mixdog 0.7.17 → 0.8.0

package/src/runtime/memory/lib/pg/supervisor.mjs

@@ -0,0 +1,495 @@
+const __mixdogMemoryStderrWrite = process.stderr.write.bind(process.stderr);
+function __mixdogMemoryLog(...args) {
+  if (process.env.MIXDOG_QUIET_MEMORY_LOG) return true;
+  return __mixdogMemoryStderrWrite(...args);
+}
+
+/**
+ * supervisor-pg — PG child process lifecycle wired into the mixdog supervisor.
+ *
+ * Public API:
+ *   ensurePgInstance(dataDir) → Promise<{ host, port, runtimeDir, pgdataDir }>
+ *   stopPgForShutdown()       → Promise<void>  — call from server-main.mjs shutdown()
+ *
+ * Depends on Track A's pg-process.mjs for startPg / stopPg / healthcheckPg.
+ * Lazy-imported so this module loads cleanly before Track A lands.
+ *
+ * active-instance.json additions:
+ *   pg_port?        number  — TCP port PG is listening on
+ *   pg_started_at?  number  — epoch ms when PG was last started
+ *   pg_pgdata?      string  — absolute path to the pgdata directory
+ *   pg_runtime_dir? string  — absolute path to the pg runtime binaries dir
+ */
+
+import { createServer }                       from 'node:net';
+import {
+  unlinkSync, readFileSync, writeFileSync,
+  renameSync, statSync, mkdirSync,
+}                                             from 'node:fs';
+import { join, resolve }                      from 'node:path';
+import { tmpdir }                             from 'node:os';
+import { updateJsonAtomicSync }               from '../../../shared/atomic-file.mjs';
+
+// ── pg-process interface (Track A) ───────────────────────────────────────────
+// Dynamic import so this module loads even before Track A's file exists.
+let _pgProc = null;
+async function _getPgProc() {
+  if (_pgProc) return _pgProc;
+  // import.meta.url is in src/memory/lib/pg/ — process.mjs lives alongside.
+  const mod = await import('./process.mjs');
+  _pgProc = {
+    startPg:          mod.startPg,
+    stopPg:           mod.stopPg,
+    healthcheckPg:    mod.healthcheckPg,
+    reconcileConfV2:  mod.reconcileConfV2,
+  };
+  return _pgProc;
+}
+
+// ── In-process state ─────────────────────────────────────────────────────────
+/** @type {{ port: number, pgdata: string, runtimeDir: string, proc: unknown } | null} */
+let _live = null;
+/** Dedup: one ensure coroutine at a time. */
+let _ensureInFlight = null;
+/** Per-process flag — postgresql.conf reconcile attempted once per supervisor. */
+let _v2ReconcileTried = false;
+
+// ── Constants ────────────────────────────────────────────────────────────────
+const PG_PORT_MIN      = 55432;
+const PG_PORT_MAX      = 55632;
+const PG_LOG_MAX_BYTES = 10 * 1024 * 1024; // 10 MB
+const SPAWN_LOCK_NAME  = 'pg-spawn.lock';
+const LOCK_WAIT_MS     = 30_000;
+const LOCK_POLL_MS     = 100;
+const LOCK_WARN_MS     = 5_000;
+const LOCK_WAIT_CODES  = new Set(['EEXIST', 'EPERM', 'EACCES', 'EBUSY']);
+
+// ── File lock (O_EXCL pattern from dispatch-persist.mjs / run-mcp.mjs) ───────
+
+/**
+ * Acquire `${dataDir}/pg-spawn.lock`.
+ * Returns the lock-file path on success, or null on timeout (caller proceeds
+ * best-effort — worst case: two supervisors race initdb on the same pgdata,
+ * but the second will fail pg_ctl start gracefully and healthcheck will
+ * converge on the first's port).
