feat: schedule persistent archive jobs with console management

scripts/test-console-e2e.sh

@@ -21,6 +21,15 @@ cleanup() {
         done
         rmdir "$E2E_DIR/archive-job" >/dev/null 2>&1 || true
     fi
+    if [ -d "$E2E_DIR/archive-volume" ]; then
+        python3 - "$E2E_DIR/archive-volume" <<'PY_CLEAN'
+import pathlib, shutil, sys
+path = pathlib.Path(sys.argv[1])
+# 仅清理由该测试 mktemp 目录下创建的独立归档卷。
+if path.name == "archive-volume" and path.parent.name.startswith("orbit-console-e2e."):
+    shutil.rmtree(path)
+PY_CLEAN
+    fi
     for e2e_artifact in "$E2E_DIR"/*; do
         if [ -f "$e2e_artifact" ]; then
             unlink "$e2e_artifact"
@@ -90,11 +99,17 @@ print(queue)
 PY
 )
 
+mkdir -m 700 "$E2E_DIR/archive-volume"
+E2E_STORAGE_ID=$(openssl rand -hex 16)
 printf '%s\n' "[4/8] start Orbit container"
 "$E2E_DOCKER" run -d --name "$E2E_ORBIT" --network bridge --link "$E2E_MYSQL:mysql" \
     --env ORBIT_MYSQL_DSN="$E2E_DSN" \
     --env ORBIT_HTTP_ADDR=0.0.0.0:8080 \
     --env ORBIT_CONSOLE_QUEUE_ID="$E2E_QUEUE" \
+    --env ORBIT_ARCHIVE_ROOT=/var/lib/orbit/archive \
+    --env ORBIT_ARCHIVE_STORAGE_ID="$E2E_STORAGE_ID" \
+    --env ORBIT_ARCHIVE_TICK_INTERVAL=1s \
+    --volume "$E2E_DIR/archive-volume:/var/lib/orbit/archive" \
     --publish 127.0.0.1::8080 "$E2E_IMAGE" >/dev/null
 E2E_PORT=
 for e2e_attempt in $(seq 1 30); do
@@ -247,12 +262,39 @@ assert content["source_retained"] is True and content["root_count"] == 0
 status, _ = request("/console/retention/export", "POST", {"kind": "tasks", "expected_revision": "0"})
 assert status == 409, status
 print("retention_policy=ok replay=ok stale_revision=409 preview_categories=7 empty_export_digest=ok")
+status, body = request("/console/archive/plans")
+assert status == 200, (status, body)
+plans = json.loads(body)["data"]
+assert len(plans["items"]) == 7 and plans["executor_configured"] is True
+plan = dict(expected_revision="0", policy_revision="1", enabled=True, interval_seconds=3600, batch_size=20, max_bytes=1048576, reason="console scheduler e2e")
+plan_headers = {"Idempotency-Key": "archive-plan-e2e"}
+status, body = request("/console/archive/plans/tasks", "PUT", plan, {**plan_headers, "Origin": "https://different-origin.invalid"})
+assert status == 403, (status, body)
+status, body = request("/console/archive/plans/tasks", "PUT", plan, plan_headers)
+assert status == 200 and json.loads(body)["data"]["revision"] == "1", (status, body)
+status, body = request("/console/archive/plans/tasks", "PUT", plan, plan_headers)
+assert status == 200 and json.loads(body)["data"]["revision"] == "1", (status, body)
+import time
+for attempt in range(30):
+    status, body = request("/console/archive/jobs")
+    assert status == 200, (status, body)
+    jobs = json.loads(body)["data"]["items"]
+    if jobs and jobs[0]["state"] in ("SUCCEEDED", "FAILED"):
+        break
+    time.sleep(1)
+assert len(jobs) == 1 and jobs[0]["state"] == "SUCCEEDED", jobs
+assert jobs[0]["batches"] == 1 and jobs[0]["roots"] == 0 and len(jobs[0]["last_sha256"]) == 64
+assert "lease_id" not in jobs[0]
+print("archive_scheduler=ok plan_replay=ok cross_origin=403 durable_job=SUCCEEDED")
+
 status, _ = request("/console/auth/logout", "POST", {})
 assert status == 200, status
 status, _ = request("/v1/tasks?" + urllib.parse.urlencode({"queue_id": queue, "limit": 1}))
 assert status == 401, status
 status_retention, _ = request("/console/retention")
 assert status_retention == 401, status_retention
+status_archive, _ = request("/console/archive/jobs")
+assert status_archive == 401, status_archive
 print("console_task_data=ok succeeded_tasks=%d logout=ok unauthorized_after_logout=%d" % (sum(item.get("state") == "SUCCEEDED" for item in items), status))
 PY