feat: 新增GPU测试节点argus-metric-test-gpu-node，测试安装包安装流程及与agent/master组件之间数据准确性；

refs #20

build/build_images.sh

@@ -10,6 +10,7 @@ Usage: $0 [OPTIONS]
 Options:
   --intranet        Use intranet mirror for log/bind builds
   --master-offline  Build master offline image (requires src/master/offline_wheels.tar.gz)
+  --no-cache        Build all images without using Docker layer cache
   -h, --help        Show this help message
 
 Examples:
@@ -23,6 +24,7 @@ EOF
 use_intranet=false
 build_master=true
 build_master_offline=false
+no_cache=false
 
 while [[ $# -gt 0 ]]; do
   case $1 in
@@ -39,6 +41,10 @@ while [[ $# -gt 0 ]]; do
       build_master_offline=true
       shift
       ;;
+    --no-cache)
+      no_cache=true
+      shift
+      ;;
     -h|--help)
       show_help
       exit 0
@@ -65,6 +71,10 @@ cd "$root"
 load_build_user
 build_args+=("--build-arg" "ARGUS_BUILD_UID=${ARGUS_BUILD_UID}" "--build-arg" "ARGUS_BUILD_GID=${ARGUS_BUILD_GID}")
 
+if [[ "$no_cache" == true ]]; then
+  build_args+=("--no-cache")
+fi
+
 master_root="$root/src/master"
 master_offline_tar="$master_root/offline_wheels.tar.gz"
 master_offline_dir="$master_root/offline_wheels"
@@ -159,6 +169,9 @@ if [[ "$build_master" == true ]]; then
   if [[ "$build_master_offline" == true ]]; then
     master_args+=("--offline")
   fi
+  if [[ "$no_cache" == true ]]; then
+    master_args+=("--no-cache")
+  fi
   if ./scripts/build_images.sh "${master_args[@]}"; then
     if [[ "$build_master_offline" == true ]]; then
       images_built+=("argus-master:offline")

src/agent/README.md

@@ -34,6 +34,18 @@ Agent 不再依赖配置文件；所有参数均由环境变量与主机名推
 | `MASTER_ENDPOINT` | 是 | N/A | Master 基础地址，可写 `http://host:3000` 或 `host:3000`（自动补全 `http://`）。 |
 | `REPORT_INTERVAL_SECONDS` | 否 | `60` | 状态上报间隔（秒）。必须为正整数。 |
 | `AGENT_HOSTNAME` | 否 | `$(hostname)` | 覆盖容器内主机名，便于测试或特殊命名需求。 |
+| `AGENT_ENV` | 否 | 来源于主机名 | 运行环境标识（如 `dev`、`prod`）。与 `AGENT_USER`、`AGENT_INSTANCE` 必须同时设置。 |
+| `AGENT_USER` | 否 | 来源于主机名 | 归属用户或团队标识。与 `AGENT_ENV`、`AGENT_INSTANCE` 必须同时设置。 |
+| `AGENT_INSTANCE` | 否 | 来源于主机名 | 实例编号或别名。与 `AGENT_ENV`、`AGENT_USER` 必须同时设置。 |
+
+主机名与元数据的解析优先级：
+
+1. 若设置 `AGENT_ENV` / `AGENT_USER` / `AGENT_INSTANCE` 且全部存在，则直接使用这些值。
+2. 否则检查历史 `node.json`（注册成功后由 Master 返回的信息），若包含 `env` / `user` / `instance` 则沿用。
+3. 若以上均不可用，则按历史约定从主机名解析 `env-user-instance` 前缀。
+4. 如果仍无法得到完整结果，Agent 启动会失败并提示需要提供上述环境变量。
+
+> 提示：在首次部署时需确保环境变量或主机名能够提供完整信息。完成注册后，Agent 会把 Master 返回的元数据写入 `node.json`，后续重启无需再次提供环境变量就能保持一致性。
 
 派生路径：
 

src/agent/app/collector.py

@@ -18,13 +18,12 @@ _HOSTNAME_PATTERN = re.compile(r"^([^-]+)-([^-]+)-([^-]+)-.*$")
 def collect_metadata(config: AgentConfig) -> Dict[str, Any]:
     """汇总节点注册需要的静态信息。"""
     hostname = config.hostname
-    env, user, instance = _parse_hostname(hostname)
     meta = {
         "hostname": hostname,
         "ip": _detect_ip_address(),
-        "env": env,
+        "env": config.environment,
-        "user": user,
+        "user": config.user,
-        "instance": instance,
+        "instance": config.instance,
         "cpu_number": _detect_cpu_count(),
         "memory_in_bytes": _detect_memory_bytes(),
         "gpu_number": _detect_gpu_count(),

src/agent/app/config.py

@@ -6,14 +6,21 @@ from dataclasses import dataclass
 from pathlib import Path
 from typing import Final
 
+from .state import load_node_state
 from .version import VERSION
+from .log import get_logger
 
 DEFAULT_REPORT_INTERVAL_SECONDS: Final[int] = 60
 
+LOGGER = get_logger("argus.agent.config")
+
 
 @dataclass(frozen=True)
 class AgentConfig:
     hostname: str
+    environment: str
+    user: str
+    instance: str
     node_file: str
     version: str
     master_endpoint: str
@@ -47,11 +54,68 @@ def _resolve_hostname() -> str:
     return os.environ.get("AGENT_HOSTNAME") or socket.gethostname()
 
 
+def _load_metadata_from_state(node_file: str) -> tuple[str, str, str] | None:
+    state = load_node_state(node_file)
+    if not state:
+        return None
+
+    meta = state.get("meta_data") or {}
+    env = meta.get("env") or state.get("env")
+    user = meta.get("user") or state.get("user")
+    instance = meta.get("instance") or state.get("instance")
+
+    if env and user and instance:
+        LOGGER.debug("Metadata resolved from node state", extra={"node_file": node_file})
+        return env, user, instance
+
+    LOGGER.warning(
+        "node.json missing metadata fields; ignoring",
+        extra={"node_file": node_file, "meta_data": meta},
+    )
+    return None
+
+
+def _resolve_metadata_fields(hostname: str, node_file: str) -> tuple[str, str, str]:
+    env = os.environ.get("AGENT_ENV")
+    user = os.environ.get("AGENT_USER")
+    instance = os.environ.get("AGENT_INSTANCE")
+
+    if env and user and instance:
+        return env, user, instance
+
+    if any([env, user, instance]):
+        LOGGER.warning(
+            "Incomplete metadata environment variables; falling back to persisted metadata",
+            extra={
+                "has_env": bool(env),
+                "has_user": bool(user),
+                "has_instance": bool(instance),
+            },
+        )
+
+    state_metadata = _load_metadata_from_state(node_file)
+    if state_metadata is not None:
+        return state_metadata
+
+    from .collector import _parse_hostname  # Local import to avoid circular dependency
+
+    env, user, instance = _parse_hostname(hostname)
+
+    if not all([env, user, instance]):
+        raise ValueError(
+            "Failed to determine metadata fields; set AGENT_ENV/USER/INSTANCE or use supported hostname pattern"
+        )
+
+    return env, user, instance
+
+
 def load_config() -> AgentConfig:
     """从环境变量推导配置，移除了外部配置文件依赖。"""
 
     hostname = _resolve_hostname()
     node_file = f"/private/argus/agent/{hostname}/node.json"
+    environment, user, instance = _resolve_metadata_fields(hostname, node_file)
+
     health_dir = f"/private/argus/agent/{hostname}/health/"
 
     master_endpoint_env = os.environ.get("MASTER_ENDPOINT")
@@ -66,6 +130,9 @@ def load_config() -> AgentConfig:
 
     return AgentConfig(
         hostname=hostname,
+        environment=environment,
+        user=user,
+        instance=instance,
         node_file=node_file,
         version=VERSION,
         master_endpoint=master_endpoint,

src/agent/tests/docker-compose.yml

@@ -60,6 +60,36 @@ services:
         ipv4_address: 172.28.0.20
     restart: always
 