+ */
+async function acquireSpawnLock(dataDir, blockedProbe = null) {
+  const lp       = join(dataDir, SPAWN_LOCK_NAME);
+  const deadline = Date.now() + LOCK_WAIT_MS;
+  const body     = JSON.stringify({ pid: process.pid, startedAt: Date.now() });
+  let warned = false;
+  for (;;) {
+    try {
+      // 'wx' = O_CREAT | O_EXCL — atomic, fails with EEXIST if held.
+      writeFileSync(lp, body, { flag: 'wx' });
+      return { lockPath: lp, reuse: null };
+    } catch (err) {
+      if (!LOCK_WAIT_CODES.has(err?.code)) {
+        __mixdogMemoryLog(`[supervisor-pg] spawn lock error: ${err?.code || err?.message}\n`);
+        return { lockPath: null, reuse: null };
+      }
+      if (blockedProbe) {
+        try {
+          const reuse = await blockedProbe();
+          if (reuse) return { lockPath: null, reuse };
+        } catch {}
+      }
+      // Dead-holder detection: same as run-mcp.mjs acquireLock().
+      try {
+        const holder = JSON.parse(readFileSync(lp, 'utf8'));
+        if (holder?.pid) {
+          try { process.kill(holder.pid, 0); }
+          catch (ke) {
+            if (ke.code === 'ESRCH') {
+              try { unlinkSync(lp); } catch {}
+              continue; // retry immediately after removing stale lock
+            }
+          }
+        }
+        const ageMs = Date.now() - Number(holder?.startedAt || 0);
+        if (!warned && ageMs >= LOCK_WARN_MS) {
+          warned = true;
+          __mixdogMemoryLog(`[supervisor-pg] waiting for spawn lock holder pid=${holder?.pid || 'unknown'} ageMs=${ageMs}\n`);
+        }
+      } catch { /* unreadable — fall through and wait */ }
+      if (Date.now() >= deadline) {
+        if (blockedProbe) {
+          try {
+            const reuse = await blockedProbe();
+            if (reuse) return { lockPath: null, reuse };
+          } catch {}
+        }
+        throw new Error(`[supervisor-pg] spawn-lock acquire timeout (concurrent supervisor risk)`);
+      }
+      await new Promise(r => setTimeout(r, LOCK_POLL_MS));
+    }
+  }
+}
+
+function releaseSpawnLock(lp) {
+  if (!lp) return;
+  try { unlinkSync(lp); } catch {}
+}
+
+// ── Port allocation ──────────────────────────────────────────────────────────
+
+function _probePort(port) {
+  return new Promise(resolve => {
+    const srv = createServer();
+    srv.once('error', () => resolve(false));
+    srv.once('listening', () => srv.close(() => resolve(true)));
+    srv.listen(port, '127.0.0.1');
+  });
+}
+
+async function allocatePort() {
+  for (let p = PG_PORT_MIN; p <= PG_PORT_MAX; p++) {
+    if (await _probePort(p)) return p;
+  }
+  throw new Error(`[supervisor-pg] no free TCP port in range ${PG_PORT_MIN}–${PG_PORT_MAX}`);
+}
+
+// ── Log rotation ─────────────────────────────────────────────────────────────
+// Keep at most 1 archive (pg.log.1). Rotate at PG_LOG_MAX_BYTES.
+
+function rotateLogIfNeeded(logPath) {
+  try {
+    const st = statSync(logPath);
+    if (st.size > PG_LOG_MAX_BYTES) {
+      const archive = logPath + '.1';
+      try { unlinkSync(archive); } catch {}
+      renameSync(logPath, archive);
+    }
+  } catch { /* log does not exist yet — nothing to rotate */ }
+}
+
+// ── active-instance.json patch ───────────────────────────────────────────────
+// Atomic read-modify-write using the same tmp+rename pattern as server.mjs.
+// Lives in MIXDOG_RUNTIME_ROOT or os.tmpdir()/mixdog/
+// (see src/channels/lib/runtime-paths.mjs).
