chore(deploy): add plan-only Core incident audit

deployment/device-control-core-incident-audit-v1.json

@@ -0,0 +1,36 @@
+{
+  "schemaVersion": "nodedc.device-plane.device-control-core-incident-audit.v1",
+  "mode": "double-rollback-failed-read-only-audit",
+  "allowedOperation": "canonical-plan-only",
+  "applyAllowed": false,
+  "failedAttempts": [
+    {
+      "patchId": "device-control-core-release-v3-20260822-040",
+      "artifactSha256": "08448a56cdf391076f92c5242e368fd0033b420167874645eebfa1f22184ee92",
+      "backupId": "device-plane-device-control-core-release-v3-20260822-040-20260822-184245"
+    },
+    {
+      "patchId": "device-control-core-release-v3-reconciliation-20260822-042",
+      "artifactSha256": "54ab243439bce724fa0a0872b76cc32e0052ea5127153214d92872f02ae831cf",
+      "backupId": "device-plane-device-control-core-release-v3-reconciliation-20260822-042-20260822-195448"
+    }
+  ],
+  "readOnlyEvidence": [
+    "device-control-core-runtime-inventory",
+    "device-control-core-bounded-container-logs",
+    "device-postgres-schema-presence",
+    "device-postgres-wait-activity"
+  ],
+  "runtimeMutation": "none",
+  "sourceMutation": "none",
+  "databaseMutation": "none",
+  "networkMutation": "none",
+  "secretRead": "none",
+  "preservedServices": [
+    "device-control-core",
+    "device-manager",
+    "device-gateway",
+    "device-postgres",
+    "device-backhaul-target"
+  ]
+}

infra/deploy-runner/build-device-control-core-incident-audit-artifact.mjs

@@ -0,0 +1,92 @@
+#!/usr/bin/env node
+import { createHash } from "node:crypto";
+import { spawnSync } from "node:child_process";
+import { cp, mkdir, mkdtemp, readFile, rm, writeFile } from "node:fs/promises";
+import { tmpdir } from "node:os";
+import { dirname, join, resolve } from "node:path";
+import { fileURLToPath } from "node:url";
+
+const scriptDir = dirname(fileURLToPath(import.meta.url));
+const sourceRoot = resolve(scriptDir, "../..");
+const artifactDir = resolve(
+  process.env.NODEDC_DEPLOY_ARTIFACT_DIR
+    || resolve(scriptDir, "../deploy-artifacts"),
+);
+const [
+  patchId = "device-control-core-incident-audit-20260822-043",
+  ...extra
+] = process.argv.slice(2);
+if (extra.length || !/^[A-Za-z0-9._-]{1,96}$/.test(patchId)) {
+  throw new Error("usage: build-device-control-core-incident-audit-artifact.mjs [patch-id]");
+}
+
+const entry = "deployment/device-control-core-incident-audit-v1.json";
+const stage = await mkdtemp(join(tmpdir(), "nodedc-control-core-incident-audit-"));
+const payload = join(stage, "payload");
+const target = resolve(artifactDir, `nodedc-device-plane-${patchId}.tgz`);
+
+try {
+  const descriptor = JSON.parse(await readFile(resolve(sourceRoot, entry), "utf8"));
+  if (
+    descriptor.schemaVersion
+      !== "nodedc.device-plane.device-control-core-incident-audit.v1"
+    || descriptor.mode !== "double-rollback-failed-read-only-audit"
+    || descriptor.allowedOperation !== "canonical-plan-only"
+    || descriptor.applyAllowed !== false
+    || descriptor.runtimeMutation !== "none"
+    || descriptor.sourceMutation !== "none"
+    || descriptor.databaseMutation !== "none"
+    || descriptor.networkMutation !== "none"
+    || descriptor.secretRead !== "none"
+  ) {
+    throw new Error("device_control_core_incident_audit_descriptor_mismatch");
+  }
+
+  await mkdir(dirname(join(payload, entry)), { recursive: true });
+  await cp(resolve(sourceRoot, entry), join(payload, entry), { force: true });
+  await writeFile(
+    join(stage, "manifest.env"),
+    `id=${patchId}\ncomponent=device-plane\ntype=app-overlay\n`,
+    "utf8",
+  );
+  await writeFile(join(stage, "files.txt"), `${entry}\n`, "utf8");
+  await mkdir(artifactDir, { recursive: true });
+  const tar = spawnSync(
+    "python3",
+    ["-c", canonicalTarScript(), target, stage],
+    { encoding: "utf8", maxBuffer: 16 * 1024 * 1024 },
+  );
+  if (tar.status !== 0) throw new Error(`tar_failed:${tar.stderr || tar.stdout}`);
+  const sha256 = createHash("sha256")
+    .update(await readFile(target))
+    .digest("hex");
+  console.log(JSON.stringify({
+    ok: true,
+    patchId,
+    component: "device-plane",
+    artifact: target,
+    sha256,
+    entries: [entry],
+    build: [],
+    services: [],
+    allowedOperation: descriptor.allowedOperation,
+    runtimeMutation: descriptor.runtimeMutation,
+  }, null, 2));
+} finally {
+  await rm(stage, { recursive: true, force: true });
+}
+
+function canonicalTarScript() {
+  return [
+    "import gzip,io,pathlib,sys,tarfile",
+    "root=pathlib.Path(sys.argv[2])",
+    "with open(sys.argv[1],'wb') as out:",
+    " with gzip.GzipFile(filename='',mode='wb',fileobj=out,compresslevel=9,mtime=0) as gz:",
+    "  with tarfile.open(fileobj=gz,mode='w',format=tarfile.PAX_FORMAT) as tar:",
+    "   for top in ('manifest.env','files.txt','payload'):",
+    "    p=root/top; paths=[p]+(sorted(p.rglob('*')) if p.is_dir() else [])",
+    "    for x in paths:",
+    "     info=tar.gettarinfo(str(x),arcname=x.relative_to(root).as_posix()); info.uid=info.gid=0; info.uname=info.gname='root'; info.mtime=0; info.mode=0o755 if info.isdir() else 0o644",
+    "     with (open(x,'rb') if info.isfile() else io.BytesIO()) as src: tar.addfile(info,src if info.isfile() else None)",
+  ].join("\n");
+}

