutMs?: number }): Promise<StrandInstance> {
+    const instance = this.instances.get(strandId);
+    if (!instance) {
+      return Promise.reject(new Error(`Cannot wait for strand ${strandId} to become writable: not tracked`));
+    }
+    if (instance.database) {
+      return Promise.resolve(instance);
+    }
+    const timeoutMs = options?.timeoutMs
+      ?? this.launchConfigs.get(strandId)?.firstSync?.timeoutMs
+      ?? DEFAULT_STRAND_FIRST_SYNC_TIMEOUT_MS;
+    const startedAt = Date.now();
+    return new Promise<StrandInstance>((resolve, reject) => {
+      const waiters = this.writableWaiters.get(strandId) ?? new Set<WritableWaiter>();
+      this.writableWaiters.set(strandId, waiters);
+      const settle = (): void => {
+        clearTimeout(timer);
+        waiters.delete(waiter);
+        if (waiters.size === 0) {
+          this.writableWaiters.delete(strandId);
+        }
+      };
+      const waiter: WritableWaiter = {
+        resolve: () => { settle(); resolve(instance); },
+        reject: (error) => { settle(); reject(error); }
+      };
+      const timer = setTimeout(
+        () => waiter.reject(new StrandAwaitingFirstSyncError(strandId, Date.now() - startedAt)),
+        timeoutMs);
+      (timer as { unref?: () => void }).unref?.();
+      waiters.add(waiter);
+    });
+  }
+
+  /**
+   * The tracked instance once no runtime build of it is in flight, or `undefined` when none
+   * is tracked by then — never launched, or the launch waited on failed and dropped its
+   * record. For a caller about to hand a tracked instance to an app: until its build settles
+   * the instance is `'starting'` with no database, and a failed launch leaves no instance.
+   * Never rejects; a failed build is reported by the call that started it.
+   */
+  async whenRuntimeBuilt(strandId: string): Promise<StrandInstance | undefined> {
+    await this.settleRuntimeBuilds(strandId);
+    return this.instances.get(strandId);
+  }
+
+  /**
+   * Hand a launch's database to the app: set `instance.database`, retire the gate that
+   * held it, release every {@link whenWritable} waiter, and — when the launch had already
+   * been reported gated — flip it `'active'` and announce it through the retained config's
+   * `onWritable`. "Reported gated" is any status but `'starting'`: `'syncing'`, or the
+   * `'idle'` the hibernation manager's idle timer flips a long-gated joiner to (see
+   * {@link isAwaitingFirstSync}) — the Header arriving after that flip must announce too,
+   * or an app hanging on `strand:writable` never learns the strand opened. A publish that
+   * lands while the instance is still `'starting'` (the Header arrived during the rest of
+   * bring-up) announces nothing: `startStrand` / `resumeStrand` are about to report the
+   * strand `'active'`, and `strand:writable` is defined as the follow-up to a gated
+   * launch, never a duplicate of `strand:started`.
+   *
+   * It is also the one seam where a WITHHELD database becomes available, so it kicks the
+   * membership reconciler: on a gated launch that loop's own immediate pass ran while
+   * `instance.database` was still unset, found no database, and would otherwise not try
+   * again until its retry timer fired. A launch that was never gated publishes here BEFORE
+   * the reconciler is constructed, so there is nothing registered to kick and the loop's own
+   * immediate pass already sees the database — which is why the kick needs no `wasGated`
+   * test of its own.
+   */
+  private publishDatabase(instance: StrandInstance, database: StrandDatabase): void {
+    const strandId = instance.strandId;
+    const gate = this.firstSyncGates.get(strandId);
+    if (gate) {
+      gate.open();
+      this.firstSyncGates.delete(strandId);
+    }
+    instance.database = database;
+    const wasGated = instance.status !== 'starting';
+    if (wasGated) {
+      // A Header delivered by a peer is strand activity; a publish during bring-up is not
+      // (see the build's end in `buildStrandRuntime`).
+      instance.lastActivity = new Date();
+      instance.status = 'active';
+      log('Strand %s is now writable: Strand.Header held', strandId);
+    }
+    const waiters = this.writableWaiters.get(strandId);
+    if (waiters) {
+      this.writableWaiters.delete(strandId);
+      waiters.forEach((waiter) => waiter.resolve());
+    }
+    if (wasGated) {
+      this.launchConfigs.get(strandId)?.onWritable?.(strandId);
+    }
+    // Not awaited: publishing must stay synchronous, and the pass chains on the loop's own
+    // tail, so it cannot overlap a pass already in flight.
