fix(deploy): add exact-image Core v3 recovery

This commit is contained in:
DCCONSTRUCTIONS
2026-08-22 19:49:36 +03:00
parent d12b172569
commit 6dad30c565
3 changed files with 414 additions and 0 deletions
@@ -0,0 +1,25 @@
{
"schemaVersion": "nodedc.device-plane.device-control-core-release-v3-reconciliation.v1",
"mode": "failed-release-v3-exact-preapply-image-restore",
"failedPatchId": "device-control-core-release-v3-20260822-040",
"failedArtifactSha256": "08448a56cdf391076f92c5242e368fd0033b420167874645eebfa1f22184ee92",
"failedArtifact": "nodedc-device-plane-device-control-core-release-v3-20260822-040.tgz.20260822-184245",
"backupId": "device-plane-device-control-core-release-v3-20260822-040-20260822-184245",
"predecessorPatchId": "device-control-core-release-v2-20260822-038",
"predecessorArtifactSha256": "e2d062b82b022dba662522b5d6e192026ac964d78950d903295ca3cbbc95ab28",
"preapplyImageId": "sha256:31d35733ee46225b487c0f02a7b52d4ba2d13f5b99f6a717b7f5e6f5460b412a",
"sourceAction": "accept-byte-exact-restored-preapply-source",
"runtimeAction": "retag-exact-preapply-image+recreate-device-control-core-only",
"preservedServices": [
"device-manager",
"device-gateway",
"device-postgres",
"device-backhaul-target"
],
"databaseVolume": "nodedc-device-plane-postgres-data",
"publicIngress": "disabled",
"edgeChannel": "core-initiated-pinned-mtls-registered-edges-only",
"commandTransport": "typed-service-ping-v1",
"gelios": "untouched-legacy-only",
"rollback": "marker+exact-preapply-image-runtime"
}
@@ -0,0 +1,114 @@
#!/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-release-v3-reconciliation-20260822-042",
...extra
] = process.argv.slice(2);
if (extra.length || !/^[A-Za-z0-9._-]{1,96}$/.test(patchId)) {
throw new Error(
"usage: build-device-control-core-release-v3-reconciliation-artifact.mjs [patch-id]",
);
}
const entry = "deployment/device-control-core-release-v3-reconciliation-v1.json";
const stage = await mkdtemp(join(tmpdir(), "nodedc-control-core-v3-reconciliation-"));
const payload = join(stage, "payload");
const target = resolve(artifactDir, `nodedc-device-plane-${patchId}.tgz`);
try {
const descriptor = JSON.parse(await readFile(resolve(sourceRoot, entry), "utf8"));
const expected = {
schemaVersion: "nodedc.device-plane.device-control-core-release-v3-reconciliation.v1",
mode: "failed-release-v3-exact-preapply-image-restore",
failedPatchId: "device-control-core-release-v3-20260822-040",
failedArtifactSha256:
"08448a56cdf391076f92c5242e368fd0033b420167874645eebfa1f22184ee92",
failedArtifact:
"nodedc-device-plane-device-control-core-release-v3-20260822-040.tgz.20260822-184245",
backupId:
"device-plane-device-control-core-release-v3-20260822-040-20260822-184245",
predecessorPatchId: "device-control-core-release-v2-20260822-038",
predecessorArtifactSha256:
"e2d062b82b022dba662522b5d6e192026ac964d78950d903295ca3cbbc95ab28",
preapplyImageId:
"sha256:31d35733ee46225b487c0f02a7b52d4ba2d13f5b99f6a717b7f5e6f5460b412a",
sourceAction: "accept-byte-exact-restored-preapply-source",
runtimeAction: "retag-exact-preapply-image+recreate-device-control-core-only",
preservedServices: [
"device-manager",
"device-gateway",
"device-postgres",
"device-backhaul-target",
],
databaseVolume: "nodedc-device-plane-postgres-data",
publicIngress: "disabled",
edgeChannel: "core-initiated-pinned-mtls-registered-edges-only",
commandTransport: "typed-service-ping-v1",
gelios: "untouched-legacy-only",
rollback: "marker+exact-preapply-image-runtime",
};
if (JSON.stringify(descriptor) !== JSON.stringify(expected)) {
throw new Error("device_control_core_v3_reconciliation_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: ["device-control-core"],
transition: descriptor.mode,
runtimeAction: descriptor.runtimeAction,
preservedServices: descriptor.preservedServices,
}, 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");
}
@@ -0,0 +1,275 @@
#!/usr/bin/env python3
import hashlib
import importlib.machinery
import importlib.util
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-release-v3-reconciliation-artifact.mjs"
)
def load_runner():
loader = importlib.machinery.SourceFileLoader(
"nodedc_control_core_v3_reconciliation_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 DeviceControlCoreV3ReconciliationArtifactTest(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_exact_deterministic_runtime_reconciliation(self):
with tempfile.TemporaryDirectory(
prefix="nodedc-control-core-v3-reconciliation-artifact-",
) as directory:
artifact_dir = Path(directory)
patch_id = "device-control-core-v3-reconciliation-unit-042"
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["component"], "device-plane")
self.assertEqual(
first["entries"],
list(RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_RECONCILIATION_ENTRIES),
)
self.assertEqual(first["build"], [])
self.assertEqual(first["services"], ["device-control-core"])
with tarfile.open(first["artifact"], "r:gz") as archive:
names = [member.name for member in archive.getmembers()]
descriptor = json.loads(
archive.extractfile(
"payload/"
+ RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_RECONCILIATION_REL
).read().decode("utf-8")
)
self.assertEqual(
descriptor,
RUNNER.expected_device_plane_control_core_v3_reconciliation_descriptor(),
)
self.assertEqual(
names,
[
"manifest.env",
"files.txt",
"payload",
"payload/deployment",
"payload/"
+ RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_RECONCILIATION_REL,
],
)
self.assertEqual(
RUNNER.component_services(
"device-plane",
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_RECONCILIATION_ENTRIES,
),
("device-control-core",),
)
self.assertEqual(
RUNNER.component_builds(
"device-plane",
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_RECONCILIATION_ENTRIES,
),
(),
)
def test_descriptor_pins_exact_failed_evidence_and_preapply_image(self):
descriptor = (
RUNNER.expected_device_plane_control_core_v3_reconciliation_descriptor()
)
self.assertEqual(
descriptor["failedArtifactSha256"],
"08448a56cdf391076f92c5242e368fd0033b420167874645eebfa1f22184ee92",
)
self.assertEqual(
descriptor["preapplyImageId"],
"sha256:31d35733ee46225b487c0f02a7b52d4ba2d13f5b99f6a717b7f5e6f5460b412a",
)
self.assertEqual(
descriptor["runtimeAction"],
"retag-exact-preapply-image+recreate-device-control-core-only",
)
def test_runtime_recovery_retags_exact_image_without_build(self):
completed = mock.Mock(returncode=0, stdout="", stderr="")
with mock.patch.object(
RUNNER,
"inspect_optional_local_image",
side_effect=[
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_PREAPPLY_IMAGE_ID,
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_PREAPPLY_IMAGE_ID,
],
), mock.patch.object(
RUNNER.subprocess,
"run",
return_value=completed,
) as run:
RUNNER.restore_device_plane_control_core_v3_preapply_image()
run.assert_called_once_with(
[
str(RUNNER.DOCKER),
"image",
"tag",
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_PREAPPLY_IMAGE_ID,
RUNNER.DEVICE_PLANE_CONTROL_CORE_IMAGE,
],
check=False,
capture_output=True,
text=True,
)
def test_reconciliation_runtime_selects_only_core_and_no_build(self):
mark_runtime_started = mock.Mock()
with mock.patch.object(
RUNNER,
"prepare_component_runtime",
) as prepare, mock.patch.object(
RUNNER,
"restore_device_plane_control_core_v3_preapply_image",
) as restore, mock.patch.object(
RUNNER,
"run_compose",
) as compose, mock.patch.object(
RUNNER,
"run_build",
) as build:
RUNNER.run_device_plane_runtime_for_apply(
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_RECONCILIATION_ENTRIES,
("device-control-core",),
mark_runtime_started,
)
prepare.assert_called_once_with(
"device-plane",
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_RECONCILIATION_ENTRIES,
)
mark_runtime_started.assert_called_once_with()
restore.assert_called_once_with()
compose.assert_called_once_with(
"device-plane",
("device-control-core",),
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_RECONCILIATION_ENTRIES,
)
build.assert_not_called()
def test_release_rollback_reuses_exact_preapply_image_without_rebuild(self):
runtime_inventory = {
"schemaVersion": "nodedc.device-plane.runtime-inventory.v1",
"composeProject": "nodedc-device-plane",
"services": [{
"service": "device-control-core",
"containerId": "1" * 64,
"imageId": RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_PREAPPLY_IMAGE_ID,
"status": "running",
"running": True,
"health": "healthy",
"restartCount": 0,
}],
}
entries = RUNNER.DEVICE_PLANE_CONTROL_CORE_RELEASE_V3_ENTRIES
with tempfile.TemporaryDirectory(
prefix="nodedc-control-core-v3-rollback-unit-",
) as directory, mock.patch.object(
RUNNER,
"read_backup_path_list",
side_effect=[list(entries), []],
), mock.patch.object(
RUNNER,
"validate_backup_partition",
return_value=(set(entries), set()),
), mock.patch.object(
RUNNER,
"read_strict_json",
return_value=runtime_inventory,
), mock.patch.object(
RUNNER,
"restore_platform_overlay",
return_value=len(entries),
), mock.patch.object(
RUNNER,
"retag_device_plane_control_core_image",
) as retag, mock.patch.object(
RUNNER,
"prepare_component_runtime",
), mock.patch.object(
RUNNER,
"run_compose",
) as compose, mock.patch.object(
RUNNER,
"run_component_runtime",
) as build_runtime, mock.patch.object(
RUNNER,
"accept_device_plane_control_core_rollback_runtime",
), mock.patch.object(
RUNNER,
"validate_device_manager_control_plane_runtime",
):
result = RUNNER.rollback_device_plane_apply(
Path(directory),
Path(directory),
entries,
"unit-stamp",
True,
("device-control-core",),
)
self.assertEqual(
result,
f"source+runtime-restored:{len(entries)}",
)
retag.assert_called_once_with(
RUNNER.DEVICE_PLANE_CONTROL_CORE_V3_PREAPPLY_IMAGE_ID,
"Device Control Core exact pre-apply rollback image",
)
compose.assert_called_once_with(
"device-plane",
("device-control-core",),
list(entries),
)
build_runtime.assert_not_called()
if __name__ == "__main__":
unittest.main()