+  agent_env:
+    image: ubuntu:22.04
+    container_name: argus-agent-env-e2e
+    hostname: host_abc
+    depends_on:
+      - master
+      - bind
+    environment:
+      - MASTER_ENDPOINT=http://master.argus.com:3000
+      - REPORT_INTERVAL_SECONDS=2
+      - AGENT_ENV=prod
+      - AGENT_USER=ml
+      - AGENT_INSTANCE=node-3
+      - AGENT_HOSTNAME=host_abc
+      - "ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}"
+      - "ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}"
+    volumes:
+      - ./private/argus/agent/host_abc:/private/argus/agent/host_abc
+      - ./private/argus/agent/host_abc/health:/private/argus/agent/host_abc/health
+      - ./private/argus/etc:/private/argus/etc
+      - ../dist/argus-agent:/usr/local/bin/argus-agent:ro
+      - ./scripts/agent_entrypoint.sh:/usr/local/bin/agent-entrypoint.sh:ro
+      - ../scripts/agent_deployment_verify.sh:/usr/local/bin/agent_deployment_verify.sh:ro
+    entrypoint:
+      - /usr/local/bin/agent-entrypoint.sh
+    networks:
+      default:
+        ipv4_address: 172.28.0.21
+    restart: always
+
 networks:
   default:
     driver: bridge

src/agent/tests/scripts/00_e2e_test.sh

@@ -7,10 +7,10 @@ SCRIPTS=(
   "02_up.sh"
   "03_wait_and_assert_registration.sh"
   "04_write_health_files.sh"
-  "08_verify_agent.sh"
+  "05_verify_agent.sh"
-  "05_assert_status_on_master.sh"
+  "06_assert_status_on_master.sh"
-  "06_restart_agent_and_reregister.sh"
+  "07_restart_agent_and_reregister.sh"
-  "07_down.sh"
+  "08_down.sh"
 )
 
 for script in "${SCRIPTS[@]}"; do

src/agent/tests/scripts/02_up.sh

@@ -41,7 +41,7 @@ compose() {
   fi
 }
 
-docker container rm -f argus-agent-e2e argus-master-agent-e2e argus-bind-agent-e2e >/dev/null 2>&1 || true
+docker container rm -f argus-agent-e2e argus-agent-env-e2e argus-master-agent-e2e argus-bind-agent-e2e >/dev/null 2>&1 || true
 
 docker network rm tests_default >/dev/null 2>&1 || true
 

src/agent/tests/scripts/03_wait_and_assert_registration.sh

@@ -6,11 +6,14 @@ TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
 TMP_ROOT="$TEST_ROOT/tmp"
 API_BASE="http://localhost:32300/api/v1/master"
 AGENT_HOSTNAME="dev-e2euser-e2einst-pod-0"
+ENV_AGENT_HOSTNAME="host_abc"
 NODE_FILE="$TEST_ROOT/private/argus/agent/$AGENT_HOSTNAME/node.json"
+ENV_NODE_FILE="$TEST_ROOT/private/argus/agent/$ENV_AGENT_HOSTNAME/node.json"
 
 mkdir -p "$TMP_ROOT"
 
-node_id=""
+primary_node_id=""
+env_node_id=""
 for _ in {1..30}; do
   sleep 2
   response=$(curl -sS "$API_BASE/nodes" || true)
@@ -19,24 +22,49 @@ for _ in {1..30}; do
   fi
   list_file="$TMP_ROOT/nodes_list.json"
   echo "$response" > "$list_file"
-  node_id=$(python3 - "$list_file" <<'PY'
+  readarray -t node_ids < <(python3 - "$list_file" "$AGENT_HOSTNAME" "$ENV_AGENT_HOSTNAME" <<'PY'
 import json, sys
+
 with open(sys.argv[1]) as handle:
     nodes = json.load(handle)
-print(nodes[0]["id"] if nodes else "")
+
+target_primary = sys.argv[2]
+target_env = sys.argv[3]
+
+primary_id = ""
+env_id = ""
+
+for node in nodes:
+    if node.get("name") == target_primary:
+        primary_id = node.get("id", "")
+    if node.get("name") == target_env:
+        env_id = node.get("id", "")
+
+print(primary_id)
+print(env_id)
 PY
-)
+  )
-  if [[ -n "$node_id" ]]; then
+
+  primary_node_id="${node_ids[0]}"
+  env_node_id="${node_ids[1]}"
+
+  if [[ -n "$primary_node_id" && -n "$env_node_id" ]]; then
     break
   fi
  done
 
-if [[ -z "$node_id" ]]; then
+if [[ -z "$primary_node_id" ]]; then
-  echo "[ERROR] Agent did not register within timeout" >&2
+  echo "[ERROR] Primary agent did not register within timeout" >&2
   exit 1
 fi
 
-echo "$node_id" > "$TMP_ROOT/node_id"
+if [[ -z "$env_node_id" ]]; then
+  echo "[ERROR] Env-variable agent did not register within timeout" >&2
+  exit 1
+fi
+
+echo "$primary_node_id" > "$TMP_ROOT/node_id"
+echo "$env_node_id" > "$TMP_ROOT/node_id_host_abc"
 
 if [[ ! -f "$NODE_FILE" ]]; then
   echo "[ERROR] node.json not created at $NODE_FILE" >&2
@@ -50,8 +78,20 @@ with open(sys.argv[1]) as handle:
 assert "id" in node and node["id"], "node.json missing id"
 PY
 
+if [[ ! -f "$ENV_NODE_FILE" ]]; then
+  echo "[ERROR] node.json not created at $ENV_NODE_FILE" >&2
+  exit 1
+fi
+
+python3 - "$ENV_NODE_FILE" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+assert "id" in node and node["id"], "env agent node.json missing id"
+PY
+
 detail_file="$TMP_ROOT/initial_detail.json"
-curl -sS "$API_BASE/nodes/$node_id" -o "$detail_file"
+curl -sS "$API_BASE/nodes/$primary_node_id" -o "$detail_file"
 python3 - "$detail_file" "$TMP_ROOT/initial_ip" <<'PY'
 import json, sys, pathlib
 with open(sys.argv[1]) as handle:
@@ -62,4 +102,5 @@ if not ip:
 pathlib.Path(sys.argv[2]).write_text(ip)
 PY
 
-echo "[INFO] Agent registered with node id $node_id"
+echo "[INFO] Agent registered with node id $primary_node_id"
+echo "[INFO] Env-variable agent registered with node id $env_node_id"

src/agent/tests/scripts/05_verify_agent.sh

@@ -0,0 +1,60 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+REPO_ROOT="$(cd "$TEST_ROOT/.." && pwd)"
+VERIFY_SCRIPT="$REPO_ROOT/scripts/agent_deployment_verify.sh"
+ENV_NODE_ID_FILE="$TEST_ROOT/tmp/node_id_host_abc"
+PRIMARY_CONTAINER="argus-agent-e2e"
+ENV_CONTAINER="argus-agent-env-e2e"
+PRIMARY_HOSTNAME="dev-e2euser-e2einst-pod-0"
+ENV_HOSTNAME="host_abc"
+
+if ! docker ps --format '{{.Names}}' | grep -q "^${PRIMARY_CONTAINER}$"; then
+  echo "[WARN] agent container not running; skip verification"
+  exit 0
+fi
+
+if docker exec -i "$PRIMARY_CONTAINER" bash -lc 'command -v curl >/dev/null && command -v jq >/dev/null'; then
+  echo "[INFO] curl/jq already installed in agent container"
+else
+  echo "[INFO] Installing curl/jq in agent container"
+  docker exec -i "$PRIMARY_CONTAINER" bash -lc 'apt-get update >/dev/null 2>&1 && apt-get install -y curl jq >/dev/null 2>&1' || true
+fi
+
+if [[ ! -f "$VERIFY_SCRIPT" ]]; then
+  echo "[ERROR] Verification script missing at $VERIFY_SCRIPT" >&2
+  exit 1
+fi
+
+run_verifier() {
+  local container="$1" hostname="$2"
+
+  if ! docker ps --format '{{.Names}}' | grep -q "^${container}$"; then
+    echo "[WARN] container $container not running; skip"
+    return
+  fi
+
+  if ! docker exec -i "$container" bash -lc 'command -v /usr/local/bin/agent_deployment_verify.sh >/dev/null'; then
+    echo "[ERROR] /usr/local/bin/agent_deployment_verify.sh missing in $container" >&2
+    exit 1
+  fi
+
+  echo "[INFO] Running verification for $hostname in $container"
+  docker exec -i "$container" env VERIFY_HOSTNAME="$hostname" /usr/local/bin/agent_deployment_verify.sh
+}
+
+run_verifier "$PRIMARY_CONTAINER" "$PRIMARY_HOSTNAME"
+
+if docker ps --format '{{.Names}}' | grep -q "^${ENV_CONTAINER}$"; then
+  if docker exec -i "$ENV_CONTAINER" bash -lc 'command -v curl >/dev/null && command -v jq >/dev/null'; then
+    echo "[INFO] curl/jq already installed in env agent container"
+  else
+    echo "[INFO] Installing curl/jq in env agent container"
+    docker exec -i "$ENV_CONTAINER" bash -lc 'apt-get update >/dev/null 2>&1 && apt-get install -y curl jq >/dev/null 2>&1' || true
+  fi
+  run_verifier "$ENV_CONTAINER" "$ENV_HOSTNAME"
+else
+  echo "[WARN] env-driven agent container not running; skip secondary verification"
+fi

src/agent/tests/scripts/06_assert_status_on_master.sh

@@ -6,6 +6,8 @@ TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
 TMP_ROOT="$TEST_ROOT/tmp"
 API_BASE="http://localhost:32300/api/v1/master"
 NODE_ID="$(cat "$TMP_ROOT/node_id")"
+ENV_NODE_ID="$(cat "$TMP_ROOT/node_id_host_abc")"
+ENV_HOSTNAME="host_abc"
 NODES_JSON="$TEST_ROOT/private/argus/metric/prometheus/nodes.json"
 
 success=false
@@ -41,13 +43,36 @@ if [[ ! -f "$NODES_JSON" ]]; then
   exit 1
 fi
 