+    void this.membershipReconcilers.get(strandId)?.reconcile();
+  }
+
+  /** Reject every {@link whenWritable} waiter for a strand that is going away. */
+  private rejectWritableWaiters(strandId: string, reason: string): void {
+    const waiters = this.writableWaiters.get(strandId);
+    if (!waiters) {
+      return;
+    }
+    this.writableWaiters.delete(strandId);
+    waiters.forEach((waiter) => waiter.reject(new Error(`Strand ${strandId} ${reason} before becoming writable`)));
+  }
+
+  /**
+   * Undo {@link foundExistingStrand}'s founder flip after the founding it promised
+   * failed, so the retained config again says what this instance actually runs as (a
+   * joiner) and the next founder request re-attempts the bootstrap. The resolved party
+   * key stays retained — it is this party's identity for the strand either way. No-op
+   * when the strand is untracked or its config does not found.
+   */
+  withdrawFounderRequest(strandId: string): void {
+    const config = this.launchConfigs.get(strandId);
+    if (config?.founder === true) {
+      this.launchConfigs.set(strandId, { ...config, founder: false });
+    }
+  }
+
+  /**
+   * Run the (idempotent) founder bootstrap against a tracked strand's LIVE
+   * database, independently of what the retained launch config says.
+   *
+   * Separate from {@link foundExistingStrand} because a caller that resolved
+   * `'needs-resume'` and woke the strand must not assume the wake's own rebuild
+   * founded it: `HibernationManager` COALESCES wakes, so a wake already in flight
+   * when the config flipped had already read the PRE-flip config and rebuilt as a
+   * joiner. Re-running the bootstrap costs one insert-if-absent probe per table
+   * and is the only thing that makes "founding resolves once the Header is
+   * written" true on that path.
+   *
+   * A strand still `'syncing'` (its database withheld behind the first-sync gate) is
+   * founded through the gate: the bootstrap writes the Header this machine was waiting
+   * to receive, so a successful run publishes the database and the strand goes
+   * `'active'` — the one legitimate way a joiner launch becomes writable without a peer.
+   *
+   * @throws when the strand is not tracked, or is still quiesced (no live
+   *   database) — both mean the bootstrap did NOT run, which a founder request
+   *   must never swallow.
+   */
+  async ensureFounderBootstrap(strandId: string): Promise<void> {
+    const instance = this.instances.get(strandId);
+    if (!instance) {
+      throw new Error(`Cannot run the founder bootstrap for strand ${strandId}: not tracked`);
+    }
+    const gate = this.firstSyncGates.get(strandId);
+    const database = instance.database ?? gate?.database;
+    if (!database) {
+      throw new Error(
+        `Cannot run the founder bootstrap for strand ${strandId}: it has no live database to ` +
+        'write to — it is quiesced (resume it first), or its database build is waiting for a ' +
+        'seed peer to answer.'
+      );
+    }
+    // Forward the RETAINED config's party key: a foundExistingStrand that flipped a
+    // joiner launch to founder may have resolved a key the live database's captured
+    // config (built at the original launch) never saw.
+    await database.ensureFounderBootstrap(this.launchConfigs.get(strandId)?.partyMemberPrivateKey);
+    if (!instance.database) {
+      this.publishDatabase(instance, database);
+    }
+  }
+
+  /**
+   * Stop a strand instance
+   */
+  async stopStrand(strandId: string): Promise<void> {
+    const instance = this.instances.get(strandId);
+    if (!instance) {
+      log('Strand %s not found', strandId);
+      return;
+    }
+
+    log('Stopping strand instance: %s', strandId);
+    instance.status = 'stopping';
+    this.rejectWritableWaiters(strandId, 'was stopped');
+
+    try {
+      await this.releaseRuntime(instance);
+      instance.status = 'stopped';
+      this.instances.delete(strandId);
+      this.launchConfigs.delete(strandId);
+      // Storage is released only here — NOT in releaseRuntime, which a quiesce shares.
+      await this.disposeStrandStorage(strandId);
+      log('Strand %s stopped successfully', strandId);
+    } catch (error) {
+      instance.status = 'error';
+      instance.error = error instanceof Error ? error.message : String(error);
+      log('Error stopping strand %s: %s', strandId, instance.error);
+      throw error;
+    }
+  }
+
+  /**
+   * Stop all strand instances
+   */
+  async stopAll(): Promise<void> {
+    this.stopping = true;
+    log('Stopping all strand instances (%d)', this.instances.size);
+    
+    const stopPromises = Array.from(this.instances.keys()).map(id => 
+      this.stopStrand(id).catch(err => {
+        log('Error stopping strand %s during shutdown: %s', id, err);
+      })
+    );
+    
+    await Promise.all(stopPromises);
+    this.stopping = false;
+    log('All strand instances stopped');
+  }
+}
+