infra/deploy-runner/test_device_control_core_incident_audit_artifact.py

@@ -0,0 +1,159 @@
+#!/usr/bin/env python3
+import hashlib
+import importlib.machinery
+import importlib.util
+import inspect
+import json
+import os
+import subprocess
+import tarfile
+import tempfile
+import unittest
+from pathlib import Path
+from unittest import mock
+
+
+SCRIPT_DIR = Path(__file__).resolve().parent
+RUNNER_PATH = Path(
+    os.environ.get(
+        "NODEDC_DEPLOY_RUNNER",
+        SCRIPT_DIR.parents[2] / "platform/infra/deploy-runner/nodedc-deploy",
+    )
+)
+BUILDER = SCRIPT_DIR / "build-device-control-core-incident-audit-artifact.mjs"
+
+
+def load_runner():
+    loader = importlib.machinery.SourceFileLoader(
+        "nodedc_control_core_incident_audit_runner_under_test",
+        str(RUNNER_PATH),
+    )
+    spec = importlib.util.spec_from_loader(loader.name, loader)
+    module = importlib.util.module_from_spec(spec)
+    loader.exec_module(module)
+    return module
+
+
+RUNNER = load_runner()
+
+
+class DeviceControlCoreIncidentAuditArtifactTest(unittest.TestCase):
+    def build(self, artifact_dir, patch_id):
+        environment = os.environ.copy()
+        environment["NODEDC_DEPLOY_ARTIFACT_DIR"] = str(artifact_dir)
+        result = subprocess.run(
+            ["node", str(BUILDER), patch_id],
+            check=True,
+            capture_output=True,
+            text=True,
+            env=environment,
+        )
+        return json.loads(result.stdout)
+
+    def test_artifact_is_deterministic_plan_only_audit(self):
+        with tempfile.TemporaryDirectory(
+            prefix="nodedc-control-core-incident-audit-artifact-",
+        ) as directory:
+            artifact_dir = Path(directory)
+            patch_id = "device-control-core-incident-audit-unit-043"
+            first = self.build(artifact_dir, patch_id)
+            first_bytes = Path(first["artifact"]).read_bytes()
+            second = self.build(artifact_dir, patch_id)
+            second_bytes = Path(second["artifact"]).read_bytes()
+
+            self.assertEqual(first_bytes, second_bytes)
+            self.assertEqual(
+                first["sha256"],
+                hashlib.sha256(first_bytes).hexdigest(),
+            )
+            self.assertEqual(first["build"], [])
+            self.assertEqual(first["services"], [])
+            self.assertEqual(first["allowedOperation"], "canonical-plan-only")
+            self.assertEqual(first["runtimeMutation"], "none")
+            with tarfile.open(first["artifact"], "r:gz") as archive:
+                descriptor = json.loads(
+                    archive.extractfile(
+                        "payload/"
+                        + RUNNER.DEVICE_PLANE_CONTROL_CORE_INCIDENT_AUDIT_REL
+                    ).read().decode("utf-8")
+                )
+            self.assertEqual(
+                descriptor,
+                RUNNER.expected_device_plane_control_core_incident_audit_descriptor(),
+            )
+            entries = RUNNER.DEVICE_PLANE_CONTROL_CORE_INCIDENT_AUDIT_ENTRIES