-python3 - "$NODES_JSON" <<'PY'
+python3 - "$NODES_JSON" "$NODE_ID" "$ENV_NODE_ID" <<'PY'
 import json, sys
 with open(sys.argv[1]) as handle:
     nodes = json.load(handle)
-assert len(nodes) == 1, nodes
+
-entry = nodes[0]
+expected_primary = sys.argv[2]
-assert entry["node_id"], entry
+expected_env = sys.argv[3]
+
+ids = {entry.get("node_id") for entry in nodes}
+assert expected_primary in ids, nodes
+assert expected_env in ids, nodes
+assert len(nodes) >= 2, nodes
 PY
 
 echo "[INFO] Master reflects agent health and nodes.json entries"
+
+env_detail_file="$TMP_ROOT/env_agent_detail.json"
+curl -sS "$API_BASE/nodes/$ENV_NODE_ID" -o "$env_detail_file"
+python3 - "$env_detail_file" "$ENV_HOSTNAME" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+
+expected_name = sys.argv[2]
+
+assert node.get("name") == expected_name, node
+meta = node.get("meta_data", {})
+assert meta.get("env") == "prod", meta
+assert meta.get("user") == "ml", meta
+assert meta.get("instance") == "node-3", meta
+PY
+
+echo "[INFO] Env-variable agent reports expected metadata"

src/agent/tests/scripts/07_restart_agent_and_reregister.sh

@@ -6,10 +6,20 @@ TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
 TMP_ROOT="$TEST_ROOT/tmp"
 API_BASE="http://localhost:32300/api/v1/master"
 NODE_ID="$(cat "$TMP_ROOT/node_id")"
+ENV_NODE_ID_FILE="$TMP_ROOT/node_id_host_abc"
+if [[ ! -f "$ENV_NODE_ID_FILE" ]]; then
+  echo "[ERROR] Env agent node id file missing at $ENV_NODE_ID_FILE" >&2
+  exit 1
+fi
+
+ENV_NODE_ID="$(cat "$ENV_NODE_ID_FILE")"
 AGENT_HOSTNAME="dev-e2euser-e2einst-pod-0"
+ENV_AGENT_HOSTNAME="host_abc"
 NETWORK_NAME="tests_default"
 NEW_AGENT_IP="172.28.0.200"
+NEW_ENV_AGENT_IP="172.28.0.210"
 ENTRYPOINT_SCRIPT="$SCRIPT_DIR/agent_entrypoint.sh"
+VERIFY_SCRIPT="$TEST_ROOT/../scripts/agent_deployment_verify.sh"
 ENV_FILE="$TEST_ROOT/.env"
 
 # 中文提示：重启场景也需要同样的入口脚本，确保 DNS 注册逻辑一致
@@ -18,6 +28,11 @@ if [[ ! -f "$ENTRYPOINT_SCRIPT" ]]; then
   exit 1
 fi
 
+if [[ ! -f "$VERIFY_SCRIPT" ]]; then
+  echo "[ERROR] agent verification script missing at $VERIFY_SCRIPT" >&2
+  exit 1
+fi
+
 if [[ ! -f "$TMP_ROOT/agent_binary_path" ]]; then
   echo "[ERROR] Agent binary path missing; rerun bootstrap" >&2
   exit 1
@@ -74,15 +89,37 @@ if [[ "$prev_ip" != "$initial_ip" ]]; then
   exit 1
 fi
 
+env_before_file="$TMP_ROOT/env_before_restart.json"
+curl -sS "$API_BASE/nodes/$ENV_NODE_ID" -o "$env_before_file"
+env_prev_last_updated=$(python3 - "$env_before_file" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+print(node.get("last_updated", ""))
+PY
+)
+env_prev_ip=$(python3 - "$env_before_file" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+print(node["meta_data"].get("ip", ""))
+PY
+)
+
 pushd "$TEST_ROOT" >/dev/null
 compose rm -sf agent
+compose rm -sf agent_env
 popd >/dev/null
 
 docker container rm -f argus-agent-e2e >/dev/null 2>&1 || true
+docker container rm -f argus-agent-env-e2e >/dev/null 2>&1 || true
 
 AGENT_DIR="$TEST_ROOT/private/argus/agent/$AGENT_HOSTNAME"
 HEALTH_DIR="$TEST_ROOT/private/argus/agent/$AGENT_HOSTNAME/health"
 
+ENV_AGENT_DIR="$TEST_ROOT/private/argus/agent/$ENV_AGENT_HOSTNAME"
+ENV_HEALTH_DIR="$TEST_ROOT/private/argus/agent/$ENV_AGENT_HOSTNAME/health"
+
 # 先以 sleep 方式启动容器，确保我们掌握注册时的网络状态
 if ! docker run -d \
   --name argus-agent-e2e \
@@ -94,6 +131,7 @@ if ! docker run -d \
   -v "$TEST_ROOT/private/argus/etc:/private/argus/etc" \
   -v "$AGENT_BINARY:/usr/local/bin/argus-agent:ro" \
   -v "$ENTRYPOINT_SCRIPT:/usr/local/bin/agent-entrypoint.sh:ro" \
+  -v "$VERIFY_SCRIPT:/usr/local/bin/agent_deployment_verify.sh:ro" \
   -e MASTER_ENDPOINT=http://master.argus.com:3000 \
   -e REPORT_INTERVAL_SECONDS=2 \
   -e ARGUS_BUILD_UID="$AGENT_UID" \
@@ -141,3 +179,76 @@ if [[ "$success" != true ]]; then
 fi
 
 echo "[INFO] Agent restart produced successful re-registration with IP change"
+
+# ---- Restart env-driven agent without metadata environment variables ----
+
+if [[ ! -d "$ENV_AGENT_DIR" ]]; then
+  echo "[ERROR] Env agent data dir missing at $ENV_AGENT_DIR" >&2
+  exit 1
+fi
+
+if [[ ! -d "$ENV_HEALTH_DIR" ]]; then
+  mkdir -p "$ENV_HEALTH_DIR"
+fi
+
+if ! docker run -d \
+  --name argus-agent-env-e2e \
+  --hostname "$ENV_AGENT_HOSTNAME" \
+  --network "$NETWORK_NAME" \
+  --ip "$NEW_ENV_AGENT_IP" \
+  -v "$ENV_AGENT_DIR:/private/argus/agent/$ENV_AGENT_HOSTNAME" \
+  -v "$ENV_HEALTH_DIR:/private/argus/agent/$ENV_AGENT_HOSTNAME/health" \
+  -v "$TEST_ROOT/private/argus/etc:/private/argus/etc" \
+  -v "$AGENT_BINARY:/usr/local/bin/argus-agent:ro" \
+  -v "$ENTRYPOINT_SCRIPT:/usr/local/bin/agent-entrypoint.sh:ro" \
+  -v "$VERIFY_SCRIPT:/usr/local/bin/agent_deployment_verify.sh:ro" \
+  -e MASTER_ENDPOINT=http://master.argus.com:3000 \
+  -e REPORT_INTERVAL_SECONDS=2 \
+  -e ARGUS_BUILD_UID="$AGENT_UID" \
+  -e ARGUS_BUILD_GID="$AGENT_GID" \
+  --entrypoint /usr/local/bin/agent-entrypoint.sh \
+  ubuntu:22.04 >/dev/null; then
+  echo "[ERROR] Failed to start env-driven agent container without metadata env" >&2
+  exit 1
+fi
+
+env_success=false
+env_detail_file="$TMP_ROOT/env_post_restart.json"
+for _ in {1..20}; do
+  sleep 3
+  if ! curl -sS "$API_BASE/nodes/$ENV_NODE_ID" -o "$env_detail_file"; then
+    continue
+  fi
+  if python3 - "$env_detail_file" "$env_prev_last_updated" "$ENV_NODE_ID" "$env_prev_ip" "$NEW_ENV_AGENT_IP" <<'PY'
+import json, sys
+with open(sys.argv[1]) as handle:
+    node = json.load(handle)
+prev_last_updated = sys.argv[2]
+expected_id = sys.argv[3]
+old_ip = sys.argv[4]
+expected_ip = sys.argv[5]
+last_updated = node.get("last_updated")
+current_ip = node["meta_data"].get("ip")
+meta = node.get("meta_data", {})
+assert node["id"] == expected_id
+if current_ip != expected_ip:
+    raise SystemExit(1)
+if current_ip == old_ip:
+    raise SystemExit(1)
+if not last_updated or last_updated == prev_last_updated:
+    raise SystemExit(1)
+if meta.get("env") != "prod" or meta.get("user") != "ml" or meta.get("instance") != "node-3":
+    raise SystemExit(1)
+PY
+  then
+    env_success=true
+    break
+  fi
+done
+
+if [[ "$env_success" != true ]]; then
+  echo "[ERROR] Env-driven agent did not reuse persisted metadata after restart" >&2
+  exit 1
+fi
+
+echo "[INFO] Env-driven agent restart succeeded with persisted metadata"

src/agent/tests/scripts/08_down.sh

@@ -13,7 +13,7 @@ compose() {
   fi
 }
 
-docker container rm -f argus-agent-e2e >/dev/null 2>&1 || true
+docker container rm -f argus-agent-e2e argus-agent-env-e2e >/dev/null 2>&1 || true
 