+
+const _RUNTIME_ROOT = process.env.MIXDOG_RUNTIME_ROOT
+  ? resolve(process.env.MIXDOG_RUNTIME_ROOT)
+  : join(tmpdir(), 'mixdog');
+const _ACTIVE_FILE = join(_RUNTIME_ROOT, 'active-instance.json');
+
+function patchActiveInstance(fields) {
+  try {
+    updateJsonAtomicSync(_ACTIVE_FILE, (curRaw) => {
+      // Drop stale fields (pid/startedAt) written by older server versions.
+      const { pid: _legacyPid, startedAt: _legacyStartedAt, ...cur } = curRaw ?? {};
+      // Omit null-valued fields (clean removal when pg is stopped).
+      const merged = { ...cur, updatedAt: Date.now() };
+      for (const [k, v] of Object.entries(fields)) {
+        if (v == null) delete merged[k];
+        else merged[k] = v;
+      }
+      return merged;
+    }, { compact: true, fsyncDir: true });
+  } catch (e) {
+    __mixdogMemoryLog(`[supervisor-pg] patchActiveInstance failed: ${e?.message}\n`);
+  }
+}
+
+// ── postmaster.pid helpers ───────────────────────────────────────────────────
+
+function readPostmasterPid(pgdata) {
+  return readPostmasterInfo(pgdata).pid;
+}
+
+function readPostmasterInfo(pgdata) {
+  try {
+    const raw = readFileSync(join(pgdata, 'postmaster.pid'), 'utf8');
+    const lines = raw.split('\n');
+    const pid = parseInt(lines[0], 10);
+    const port = parseInt(lines[3], 10);
+    return {
+      pid: Number.isFinite(pid) && pid > 0 ? pid : null,
+      port: Number.isFinite(port) && port > 0 ? port : null,
+    };
+  } catch { return { pid: null, port: null }; }
+}
+
+function isPidAlive(pid) {
+  try { process.kill(pid, 0); return true; }
+  catch (e) {
+    // EPERM: process exists but we lack permission → alive.
+    // ESRCH: no such process → dead.
+    return e.code === 'EPERM';
+  }
+}
+
+/**
+ * Best-effort process-name check: confirms the pid's comm name contains 'postgres'.
+ * Falls back to true (alive) when the name cannot be determined.
+ */
+async function isPostgresPid(pid) {
+  try {
+    if (process.platform === 'linux') {
+      const { readFileSync: rfs } = await import('node:fs');
+      const comm = rfs(`/proc/${pid}/comm`, 'utf8').trim();
+      return comm.includes('postgres');
+    }
+    if (process.platform === 'darwin') {
+      const { spawnSync } = await import('node:child_process');
+      const r = spawnSync('ps', ['-o', 'comm=', '-p', String(pid)], { encoding: 'utf8', windowsHide: true });
+      if (r.status === 0) return (r.stdout || '').trim().includes('postgres');
+      return true;
+    }
+    if (process.platform === 'win32') {
+      const { spawnSync } = await import('node:child_process');
+      const r = spawnSync('tasklist', ['/FI', `PID eq ${pid}`, '/FO', 'CSV', '/NH'], { encoding: 'utf8', windowsHide: true });
+      if (r.status === 0) return (r.stdout || '').toLowerCase().includes('postgres');
+      return true;
+    }
+  } catch { /* cannot read — fall back to alive */ }
+  return true;
+}
+
+async function isPostmasterAlive(pid) {
+  if (!isPidAlive(pid)) return false;
+  return isPostgresPid(pid);
+}
+
+// ── Internal: spawn a fresh PG instance ─────────────────────────────────────
+
+async function _startFresh(dataDir, pgdata, port, runtimeDir) {
+  const { startPg } = await _getPgProc();
+  const logPath = join(dataDir, 'pg.log');
+  rotateLogIfNeeded(logPath);
+  mkdirSync(pgdata, { recursive: true });
+  // Track A's startPg handles initdb (if needed) then pg_ctl start.
