feat(1d): migrate keycloak/cryptpad/matrix-synapse/n8n/lasuite-docs overlays to deploy-once contract (DG7)

Mechanical port to the assertion-only contract (no softened/skipped assertions): install uses
live_app + generic.assert_serving (extend) + the recipe's http/playwright/api checks; upgrade seeds
its data marker then generic.do_upgrade + asserts survival; backup/restore split into test_backup.py
(seed->do_backup->mutate) + new test_restore.py (do_restore->assert original). Recipe-specifics
preserved verbatim (keycloak realm+admin-console+kc_admin, matrix/lasuite db-service psql markers,
cryptpad/n8n volume markers). No recipe now double-deploys under the deploy-once orchestrator.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-05-28 01:32:53 +01:00
parent 9b5bcff92a
commit afd75a48db
21 changed files with 315 additions and 325 deletions

View File

@ -1,5 +1,7 @@
"""cryptpad — backup/restore stage (D2): write a marker into the backed-up cryptpad_data volume,
backup, mutate, restore, assert the restored state matches the pre-mutation (backed-up) state.
"""cryptpad — BACKUP overlay (Phase 1d, DG4): seed a known state into the backed-up cryptpad_data
volume, back it up (assert a snapshot artifact), then mutate so the RESTORE overlay (test_restore.py)
can prove the backed-up state returns. Runs on the shared deployment; the mutated marker persists for
the restore tier.
The cryptpad `app` service is labelled `backupbot.backup=true`, so its volumes (incl. cryptpad_data)
are backed up. Marker is checked via `exec_in_app` (data isn't HTTP-served)."""
@ -8,32 +10,21 @@ import os
import sys
sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "..", "runner"))
from harness import lifecycle # noqa: E402
from harness import generic, lifecycle # noqa: E402
MARKER = "/cryptpad/data/ci-marker.txt"
def test_backup_mutate_restore(deployed, meta):
domain = deployed
def test_backup_captures_state(live_app, meta):
domain = live_app
# 1) establish original state in the backed-up volume, then back it up
# 1) establish original state in the backed-up volume, then back it up (reuse the generic op:
# backup + assert a snapshot artifact was produced)
lifecycle.exec_in_app(domain, ["sh", "-c", f"echo original > {MARKER}"])
assert lifecycle.exec_in_app(domain, ["cat", MARKER]).strip() == "original"
lifecycle.backup_app(domain)
snap = generic.do_backup(domain)
assert snap, "backup produced no snapshot artifact"
# 2) mutate state (diverge from the backup)
lifecycle.exec_in_app(domain, ["sh", "-c", f"echo mutated > {MARKER}"])
assert lifecycle.exec_in_app(domain, ["cat", MARKER]).strip() == "mutated"
# 3) restore -> state returns to the backed-up "original"
lifecycle.restore_app(domain)
lifecycle.wait_healthy(
domain,
ok_codes=tuple(meta["HEALTH_OK"]),
path=meta["HEALTH_PATH"],
deploy_timeout=meta["DEPLOY_TIMEOUT"],
http_timeout=meta["HTTP_TIMEOUT"],
)
assert (
lifecycle.exec_in_app(domain, ["cat", MARKER]).strip() == "original"
), "restore did not return the pre-mutation state"

View File

@ -1,23 +1,29 @@
"""cryptpad — install stage (recipe #3, stateful/no-DB). D2 install + D3 Playwright."""
"""cryptpad — INSTALL overlay (Phase 1d, DG4): override + extend-by-composition.
Reuses the generic "really serving" assertion, then ADDS the recipe-specific checks: cryptpad answers
over real HTTPS through the gateway, and a real browser loads the live cryptpad landing page and sees
its served app (D2 install + D3 Playwright). Assertion-only on the shared deployment."""
import os
import sys
sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "..", "runner"))
from harness import lifecycle # noqa: E402
from harness import generic, lifecycle # noqa: E402
def test_http_reachable(deployed_app):
"""cryptpad answers over real HTTPS through the gateway (nginx -> cryptpad app)."""
status = lifecycle.http_get(deployed_app, "/")
assert status in (200, 301, 302), f"expected 2xx/3xx from {deployed_app}, got {status}"
def test_serving_and_content(live_app, meta):
# extend-by-composition: reuse the generic "really serving" assertion first ...
generic.assert_serving(live_app, meta)
# ... then the recipe-specific assertions.
# cryptpad answers over real HTTPS through the gateway (nginx -> cryptpad app).
status = lifecycle.http_get(live_app, "/")
assert status in (200, 301, 302), f"expected 2xx/3xx from {live_app}, got {status}"
def test_playwright_loads_cryptpad(deployed_app):
"""A real browser loads the live cryptpad landing page and sees its served app."""
# A real browser loads the live cryptpad landing page and sees its served app.
from playwright.sync_api import sync_playwright
url = f"https://{deployed_app}/"
url = f"https://{live_app}/"
with sync_playwright() as p:
browser = p.chromium.launch(args=["--no-sandbox"])
try:

View File

@ -0,0 +1,24 @@
"""cryptpad — RESTORE overlay (Phase 1d, DG4): data-integrity, extends the generic restore.
Runs after the backup overlay (test_backup.py) on the SAME shared deployment, which left the
cryptpad_data marker mutated to "mutated" after backing up "original". This restores the snapshot via
the shared op helper (`generic.do_restore`, which also asserts the app is healthy + serving
afterwards), then asserts the volume data returned to the pre-mutation "original" — the app-specific
data integrity the generic restore cannot check. Reads the marker via `exec_in_app` (data isn't
HTTP-served). Assertion-only (no deploy/teardown)."""
import os
import sys
sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "..", "runner"))
from harness import generic, lifecycle # noqa: E402
MARKER = "/cryptpad/data/ci-marker.txt"
def test_restore_returns_state(live_app, meta):
domain = live_app
generic.do_restore(domain, meta) # restore + assert healthy/serving
assert (
lifecycle.exec_in_app(domain, ["cat", MARKER]).strip() == "original"
), "restore did not return the pre-mutation state"

View File

@ -1,53 +1,28 @@
"""cryptpad — upgrade stage (D2): deploy the previous published version, write a data marker into a
persistent volume, upgrade to current/$REF, assert the app stays healthy and the data survives.
"""cryptpad — UPGRADE overlay (Phase 1d, DG4): data-continuity, extends the generic upgrade.
cryptpad data isn't HTTP-served as a static file (it's an encrypted datastore), so the marker is
written into the cryptpad_data volume and read back via `exec_in_app` (docker exec), not HTTP."""
The orchestrator deployed the previous published version ONCE; this overlay writes a marker into the
persistent cryptpad_data volume (cryptpad data isn't HTTP-served as a static file — it's an encrypted
datastore — so the marker is read back via `exec_in_app`, not HTTP), performs the in-place upgrade via
the shared op helper (`generic.do_upgrade`, which also asserts reconverge + serving + that the
deployment moved), then asserts the data SURVIVED. Assertion-only on the shared deployment."""
import os
import sys
import pytest
sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "..", "runner"))
from harness import lifecycle # noqa: E402
from harness import generic, lifecycle # noqa: E402
MARKER = "/cryptpad/data/ci-marker.txt"
@pytest.fixture
def old_app(recipe, app_domain, meta, request):
prev = lifecycle.previous_version(recipe)
if not prev:
pytest.skip(f"{recipe}: no previous published version to upgrade from")
lifecycle.janitor()
request.addfinalizer(lambda: lifecycle.teardown_app(app_domain))
lifecycle.deploy_app(recipe, app_domain, version=prev)
lifecycle.wait_healthy(
app_domain,
ok_codes=tuple(meta["HEALTH_OK"]),
path=meta["HEALTH_PATH"],
deploy_timeout=meta["DEPLOY_TIMEOUT"],
http_timeout=meta["HTTP_TIMEOUT"],
)
return app_domain, prev
def test_upgrade_preserves_data(old_app, meta):
domain, prev = old_app
def test_upgrade_preserves_data(live_app, meta):
domain = live_app
# write a data marker into the persistent cryptpad_data volume
lifecycle.exec_in_app(domain, ["sh", "-c", f"echo upgrade-survives > {MARKER}"])
assert lifecycle.exec_in_app(domain, ["cat", MARKER]).strip() == "upgrade-survives"
# upgrade previous -> current/$REF
lifecycle.upgrade_app(domain, version=os.environ.get("VERSION") or None)
lifecycle.wait_healthy(
domain,
ok_codes=tuple(meta["HEALTH_OK"]),
path=meta["HEALTH_PATH"],
deploy_timeout=meta["DEPLOY_TIMEOUT"],
http_timeout=meta["HTTP_TIMEOUT"],
)
# in-place upgrade previous -> target (reuses the generic op: upgrade + assert reconverge/serving)
generic.do_upgrade(domain, os.environ.get("VERSION") or None, meta)
# app healthy and the data written before the upgrade is still there
assert lifecycle.http_get(domain, "/") in (200, 301, 302)