 pushd "$TEST_ROOT" >/dev/null
 compose down --remove-orphans

src/agent/tests/scripts/08_verify_agent.sh

@@ -1,26 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-TEST_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
-VERIFY_SCRIPT="$(cd "$TEST_ROOT/.." && pwd)/scripts/agent_deployment_verify.sh"
-
-if ! docker ps --format '{{.Names}}' | grep -q '^argus-agent-e2e$'; then
-  echo "[WARN] agent container not running; skip verification"
-  exit 0
-fi
-
-if docker exec -i argus-agent-e2e bash -lc 'command -v curl >/dev/null && command -v jq >/dev/null'; then
-  echo "[INFO] curl/jq already installed in agent container"
-else
-  echo "[INFO] Installing curl/jq in agent container"
-  docker exec -i argus-agent-e2e bash -lc 'apt-get update >/dev/null 2>&1 && apt-get install -y curl jq >/dev/null 2>&1' || true
-fi
-
-if docker exec -i argus-agent-e2e bash -lc 'command -v /usr/local/bin/agent_deployment_verify.sh >/dev/null'; then
-  docker exec -i argus-agent-e2e /usr/local/bin/agent_deployment_verify.sh
-elif [[ -x "$VERIFY_SCRIPT" ]]; then
-  docker exec -i argus-agent-e2e "$VERIFY_SCRIPT"
-else
-  echo "[WARN] agent_deployment_verify.sh not found"
-fi

src/agent/tests/test_config_metadata.py

@@ -0,0 +1,151 @@
+from __future__ import annotations
+
+import os
+import unittest
+from contextlib import contextmanager
+from unittest.mock import patch
+
+from app.config import AgentConfig, load_config
+
+
+@contextmanager
+def temp_env(**overrides: str | None):
+    originals: dict[str, str | None] = {}
+    try:
+        for key, value in overrides.items():
+            originals[key] = os.environ.get(key)
+            if value is None:
+                os.environ.pop(key, None)
+            else:
+                os.environ[key] = value
+        yield
+    finally:
+        for key, original in originals.items():
+            if original is None:
+                os.environ.pop(key, None)
+            else:
+                os.environ[key] = original
+
+
+class LoadConfigMetadataTests(unittest.TestCase):
+    @patch("app.config.Path.mkdir")
+    def test_metadata_from_environment_variables(self, mock_mkdir):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="dev-user-one-pod",
+            AGENT_ENV="prod",
+            AGENT_USER="ops",
+            AGENT_INSTANCE="node-1",
+        ):
+            config = load_config()
+
+        self.assertEqual(config.environment, "prod")
+        self.assertEqual(config.user, "ops")
+        self.assertEqual(config.instance, "node-1")
+        mock_mkdir.assert_called()
+
+    @patch("app.config.Path.mkdir")
+    def test_metadata_falls_back_to_hostname(self, mock_mkdir):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="qa-team-abc-pod-2",
+            AGENT_ENV=None,
+            AGENT_USER=None,
+            AGENT_INSTANCE=None,
+        ):
+            config = load_config()
+
+        self.assertEqual(config.environment, "qa")
+        self.assertEqual(config.user, "team")
+        self.assertEqual(config.instance, "abc")
+        mock_mkdir.assert_called()
+
+    @patch("app.config._load_metadata_from_state", return_value=("prod", "ops", "node-1"))
+    @patch("app.config.Path.mkdir")
+    def test_metadata_from_node_state(self, mock_mkdir, mock_state):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="host_abc",
+            AGENT_ENV=None,
+            AGENT_USER=None,
+            AGENT_INSTANCE=None,
+        ):
+            config = load_config()
+
+        self.assertEqual(config.environment, "prod")
+        self.assertEqual(config.user, "ops")
+        self.assertEqual(config.instance, "node-1")
+        mock_state.assert_called_once()
+        mock_mkdir.assert_called()
+
+    @patch("app.config.Path.mkdir")
+    def test_partial_environment_variables_fallback(self, mock_mkdir):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="stage-ml-001-node",
+            AGENT_ENV="prod",
+            AGENT_USER=None,
+            AGENT_INSTANCE=None,
+        ):
+            config = load_config()
+
+        self.assertEqual(config.environment, "stage")
+        self.assertEqual(config.user, "ml")
+        self.assertEqual(config.instance, "001")
+        mock_mkdir.assert_called()
+
+    @patch("app.config.Path.mkdir")
+    def test_invalid_hostname_raises_error(self, mock_mkdir):
+        with temp_env(
+            MASTER_ENDPOINT="http://master.local",
+            AGENT_HOSTNAME="invalidhostname",
+            AGENT_ENV=None,
+            AGENT_USER=None,
+            AGENT_INSTANCE=None,
+        ):
+            with self.assertRaises(ValueError):
+                load_config()
+
+        mock_mkdir.assert_not_called()
+
+
+class CollectMetadataTests(unittest.TestCase):
+    @patch("app.collector._detect_ip_address", return_value="127.0.0.1")
+    @patch("app.collector._detect_gpu_count", return_value=0)
+    @patch("app.collector._detect_memory_bytes", return_value=1024)
+    @patch("app.collector._detect_cpu_count", return_value=8)
+    def test_collect_metadata_uses_config_fields(
+        self,
+        mock_cpu,
+        mock_memory,
+        mock_gpu,
+        mock_ip,
+    ):
+        config = AgentConfig(
+            hostname="dev-user-001-pod",
+            environment="prod",
+            user="ops",
+            instance="node-1",
+            node_file="/tmp/node.json",
+            version="1.0.0",
+            master_endpoint="http://master.local",
+            report_interval_seconds=60,
+            health_dir="/tmp/health",
+        )
+
+        from app.collector import collect_metadata
+
+        metadata = collect_metadata(config)
+
+        self.assertEqual(metadata["env"], "prod")
+        self.assertEqual(metadata["user"], "ops")
+        self.assertEqual(metadata["instance"], "node-1")
+        self.assertEqual(metadata["hostname"], "dev-user-001-pod")
+        self.assertEqual(metadata["ip"], "127.0.0.1")
+        self.assertEqual(metadata["cpu_number"], 8)
+        self.assertEqual(metadata["memory_in_bytes"], 1024)
+        self.assertEqual(metadata["gpu_number"], 0)
+
+
+if __name__ == "__main__":
+    unittest.main()

src/bind/build/dns-monitor.sh

@@ -17,6 +17,9 @@ log_message() {
 
 log_message "DNS监控脚本启动"
 
+log_message "删除DNS备份文件（如果存在）"
+rm -f $DNS_BACKUP
+
 while true; do
     if [ -f "$DNS_CONF" ]; then
         if [ -f "$DNS_BACKUP" ]; then

src/log/kibana/build/kibana-post-start.sh

@@ -2,7 +2,7 @@
 set -euo pipefail
 
 ES_HOST="${ELASTICSEARCH_HOSTS:-http://es:9200}"
-KB_HOST="http://localhost:5601"
+KB_HOST="${KB_HOST:-http://127.0.0.1:5601}"
 
 echo "[INFO] Starting Kibana post-start configuration..."
 
@@ -83,50 +83,37 @@ fix_replicas_idempotent() {
 }
 
 # 幂等创建数据视图
+create_or_ensure_data_view() {
+    local name="$1"
+    local title="$2"
+
+    local list_response
+    list_response=$(curl -fsS "$KB_HOST/api/data_views" -H 'kbn-xsrf: true' 2>/dev/null || echo "")
+
+    if [ -z "$list_response" ]; then
+        echo "[WARN] Failed to list data views, skipping creation check for $title"
+        return
+    fi
+
+    if echo "$list_response" | grep -Fq "\"title\":\"$title\""; then
+        echo "[INFO] Data view $title already exists, skipping"
+        return
+    fi
+
+    echo "[INFO] Creating data view for $title indices (allowNoIndex)"
+
+    curl -fsS -X POST "$KB_HOST/api/data_views/data_view?allowNoIndex=true" \
+        -H 'kbn-xsrf: true' \
+        -H 'Content-Type: application/json' \
+        -d "{\"data_view\":{\"name\":\"$name\",\"title\":\"$title\",\"timeFieldName\":\"@timestamp\",\"allowNoIndex\":true}}" \
+        >/dev/null && echo "[OK] Created $name data view" || echo "[WARN] Failed to create $name data view"
+}
+
 create_data_views_idempotent() {
     echo "[INFO] Checking and creating data views..."
 