+  // stdout/stderr are directed to logPath by Track A's implementation.
+  const pgdataDir = pgdata;
+  const proc = await startPg({ runtimeDir, pgdataDir, port, logPath });
+  const actualPort = proc?.port ?? port;
+  _live = { port: actualPort, pgdata, runtimeDir, proc };
+  patchActiveInstance({
+    pg_port: actualPort, pg_started_at: Date.now(),
+    pg_pgdata: pgdata, pg_runtime_dir: runtimeDir,
+  });
+  __mixdogMemoryLog(`[supervisor-pg] ${proc?.attached ? 'attached to' : 'started'} PG port=${actualPort} pgdata=${pgdata}\n`);
+  return { host: '127.0.0.1', port: actualPort, runtimeDir, pgdataDir };
+}
+
+async function tryReusePgInstance({ pgdata, runtimeDir, healthcheckPg, source = 'reuse' }) {
+  let ai = null;
+  let existingPort = null;
+  let existingRtDir = runtimeDir;
+  try {
+    ai = JSON.parse(readFileSync(_ACTIVE_FILE, 'utf8'));
+    if (ai?.pg_port && ai?.pg_pgdata && resolve(ai.pg_pgdata) === resolve(pgdata)) {
+      existingPort = ai.pg_port;
+      existingRtDir = ai?.pg_runtime_dir ?? runtimeDir;
+    }
+  } catch {}
+
+  if (existingPort) {
+    try {
+      if (await healthcheckPg({ port: existingPort })) {
+        __mixdogMemoryLog(`[supervisor-pg] reusing PG on port ${existingPort} (${source}:active-instance)\n`);
+        _live = { port: existingPort, pgdata, runtimeDir: existingRtDir, proc: null };
+        return { host: '127.0.0.1', port: existingPort, runtimeDir: existingRtDir, pgdataDir: pgdata };
+      }
+    } catch {}
+  }
+
+  const pm = readPostmasterInfo(pgdata);
+  if (pm.pid && pm.port && await isPostmasterAlive(pm.pid)) {
+    try {
+      if (await healthcheckPg({ port: pm.port })) {
+        __mixdogMemoryLog(`[supervisor-pg] attaching to PG pid=${pm.pid} port=${pm.port} (${source}:postmaster.pid)\n`);
+        _live = { port: pm.port, pgdata, runtimeDir, proc: null };
+        patchActiveInstance({
+          pg_port: pm.port,
+          pg_started_at: ai?.pg_started_at ?? Date.now(),
+          pg_pgdata: pgdata,
+          pg_runtime_dir: runtimeDir,
+        });
+        return { host: '127.0.0.1', port: pm.port, runtimeDir, pgdataDir: pgdata };
+      }
+    } catch {}
+  }
+
+  return null;
+}
+
+// ── Internal: full ensure logic (runs exclusively via _ensureInFlight) ────────
+
+async function _doEnsure(dataDir) {
+  const { healthcheckPg, stopPg } = await _getPgProc();
+  const pgdata = join(dataDir, 'pgdata');
+
+  // Resolve runtimeDir via runtime-fetcher (cache-hits immediately when already downloaded).
+  const { ensureRuntime } = await import('../runtime-fetcher.mjs');
+  const { runtimeDir } = await ensureRuntime(dataDir);
+
+  // One-shot v2 conf reconcile — idempotent. Covers attach paths (in-process
+  // fast path / cross-process active-instance reuse) where startPg is never
+  // re-invoked on already-running PG instances.