+            self.assertEqual(RUNNER.component_services("device-plane", entries), ())
+            self.assertEqual(RUNNER.component_builds("device-plane", entries), ())
+
+    def test_descriptor_pins_both_rollback_failures(self):
+        descriptor = (
+            RUNNER.expected_device_plane_control_core_incident_audit_descriptor()
+        )
+        self.assertFalse(descriptor["applyAllowed"])
+        self.assertEqual(descriptor["runtimeMutation"], "none")
+        self.assertEqual(
+            [item["artifactSha256"] for item in descriptor["failedAttempts"]],
+            [
+                "08448a56cdf391076f92c5242e368fd0033b420167874645eebfa1f22184ee92",
+                "54ab243439bce724fa0a0872b76cc32e0052ea5127153214d92872f02ae831cf",
+            ],
+        )
+
+    def test_runtime_audit_reads_logs_and_database_without_mutation(self):
+        log_result = mock.Mock(
+            returncode=0,
+            stdout="Error: device_database_migration_blocked\n",
+            stderr="at PostgresDeviceRepository.migrate (postgres-repository.mjs:110)\n",
+        )
+        database_result = mock.Mock(
+            returncode=0,
+            stdout=(
+                "activity\t42,active,Lock,relation,schema-ddl,181\n"
+                "schema\thost-telemetry-table-present\n"
+            ),
+            stderr="",
+        )
+        with mock.patch.object(
+            RUNNER,
+            "device_plane_service_container_ids",
+            side_effect=[("c" * 64,), ("p" * 64,)],
+        ), mock.patch.object(
+            RUNNER.subprocess,
+            "run",
+            side_effect=[log_result, database_result],
+        ) as run:
+            evidence = RUNNER.collect_device_plane_control_core_incident_audit()
+
+        self.assertEqual(evidence["coreContainerId"], "c" * 64)
+        self.assertIn(
+            "Error: device_database_migration_blocked",
+            evidence["logErrors"],
+        )
+        self.assertEqual(
+            evidence["database"],
+            [
+                "activity\t42,active,Lock,relation,schema-ddl,181",
+                "schema\thost-telemetry-table-present",
+            ],
+        )
+        self.assertEqual(run.call_count, 2)
+        log_command = run.call_args_list[0].args[0]
+        self.assertEqual(log_command[1:4], ["logs", "--tail", "240"])
+        database_command = run.call_args_list[1].args[0]
+        self.assertEqual(database_command[1], "exec")
+        query = database_command[database_command.index("-c") + 1]
+        self.assertTrue(query.lstrip().startswith("select 'schema'"))
+        self.assertNotIn(";", query)
+
+    def test_apply_path_rejects_plan_only_audit_before_state_mutation(self):
+        source = inspect.getsource(RUNNER.apply_artifact)
+        self.assertIn("canonical-plan-only", source)
+        self.assertIn("apply is forbidden", source)
+        self.assertLess(
+            source.index("canonical-plan-only"),
+            source.index("state_has_sha"),
+        )
+
+
+if __name__ == "__main__":
+    unittest.main()