-    # 检查是否存在匹配的索引
+    create_or_ensure_data_view "train" "train-*"
-    local train_indices=$(curl -s "$ES_HOST/_cat/indices/train-*?h=index" 2>/dev/null | wc -l || echo "0")
+    create_or_ensure_data_view "infer" "infer-*"
-    local infer_indices=$(curl -s "$ES_HOST/_cat/indices/infer-*?h=index" 2>/dev/null | wc -l || echo "0")
-
-    # 创建 train 数据视图
-    if [ "$train_indices" -gt 0 ]; then
-        # 检查数据视图是否已存在
-        local train_exists=$(curl -s "$KB_HOST/api/data_views" -H 'kbn-xsrf: true' 2>/dev/null | grep '"title":"train-\*"' | wc -l )
-
-        if [ "$train_exists" -eq 0 ]; then
-            echo "[INFO] Creating data view for train-* indices"
-            curl -fsS -X POST "$KB_HOST/api/data_views/data_view" \
-                -H 'kbn-xsrf: true' \
-                -H 'Content-Type: application/json' \
-                -d '{"data_view":{"name":"train","title":"train-*","timeFieldName":"@timestamp"}}' \
-                >/dev/null && echo "[OK] Created train data view" || echo "[WARN] Failed to create train data view"
-        else
-            echo "[INFO] Train data view already exists, skipping"
-        fi
-    else
-        echo "[INFO] No train-* indices found, skipping train data view creation"
-    fi
-
-    # 创建 infer 数据视图
-    if [ "$infer_indices" -gt 0 ]; then
-        # 检查数据视图是否已存在
-        local infer_exists=$(curl -s "$KB_HOST/api/data_views" -H 'kbn-xsrf: true' 2>/dev/null | grep '"title":"infer-\*"' | wc -l )
-
-        if [ "$infer_exists" -eq 0 ]; then
-            echo "[INFO] Creating data view for infer-* indices"
-            curl -fsS -X POST "$KB_HOST/api/data_views/data_view" \
-                -H 'kbn-xsrf: true' \
-                -H 'Content-Type: application/json' \
-                -d '{"data_view":{"name":"infer","title":"infer-*","timeFieldName":"@timestamp"}}' \
-                >/dev/null && echo "[OK] Created infer data view" || echo "[WARN] Failed to create infer data view"
-        else
-            echo "[INFO] Infer data view already exists, skipping"
-        fi
-    else
-        echo "[INFO] No infer-* indices found, skipping infer data view creation"
-    fi
 }
 
 # 主逻辑

src/log/tests/scripts/e2e_test.sh

@@ -115,20 +115,32 @@ show_step "Health" "Check service health"
 echo "[INFO] Checking service health..."
 
 # 检查 Elasticsearch 健康状态
+health_check_ok=1
 es_health=$(curl -s "http://localhost:9200/_cluster/health" | grep -o '"status":"[^"]*"' | cut -d'"' -f4)
 if [ "$es_health" = "green" ] || [ "$es_health" = "yellow" ]; then
     echo "✅ Elasticsearch health: $es_health"
 else
     echo "❌ Elasticsearch health: $es_health"
+    health_check_ok=0
 fi
 
 # 检查 Kibana 状态
 if curl -fs "http://localhost:5601/api/status" >/dev/null 2>&1; then
     kb_status="available"
     echo "✅ Kibana status: $kb_status"
+
+    data_views_json=$(curl -fs "http://localhost:5601/api/data_views" -H 'kbn-xsrf: true' 2>/dev/null || true)
+    if echo "$data_views_json" | grep -F '"title":"train-*"' >/dev/null 2>&1 && \
+       echo "$data_views_json" | grep -F '"title":"infer-*"' >/dev/null 2>&1; then
+        echo "✅ Kibana data views: train-* and infer-* present"
+    else
+        echo "❌ Kibana data views missing: train-* or infer-*"
+        health_check_ok=0
+    fi
 else
     kb_status="unavailable"
     echo "⚠️  Kibana status: $kb_status"
+    health_check_ok=0
 fi
 
 # 检查 Fluent-Bit 指标
@@ -139,6 +151,13 @@ if [ "$fb_host01_uptime" -gt 0 ] && [ "$fb_host02_uptime" -gt 0 ]; then
     echo "✅ Fluent-Bit services: host01 uptime=${fb_host01_uptime}s, host02 uptime=${fb_host02_uptime}s"
 else
     echo "⚠️  Fluent-Bit services: host01 uptime=${fb_host01_uptime}s, host02 uptime=${fb_host02_uptime}s"
+    health_check_ok=0
+fi
+
+if [ "$health_check_ok" -eq 1 ]; then
+    true
+else
+    false
 fi
 
 verify_step "Service health check"

src/master/scripts/build_images.sh

@@ -3,12 +3,13 @@ set -euo pipefail
 
 usage() {
   cat >&2 <<'USAGE'
-Usage: $0 [--intranet] [--offline] [--tag <image_tag>]
+Usage: $0 [--intranet] [--offline] [--tag <image_tag>] [--no-cache]
 
 Options:
   --intranet           使用指定的 PyPI 镜像源（默认清华镜像）。
   --offline            完全离线构建，依赖 offline_wheels/ 目录中的离线依赖包。
   --tag <image_tag>    自定义镜像标签，默认 argus-master:latest。
+  --no-cache           不使用 Docker 构建缓存。
 USAGE
 }
 
@@ -19,6 +20,7 @@ IMAGE_TAG="${IMAGE_TAG:-argus-master:latest}"
 DOCKERFILE="src/master/Dockerfile"
 BUILD_ARGS=()
 OFFLINE_MODE=0
+NO_CACHE=0
 
 source "$PROJECT_ROOT/scripts/common/build_user.sh"
 load_build_user
@@ -45,6 +47,11 @@ while [[ "$#" -gt 0 ]]; do
       IMAGE_TAG="$2"
       shift 2
       ;;
+    --no-cache)
+      NO_CACHE=1
+      BUILD_ARGS+=("--no-cache")
+      shift
+      ;;
     -h|--help)
       usage
       exit 0

src/metric/client-plugins/all-in-one-demo/scripts/check_health.sh

@@ -4,6 +4,15 @@
 
 set -e
 
+# PID 文件检测，防止重复执行
+PIDFILE="/var/run/check_health.pid"
+if [ -f "$PIDFILE" ] && kill -0 $(cat "$PIDFILE") 2>/dev/null; then
+    echo "健康检查脚本已在运行中，跳过本次执行" >&2
+    exit 0
+fi
+echo $$ > "$PIDFILE"
+trap "rm -f $PIDFILE" EXIT
+
 # 获取脚本所在目录
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 HEALTH_LOG_FILE="$SCRIPT_DIR/.health_log"

src/metric/client-plugins/all-in-one-demo/scripts/install_artifact.sh

@@ -200,22 +200,22 @@ parse_version_info() {
         VERSION=$(grep '"version"' "$VERSION_FILE_PATH" | sed 's/.*"version": *"\([^"]*\)".*/\1/')
         BUILD_TIME=$(grep '"build_time"' "$VERSION_FILE_PATH" | sed 's/.*"build_time": *"\([^"]*\)".*/\1/')
         
-        # 解析 artifact_list
+        # 解析 artifact_list（跳过字段名本身）
-        grep -A 100 '"artifact_list"' "$VERSION_FILE_PATH" | grep -E '^\s*"[^"]+":\s*"[^"]+"' | while read line; do
+        grep -A 100 '"artifact_list"' "$VERSION_FILE_PATH" | grep -v '"artifact_list"' | grep -E '^\s*"[^"]+":\s*"[^"]+"' | while read line; do
             component=$(echo "$line" | sed 's/.*"\([^"]*\)":\s*"[^"]*".*/\1/')
             version=$(echo "$line" | sed 's/.*"[^"]*":\s*"\([^"]*\)".*/\1/')
             echo "$component:$version" >> "$TEMP_DIR/components.txt"
         done
         
-        # 解析 checksums
+        # 解析 checksums（跳过字段名本身）
-        grep -A 100 '"checksums"' "$VERSION_FILE_PATH" | grep -E '^\s*"[^"]+":\s*"[^"]+"' | while read line; do
+        grep -A 100 '"checksums"' "$VERSION_FILE_PATH" | grep -v '"checksums"' | grep -E '^\s*"[^"]+":\s*"[^"]+"' | while read line; do
             component=$(echo "$line" | sed 's/.*"\([^"]*\)":\s*"[^"]*".*/\1/')
             checksum=$(echo "$line" | sed 's/.*"[^"]*":\s*"\([^"]*\)".*/\1/')
             echo "$component:$checksum" >> "$TEMP_DIR/checksums.txt"
         done
         
-        # 解析 install_order
+        # 解析 install_order（跳过字段名本身，只取数组元素）
-        grep -A 100 '"install_order"' "$VERSION_FILE_PATH" | grep -E '^\s*"[^"]+"' | while read line; do
+        grep -A 100 '"install_order"' "$VERSION_FILE_PATH" | grep -v '"install_order"' | grep -E '^\s*"[^"]+"' | while read line; do
             component=$(echo "$line" | sed 's/.*"\([^"]*\)".*/\1/')
             echo "$component" >> "$TEMP_DIR/install_order.txt"
         done
@@ -317,85 +317,152 @@ create_install_dirs() {
     log_success "安装目录创建完成: $INSTALL_DIR"
 }
 
+# 获取系统版本
+get_system_version() {
+    if [[ ! -f /etc/os-release ]]; then
+        log_error "无法检测操作系统版本"
+        return 1
+    fi
+    
+    source /etc/os-release
+    
+    # 提取主版本号
+    case "$VERSION_ID" in
+        "20.04")
+            echo "ubuntu20"
+            ;;
+        "22.04")
+            echo "ubuntu22"
+            ;;
+        *)
+            log_warning "未识别的Ubuntu版本: $VERSION_ID，尝试使用ubuntu22"
+            echo "ubuntu22"
+            ;;
+    esac
+}
+
 # 安装系统依赖包
 install_system_deps() {
     log_info "检查系统依赖包..."
-    
+
     local script_dir="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
     local deps_dir="$script_dir/deps"
-    
+
     # 检查deps目录是否存在
     if [[ ! -d "$deps_dir" ]]; then
         log_info "deps 目录不存在，跳过系统依赖包安装"
         return 0
     fi
+
+    # 获取系统版本对应的依赖目录
+    local system_version=$(get_system_version)
+    local version_deps_dir="$deps_dir/$system_version"
     