+  if (!_v2ReconcileTried) {
+    _v2ReconcileTried = true;
+    try {
+      const { reconcileConfV2 } = await _getPgProc();
+      reconcileConfV2(runtimeDir, pgdata);
+    } catch (e) {
+      __mixdogMemoryLog(`[supervisor-pg] reconcileConfV2 error (non-fatal): ${e?.message}\n`);
+    }
+  }
+
+  // ── Fast path: already live in this process ──────────────────────────────
+  if (_live) {
+    try {
+      if (await healthcheckPg({ port: _live.port })) {
+        return { host: '127.0.0.1', port: _live.port, runtimeDir: _live.runtimeDir, pgdataDir: _live.pgdata };
+      }
+    } catch {}
+    // healthcheck failed — fall through to recovery under lock
+    __mixdogMemoryLog(`[supervisor-pg] in-process PG failed healthcheck — recovering\n`);
+    _live = null;
+  }
+
+  const prelockReuse = await tryReusePgInstance({
+    pgdata,
+    runtimeDir,
+    healthcheckPg,
+    source: 'prelock',
+  });
+  if (prelockReuse) return prelockReuse;
+
+  // ── Acquire spawn lock to serialize initdb races ─────────────────────────
+  const acquired = await acquireSpawnLock(dataDir, () => tryReusePgInstance({
+    pgdata,
+    runtimeDir,
+    healthcheckPg,
+    source: 'lock-wait',
+  }));
+  if (acquired?.reuse) return acquired.reuse;
+  const lp = acquired?.lockPath ?? null;
+  try {
+    // ── Reuse path: another supervisor already started PG ─────────────────
+    const lockedReuse = await tryReusePgInstance({
+      pgdata,
+      runtimeDir,
+      healthcheckPg,
+      source: 'locked',
+    });
+    if (lockedReuse) return lockedReuse;
+
+    let existingPort = null;
+    let ai = null;
+    try {
+      ai = JSON.parse(readFileSync(_ACTIVE_FILE, 'utf8'));
+      // Only reuse a recorded instance when it was started for THIS pgdata.
+      // active-instance.json is process-global; without matching pg_pgdata a
+      // healthy PG serving a different data directory would be reused, binding
+      // this dataDir's memory to the wrong cluster. A missing/mismatched
+      // pg_pgdata falls through to a fresh start for the requested pgdata.
+      if (ai?.pg_port && ai?.pg_pgdata && resolve(ai.pg_pgdata) === resolve(pgdata)) {
+        existingPort = ai.pg_port;
+      }
+    } catch {}
+
+    if (existingPort) {
+      try {
+        if (await healthcheckPg({ port: existingPort })) {
+          __mixdogMemoryLog(`[supervisor-pg] reusing PG on port ${existingPort}\n`);
+          const existingRtDir = ai?.pg_runtime_dir ?? runtimeDir;
+          _live = { port: existingPort, pgdata, runtimeDir: existingRtDir, proc: null };
+          return { host: '127.0.0.1', port: existingPort, runtimeDir: existingRtDir, pgdataDir: pgdata };
+        }
+      } catch {}
+
+      // ── Stale detection: pg_port recorded but healthcheck failing ─────────
+      __mixdogMemoryLog(
+        `[supervisor-pg] pg_port=${existingPort} recorded but healthcheck failed — recovering\n`,
+      );
+      const pmPid = readPostmasterPid(pgdata);
+      if (pmPid && await isPostmasterAlive(pmPid)) {
+        // postmaster alive but unhealthy: attempt graceful stop first
+        __mixdogMemoryLog(`[supervisor-pg] postmaster PID ${pmPid} alive — attempting graceful stopPg\n`);
+        try { await stopPg({ runtimeDir, pgdataDir: pgdata }); } catch (e) {
+          __mixdogMemoryLog(`[supervisor-pg] graceful stopPg failed: ${e?.message} — continuing to fresh start\n`);
+        }
+      } else if (pmPid) {
+        // postmaster dead: remove stale postmaster.pid so initdb/start is not blocked
+        __mixdogMemoryLog(`[supervisor-pg] postmaster PID ${pmPid} dead — removing stale postmaster.pid\n`);
+        try { unlinkSync(join(pgdata, 'postmaster.pid')); } catch {}
+      }
+      // Clear stale pg fields before restart
+      patchActiveInstance({ pg_port: null, pg_started_at: null, pg_pgdata: null });
+    }
+
+    // ── Allocate a fresh port and spawn ───────────────────────────────────
+    const port = await allocatePort();
+    return await _startFresh(dataDir, pgdata, port, runtimeDir);
+  } finally {
+    releaseSpawnLock(lp);
+  }
+}
+
+// ── Public API ───────────────────────────────────────────────────────────────
+
+/**
+ * Ensure a live PG instance exists for `dataDir`.