-    # 检查是否有tar.gz文件
+    log_info "检测到系统版本: $system_version"
-    local deps_count=$(find "$deps_dir" -name "*.tar.gz" | wc -l)
+    
-    if [[ $deps_count -eq 0 ]]; then
+    # 检查版本特定的依赖目录是否存在
-        log_info "deps 目录中没有 tar.gz 文件，跳过系统依赖包安装"
+    if [[ ! -d "$version_deps_dir" ]]; then
-        return 0
+        log_warning "未找到 $system_version 版本的依赖目录: $version_deps_dir"
+        # 回退到旧的逻辑，检查根deps目录
+        local deps_count=$(find "$deps_dir" -name "*.tar.gz" | wc -l)
+        if [[ $deps_count -eq 0 ]]; then
+            log_info "deps 目录中没有 tar.gz 文件，跳过系统依赖包安装"
+            return 0
+        fi
+        version_deps_dir="$deps_dir"
+    else
+        # 检查版本目录中是否有tar.gz文件
+        local deps_count=$(find "$version_deps_dir" -name "*.tar.gz" | wc -l)
+        if [[ $deps_count -eq 0 ]]; then
+            log_info "$system_version 版本目录中没有 tar.gz 文件，跳过系统依赖包安装"
+            return 0
+        fi
     fi
-    
+
-    log_info "找到 $deps_count 个系统依赖包，开始安装..."
+    log_info "找到 $system_version 版本的依赖包，开始安装..."
-    
+
     # 创建临时目录用于解压依赖包
-    local deps_temp_dir="$TEMP_DIR/deps"
+    local deps_temp_dir="${TEMP_DIR:-/tmp}/deps"
     mkdir -p "$deps_temp_dir"
-    
+
+    # 定义要检查的核心依赖
+    local CORE_DEPS=(jq cron curl)
+    local FAILED_DEPS=()
+
     # 处理每个tar.gz文件
-    find "$deps_dir" -name "*.tar.gz" | while read tar_file; do
+    find "$version_deps_dir" -name "*.tar.gz" | while read tar_file; do
         local tar_basename=$(basename "$tar_file")
         local extract_name="${tar_basename%.tar.gz}"
-        
+
         log_info "处理依赖包: $tar_basename"
-        
+
         # 解压到临时目录
         local extract_dir="$deps_temp_dir/$extract_name"
         mkdir -p "$extract_dir"
-        
+
         if tar -xzf "$tar_file" -C "$extract_dir" 2>/dev/null; then
             log_success "  $tar_basename 解压完成"
         else
             log_error "  $tar_basename 解压失败"
             continue
         fi
-        
+
         # 进入解压目录，查找deb包
-        cd "$extract_dir"
+        cd "$extract_dir" || continue
-        local deb_count=$(find . -name "*.deb" | wc -l)
+        local deb_files=(*.deb)
-        
+        if [[ ${#deb_files[@]} -gt 0 ]]; then
-        if [[ $deb_count -gt 0 ]]; then
+            log_info "  找到 ${#deb_files[@]} 个 deb 包，开始安装..."
-            log_info "  找到 $deb_count 个 deb 包，开始安装..."
+
-            
+            for deb in "${deb_files[@]}"; do
-            # 1. 先尝试安装所有deb包
+                local pkg_name
-            log_info "  第1步：批量安装deb包..."
+                pkg_name=$(dpkg-deb -f "$deb" Package 2>/dev/null)
-            if dpkg -i *.deb 2>/dev/null; then
+
-                log_success "  所有deb包安装成功"
+                # 如果已安装，则跳过
-            else
+                if dpkg -s "$pkg_name" &>/dev/null; then
-                log_warning "  部分deb包安装失败，可能存在依赖问题"
+                    log_success "  $pkg_name 已安装，跳过"
-                
+                    continue
-                # 2. 使用apt-get修复依赖
-                log_info "  第2步：修复依赖关系..."
-                if apt-get install -f -y; then
-                    log_success "  依赖关系修复完成"
-                else
-                    log_error "  依赖关系修复失败"
-                    # 继续处理其他包，不退出
                 fi
-            fi
+
+                # 尝试安装
+                log_info "  安装 $pkg_name..."
+                if DEBIAN_FRONTEND=noninteractive dpkg -i "$deb" &>/dev/null; then
+                    log_success "  $pkg_name 安装成功"
+                else
+                    log_warning "  $pkg_name 安装失败，尝试修复依赖..."
+                    if DEBIAN_FRONTEND=noninteractive apt-get install -f -y &>/dev/null; then
+                        if dpkg -s "$pkg_name" &>/dev/null; then
+                            log_success "  $pkg_name 修复安装成功"
+                        else
+                            log_error "  $pkg_name 仍未安装成功"
+                            FAILED_DEPS+=("$pkg_name")
+                        fi
+                    else
+                        log_error "  $pkg_name 自动修复失败"
+                        FAILED_DEPS+=("$pkg_name")
+                    fi
+                fi
+            done
         else
             log_info "  $tar_basename 中没有找到deb包，跳过"
         fi
-        
+
         # 返回到依赖临时目录
-        cd "$deps_temp_dir"
+        cd "$deps_temp_dir" || continue
     done
-    
+
     # 检查并启动 cron 服务
     start_cron_service
-    
+
-    log_success "系统依赖包安装完成"
+    # 总结安装结果
+    if [[ ${#FAILED_DEPS[@]} -gt 0 ]]; then
+        log_error "以下系统依赖未能成功安装，安装终止，请手动安装后重试："
+        for f in "${FAILED_DEPS[@]}"; do
+            echo "  - $f"
+        done
+        exit 1
+    else
+        log_success "系统依赖包安装完成，全部就绪"
+    fi
 }
 
 # 启动 cron 服务
@@ -637,6 +704,18 @@ EOF
     log_success "安装记录已创建: $install_record_file"
 }
 
+# 检查cron任务是否已存在
+check_cron_task_exists() {
+    local task_pattern="$1"
+    local temp_cron="$2"
+    
+    if grep -q "$task_pattern" "$temp_cron"; then
+        return 0  # 任务已存在
+    else
+        return 1  # 任务不存在
+    fi
+}
+
 # 设置健康检查定时任务
 setup_health_check_cron() {
     log_info "设置健康检查定时任务..."
@@ -661,7 +740,7 @@ setup_health_check_cron() {
     crontab -l 2>/dev/null > "$temp_cron" || touch "$temp_cron"
     
     # 检查并删除旧的健康检查任务
-    if grep -q "check_health.sh" "$temp_cron"; then
+    if check_cron_task_exists "check_health.sh" "$temp_cron"; then
         log_info "发现旧的健康检查定时任务，正在更新..."
         # 删除所有包含check_health.sh的行
         grep -v "check_health.sh" "$temp_cron" > "$temp_cron.new"
@@ -716,7 +795,7 @@ setup_dns_sync_cron() {
     crontab -l 2>/dev/null > "$temp_cron" || touch "$temp_cron"
     
     # 检查并删除旧的 DNS 同步任务
-    if grep -q "sync_dns.sh" "$temp_cron"; then
+    if check_cron_task_exists "sync_dns.sh" "$temp_cron"; then
         log_info "发现旧的 DNS 同步定时任务，正在更新..."
         # 删除所有包含sync_dns.sh的行
         grep -v "sync_dns.sh" "$temp_cron" > "$temp_cron.new"
@@ -724,16 +803,15 @@ setup_dns_sync_cron() {
         log_info "旧的 DNS 同步定时任务已删除"
     fi
     
-    # 添加新的定时任务（每30秒执行一次）
+    # 添加新的定时任务（每1分钟执行一次）
     # 直接使用版本目录中的 DNS 同步脚本
     echo "# Argus-Metrics DNS 同步定时任务" >> "$temp_cron"
     echo "* * * * * $sync_dns_script >> $INSTALL_DIR/.dns_sync.log 2>&1" >> "$temp_cron"
-    echo "* * * * * sleep 30; $sync_dns_script >> $INSTALL_DIR/.dns_sync.log 2>&1" >> "$temp_cron"
     
     # 安装新的crontab
     if crontab "$temp_cron"; then
         log_success "DNS 同步定时任务设置成功"
-        log_info "  执行频率: 每30秒"
+        log_info "  执行频率: 每1分钟"
         log_info "  日志文件: $INSTALL_DIR/.dns_sync.log"
         log_info "  查看定时任务: crontab -l"
         log_info "  删除定时任务: crontab -e"
@@ -771,7 +849,7 @@ setup_version_check_cron() {
     crontab -l > "$temp_cron" 2>/dev/null || touch "$temp_cron"
     
     # 检查是否已存在版本校验定时任务
-    if grep -q "check_version.sh" "$temp_cron"; then
+    if check_cron_task_exists "check_version.sh" "$temp_cron"; then
         log_info "发现旧的版本校验定时任务，正在更新..."
         # 删除所有包含check_version.sh的行
         grep -v "check_version.sh" "$temp_cron" > "$temp_cron.new"
@@ -824,7 +902,7 @@ setup_restart_cron() {
     crontab -l > "$temp_cron" 2>/dev/null || touch "$temp_cron"
     
     # 检查是否已存在自动重启定时任务
-    if grep -q "restart_unhealthy.sh" "$temp_cron"; then
+    if check_cron_task_exists "restart_unhealthy.sh" "$temp_cron"; then
         log_info "发现旧的自动重启定时任务，正在更新..."
         # 删除所有包含restart_unhealthy.sh的行
         grep -v "restart_unhealthy.sh" "$temp_cron" > "$temp_cron.new"
@@ -885,9 +963,9 @@ main() {
     check_system
     find_version_file
     create_install_dirs
-    parse_version_info
+    install_system_deps     
+    parse_version_info    
     verify_checksums
-    install_system_deps
     install_components
     copy_config_files
     create_install_record
@@ -895,6 +973,20 @@ main() {
     setup_dns_sync_cron
     setup_version_check_cron
     setup_restart_cron
+    
+    # 注释掉立即执行健康检查，避免与cron任务重复执行
+    # log_info "立即执行一次健康检查..."
+    # local check_health_script="$INSTALL_DIR/check_health.sh"
+    # if [[ -f "$check_health_script" ]]; then
+    #     if "$check_health_script" >> "$INSTALL_DIR/.health_check.log" 2>&1; then
+    #         log_success "健康检查执行完成"
+    #     else
+    #         log_warning "健康检查执行失败，请检查日志: $INSTALL_DIR/.health_check.log"
+    #     fi
+    # else
+    #     log_warning "健康检查脚本不存在: $check_health_script"
+    # fi
+    
     show_install_info
 }
 

src/metric/client-plugins/all-in-one-demo/scripts/publish_artifact.sh

@@ -29,26 +29,68 @@ log_error() {
 show_help() {
     echo "Argus-Metric Artifact 发布脚本"
     echo
-    echo "用法: $0 <版本号>"
+    echo "用法: $0 <版本号> [选项]"
     echo
     echo "参数:"
-    echo "  <版本号>    要发布的版本号，对应 artifact 目录中的版本"
+    echo "  <版本号>              要发布的版本号，对应 artifact 目录中的版本"
+    echo
+    echo "选项:"
+    echo "  --output-dir <路径>   指定输出目录 (默认: /private/argus/ftp/share/)"
+    echo "  --owner <uid:gid>     指定文件所有者 (默认: 2133:2015)"
+    echo "  -h, --help           显示此帮助信息"
     echo
     echo "示例:"
-    echo "  $0 1.20.0    # 发布 1.20.0 版本"
+    echo "  $0 1.20.0                                    # 使用默认配置发布"
+    echo "  $0 1.20.0 --output-dir /tmp/publish          # 指定输出目录"
+    echo "  $0 1.20.0 --owner 1000:1000                  # 指定文件所有者"
+    echo "  $0 1.20.0 --output-dir /srv/ftp --owner root:root  # 同时指定两者"
     echo
 }
 
-# 检查参数
+# 默认配置
-if [[ $# -ne 1 ]]; then
+DEFAULT_PUBLISH_DIR="/private/argus/ftp/share/"
+DEFAULT_OWNER="2133:2015"
+
+# 解析参数
+VERSION=""
+PUBLISH_DIR="$DEFAULT_PUBLISH_DIR"
+OWNER="$DEFAULT_OWNER"
+
+while [[ $# -gt 0 ]]; do
+    case $1 in
+        -h|--help)
+            show_help
+            exit 0
+            ;;
+        --output-dir)
+            PUBLISH_DIR="$2"
+            shift 2
+            ;;
+        --owner)
+            OWNER="$2"
+            shift 2
+            ;;
+        *)
+            if [[ -z "$VERSION" ]]; then
+                VERSION="$1"
+                shift
+            else
+                log_error "未知参数: $1"
+                show_help
+                exit 1
+            fi
+            ;;
+    esac
+done
+
+# 检查版本号是否提供
+if [[ -z "$VERSION" ]]; then
     log_error "请提供版本号参数"
     show_help
     exit 1
 fi
 
-VERSION="$1"
 ARTIFACT_DIR="artifact/$VERSION"
-PUBLISH_DIR="/Users/sundapeng/Project/nlp/aiops/client-plugins/all-in-one/publish/"
 
 # 检查版本目录是否存在
 if [[ ! -d "$ARTIFACT_DIR" ]]; then
@@ -57,10 +99,12 @@ if [[ ! -d "$ARTIFACT_DIR" ]]; then
 fi
 
 log_info "开始发布版本: $VERSION"
+log_info "输出目录: $PUBLISH_DIR"
+log_info "文件所有者: $OWNER"
 
 # 确保发布目录存在
 log_info "确保发布目录存在: $PUBLISH_DIR"
-mkdir -p "$PUBLISH_DIR"
+sudo mkdir -p "$PUBLISH_DIR"
 
 # 创建临时目录用于打包
 TEMP_PACKAGE_DIR="/tmp/argus-metric-package-$$"
@@ -164,20 +208,26 @@ fi
 TAR_NAME="argus-metric_$(echo $VERSION | tr '.' '_').tar.gz"
 log_info "创建发布包: $TAR_NAME"
 cd "$TEMP_PACKAGE_DIR"
-tar -czf "$PUBLISH_DIR/$TAR_NAME" *
+sudo tar -czf "$PUBLISH_DIR/$TAR_NAME" *
 cd - > /dev/null
 
+# 设置文件所有者
+log_info "设置文件所有者为: $OWNER"
+sudo chown "$OWNER" "$PUBLISH_DIR/$TAR_NAME"
+
 # 清理临时目录
 rm -rf "$TEMP_PACKAGE_DIR"
 
 # 更新 LATEST_VERSION 文件
 log_info "更新 LATEST_VERSION 文件..."
-echo "$VERSION" > "$PUBLISH_DIR/LATEST_VERSION"
+echo "$VERSION" | sudo tee "$PUBLISH_DIR/LATEST_VERSION" > /dev/null
+sudo chown "$OWNER" "$PUBLISH_DIR/LATEST_VERSION"
 
 # 复制 DNS 配置文件到发布目录根目录（直接从 config 目录复制）
 if [[ -f "config/dns.conf" ]]; then
     log_info "复制 DNS 配置文件到发布目录根目录..."
-    cp "config/dns.conf" "$PUBLISH_DIR/"
+    sudo cp "config/dns.conf" "$PUBLISH_DIR/"
+    sudo chown "$OWNER" "$PUBLISH_DIR/dns.conf"
     log_success "DNS 配置文件复制完成: $PUBLISH_DIR/dns.conf"
 else
     log_warning "未找到 config/dns.conf 文件，跳过 DNS 配置文件复制"
@@ -186,7 +236,8 @@ fi
 # 复制 setup.sh 到发布目录
 if [[ -f "scripts/setup.sh" ]]; then
     log_info "复制 setup.sh 到发布目录..."
-    cp "scripts/setup.sh" "$PUBLISH_DIR/"
+    sudo cp "scripts/setup.sh" "$PUBLISH_DIR/"
+    sudo chown "$OWNER" "$PUBLISH_DIR/setup.sh"
 fi
 
 # 显示发布结果

src/metric/client-plugins/all-in-one-demo/scripts/restart_unhealthy.sh

@@ -2,6 +2,15 @@
 
 # 此脚本会检查各组件的健康状态，并重启不健康的组件
 
+# PID 文件检测，防止重复执行
+PIDFILE="/var/run/restart_unhealthy.pid"
+if [ -f "$PIDFILE" ] && kill -0 $(cat "$PIDFILE") 2>/dev/null; then
+    echo "自动重启脚本已在运行中，跳过本次执行" >&2
+    exit 0
+fi
+echo $$ > "$PIDFILE"
+trap "rm -f $PIDFILE" EXIT
+
 # 获取脚本所在目录
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 INSTALL_RECORD_FILE="$SCRIPT_DIR/.install_record"

src/metric/client-plugins/all-in-one-demo/scripts/sync_dns.sh

@@ -1,257 +1,143 @@
 #!/bin/bash
-
-# DNS 同步脚本
-# 比较 FTP 根目录的 dns.conf 和本地的 dns.conf，如果有变化则同步到 /etc/resolv.conf
-
 set -e
 
-# 颜色定义
+# 颜色
-RED='\033[0;31m'
+RED='\033[0;31m'; GREEN='\033[0;32m'; YELLOW='\033[1;33m'; BLUE='\033[0;34m'; NC='\033[0m'
-GREEN='\033[0;32m'
-YELLOW='\033[1;33m'
-BLUE='\033[0;34m'
-NC='\033[0m' # No Color
 
-# 日志函数 - 输出到 stderr 避免影响函数返回值
+# 日志函数
-log_info() {
+log_info() { echo -e "${BLUE}[INFO]${NC} $1" >&2; }
-    echo -e "${BLUE}[INFO]${NC} $1" >&2
+log_success() { echo -e "${GREEN}[SUCCESS]${NC} $1" >&2; }
-}
+log_warning() { echo -e "${YELLOW}[WARNING]${NC} $1" >&2; }
+log_error() { echo -e "${RED}[ERROR]${NC} $1" >&2; }
 