+ * Lazy: does nothing until first call — sessions that never touch memory pay
+ * zero cost.
+ *
+ * Concurrent calls are deduplicated — exactly one spawn/recover sequence runs
+ * at a time; additional callers await the same promise.
+ *
+ * @param {string} dataDir  Mixdog data directory.
+ * @returns {Promise<{ host: string, port: number, runtimeDir: string, pgdataDir: string }>}
+ */
+export function ensurePgInstance(dataDir) {
+  if (!_ensureInFlight) {
+    _ensureInFlight = _doEnsure(dataDir).finally(() => { _ensureInFlight = null; });
+  }
+  return _ensureInFlight;
+}
+
+/**
+ * Graceful PG shutdown — call from server-main.mjs shutdown() after workers
+ * have stopped.  Sends pg_ctl stop to the pgdata directory; clears pg_port
+ * from active-instance.json.
+ *
+ * On the supervisor detached-killer path (run-mcp.mjs killChild → SIGTERM
+ * with no time for graceful stop) this function is never called — that is
+ * intentional.  The next ensurePgInstance call will detect the stale
+ * postmaster.pid via isPidAlive() and recover automatically.
+ */
+export async function stopPgForShutdown() {
+  if (!_live) {
+    // _live may be null if PG was started by another process or adapter call.
+    // Attempt graceful stop via active-instance.json.
+    let ai = null;
+    try { ai = JSON.parse(readFileSync(_ACTIVE_FILE, 'utf8')); } catch {}
+    if (!ai?.pg_port || !ai?.pg_pgdata) return;
+    const pgdataDir2  = ai.pg_pgdata;
+    const runtimeDir2 = ai.pg_runtime_dir;
+    if (!runtimeDir2) {
+      __mixdogMemoryLog(`[supervisor-pg] stopPgForShutdown: pg_runtime_dir missing from active-instance.json — skipping\n`);
+      return;
+    }
+    try {
+      const { stopPg } = await _getPgProc();
+      await stopPg({ runtimeDir: runtimeDir2, pgdataDir: pgdataDir2 });
+      __mixdogMemoryLog(`[supervisor-pg] PG stopped (via active-instance.json) on shutdown\n`);
+    } catch (e) {
+      __mixdogMemoryLog(`[supervisor-pg] stopPg error on shutdown (no _live): ${e?.message}\n`);
+    }
+    patchActiveInstance({ pg_port: null, pg_started_at: null, pg_pgdata: null, pg_runtime_dir: null });
+    return;
+  }
+  const snap = _live;
+  _live = null;
+  _ensureInFlight = null;
+  try {
+    const { stopPg } = await _getPgProc();
+    await stopPg({ runtimeDir: snap.runtimeDir, pgdataDir: snap.pgdata });
+    __mixdogMemoryLog(`[supervisor-pg] PG stopped gracefully on shutdown\n`);
+  } catch (e) {
+    __mixdogMemoryLog(`[supervisor-pg] stopPg error on shutdown: ${e?.message}\n`);
+  }
+  patchActiveInstance({ pg_port: null, pg_started_at: null, pg_pgdata: null, pg_runtime_dir: null });
+}