-log_success() {
-    echo -e "${GREEN}[SUCCESS]${NC} $1" >&2
-}
-
-log_warning() {
-    echo -e "${YELLOW}[WARNING]${NC} $1" >&2
-}
-
-log_error() {
-    echo -e "${RED}[ERROR]${NC} $1" >&2
-}
-
-# 获取脚本所在目录
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 LOCAL_DNS_CONF="/opt/argus-metric/dns.conf"
-REMOTE_DNS_CONF_URL=""
 RESOLV_CONF="/etc/resolv.conf"
+ALT_RESOLV_CONF="/run/resolv.conf"
 LOG_FILE="/opt/argus-metric/.dns_sync.log"
+REMOTE_DNS_CONF_URL=""
 
-# 从环境变量或配置文件获取 FTP 服务器信息
+# 获取 FTP 配置
 get_ftp_config() {
-    # 优先从环境变量获取配置
     log_info "获取 FTP 配置信息..."
-    
-    # 如果环境变量中没有设置，则尝试从配置文件读取
     if [[ -z "$FTP_SERVER" || -z "$FTP_USER" || -z "$FTP_PASSWORD" ]]; then
-        local config_file="$SCRIPT_DIR/config.env"
+        [[ -f "$SCRIPT_DIR/config.env" ]] && source "$SCRIPT_DIR/config.env"
-        if [[ -f "$config_file" ]]; then
-            log_info "从配置文件读取 FTP 配置: $config_file"
-            source "$config_file"
-        fi
-    else
-        log_info "使用环境变量中的 FTP 配置"
     fi
-    
-    # 设置默认值（如果环境变量和配置文件都没有设置）
     FTP_SERVER="${FTP_SERVER:-localhost}"
     FTP_USER="${FTP_USER:-ftpuser}"
     FTP_PASSWORD="${FTP_PASSWORD:-ZGClab1234!}"
-    
-    # 构建远程 DNS 配置文件 URL
     REMOTE_DNS_CONF_URL="ftp://${FTP_USER}:${FTP_PASSWORD}@${FTP_SERVER}/dns.conf"
-    
-    log_info "FTP 配置来源: ${FTP_CONFIG_SOURCE:-环境变量/配置文件}"
 }
 
-# 下载远程 DNS 配置文件
+# 下载远程 dns.conf
 download_remote_dns_conf() {
-    local temp_file="/tmp/dns.conf.remote.$$"
+    local tmp="/tmp/dns.remote.$$"
-    
-    log_info "从 FTP 服务器下载 DNS 配置文件..."
-    log_info "远程地址: $REMOTE_DNS_CONF_URL"
-    log_info "FTP 服务器: $FTP_SERVER"
-    log_info "FTP 用户: $FTP_USER"
-    
-    # 先测试 FTP 连接
     log_info "测试 FTP 连接..."
-    if curl -u "${FTP_USER}:${FTP_PASSWORD}" -sfI "ftp://${FTP_SERVER}/" >/dev/null 2>&1; then
+    if ! curl -u "${FTP_USER}:${FTP_PASSWORD}" -sfI "ftp://${FTP_SERVER}/" >/dev/null; then
-        log_success "FTP 服务器连接成功"
+        log_error "无法连接到 FTP 服务器: $FTP_SERVER"; return 1
-    else
-        log_error "无法连接到 FTP 服务器: $FTP_SERVER"
-        log_error "请检查:"
-        log_error "  1. FTP 服务器是否运行"
-        log_error "  2. 网络连接是否正常"
-        log_error "  3. 服务器地址是否正确"
-        return 1
     fi
-    
+    if ! curl -u "${FTP_USER}:${FTP_PASSWORD}" -sf "ftp://${FTP_SERVER}/dns.conf" -o "$tmp" 2>/dev/null; then
-    # 测试 dns.conf 文件是否存在
+        log_error "下载 dns.conf 失败"; rm -f "$tmp"; return 1
-    log_info "检查远程 dns.conf 文件是否存在..."
-    if curl -u "${FTP_USER}:${FTP_PASSWORD}" -sfI "ftp://${FTP_SERVER}/dns.conf" >/dev/null 2>&1; then
-        log_success "远程 dns.conf 文件存在"
-    else
-        log_error "远程 dns.conf 文件不存在或无法访问"
-        log_error "请检查 FTP 服务器根目录下是否有 dns.conf 文件"
-        return 1
-    fi
-    
-    # 尝试下载文件
-    log_info "开始下载 dns.conf 文件..."
-    if curl -u "${FTP_USER}:${FTP_PASSWORD}" -sf "ftp://${FTP_SERVER}/dns.conf" -o "$temp_file" 2>/dev/null; then
-        log_success "远程 DNS 配置文件下载成功"
-        echo "$temp_file"
-    else
-        log_error "下载 dns.conf 文件失败"
-        log_error "尝试手动测试命令:"
-        log_error "  curl -u ${FTP_USER}:${FTP_PASSWORD} ftp://${FTP_SERVER}/dns.conf"
-        rm -f "$temp_file"
-        return 1
     fi
+    echo "$tmp"
 }
 
-# 比较两个文件是否相同
+# 文件比较
-compare_files() {
+compare_files() { diff -q "$1" "$2" >/dev/null 2>&1; }
-    local file1="$1"
+
-    local file2="$2"
+# 从 dns.conf 提取有效 IP
-    
+get_dns_ips() {
-    if [[ ! -f "$file1" || ! -f "$file2" ]]; then
+    grep -Eo '^[0-9]{1,3}(\.[0-9]{1,3}){3}$' "$1" | sort -u
-        return 1
-    fi
-    
-    # 使用 diff 比较文件内容
-    if diff -q "$file1" "$file2" >/dev/null 2>&1; then
-        return 0  # 文件相同
-    else
-        return 1  # 文件不同
-    fi
 }
 
-# 将 DNS 配置添加到 /etc/resolv.conf 开头
+# 安全更新 resolv.conf（保留符号链接）
 update_resolv_conf() {
-    local dns_conf_file="$1"
+    local dns_conf="$1"
-    
+    local dns_ips
-    log_info "更新 /etc/resolv.conf 文件..."
+    mapfile -t dns_ips < <(get_dns_ips "$dns_conf")
+    [[ ${#dns_ips[@]} -eq 0 ]] && { log_warning "未检测到有效 DNS"; return; }
 
-    # 创建临时文件
+    local target_file="$RESOLV_CONF"
-    local temp_resolv="/tmp/resolv.conf.$$"
+    if [[ ! -w "$RESOLV_CONF" ]]; then
-
+        log_warning "/etc/resolv.conf 不可写，使用兜底路径 $ALT_RESOLV_CONF"
-    # 将 dns.conf 内容转换为 nameserver 添加到临时文件开头
+        target_file="$ALT_RESOLV_CONF"
-    while IFS= read -r line; do
-        # 跳过空行和注释
-        [[ -z "$line" || "$line" =~ ^[[:space:]]*# ]] && continue
-        # 验证 IP 格式
-        if [[ "$line" =~ ^[0-9]{1,3}(\.[0-9]{1,3}){3}$ ]]; then
-            echo "nameserver $line" >> "$temp_resolv"
-            log_info "添加 DNS 到临时文件: $line"
-        else
-            log_warning "跳过无效 DNS: $line"
-        fi
-    done < "$dns_conf_file"
-
-    # 将原 resolv.conf 内容追加到临时文件后面
-    if [[ -f "$RESOLV_CONF" ]]; then
-        cat "$RESOLV_CONF" >> "$temp_resolv"
     fi
 
-    # 判断是否是 root
+    local temp="/tmp/resolv.new.$$"
-    if [[ $(id -u) -eq 0 ]]; then
+    cp "$target_file" "${target_file}.backup.$(date +%Y%m%d_%H%M%S)" 2>/dev/null || true
-        # root 直接写入
+    log_info "更新 DNS 配置文件: $target_file"
-        tee "$RESOLV_CONF" < "$temp_resolv" >/dev/null
+
-        chmod 644 "$RESOLV_CONF"
+    # 写入新的 nameserver 行
+    for ip in "${dns_ips[@]}"; do
+        echo "nameserver $ip"
+    done >"$temp"
+
+    # 追加原内容（去掉重复 nameserver）
+    grep -v '^nameserver' "$target_file" >>"$temp" 2>/dev/null || true
+    awk '!a[$0]++' "$temp" >"${temp}.uniq"
+
+    # ⚙️ 使用 cat 原地覆盖，避免 mv 引发 “设备忙”
+    if cat "${temp}.uniq" >"$target_file" 2>/dev/null; then
+        chmod 644 "$target_file"
+        log_success "DNS 更新完成: ${dns_ips[*]}"
     else
-        # 非 root 尝试使用 sudo
+        log_error "无法写入 $target_file，可能被系统锁定"
-        if command -v sudo >/dev/null 2>&1; then
-            sudo tee "$RESOLV_CONF" < "$temp_resolv" >/dev/null
-            sudo chmod 644 "$RESOLV_CONF"
-        else
-            log_error "非 root 用户且系统未安装 sudo，无法更新 /etc/resolv.conf"
-            rm -f "$temp_resolv"
-            exit 1
-        fi
     fi
 
-    # 清理临时文件
+    rm -f "$temp" "${temp}.uniq"
-    rm -f "$temp_resolv"
-
-    log_success "/etc/resolv.conf 已更新"
 }
 
-# 记录同步日志
+# 检查 resolv.conf 是否包含 dns.conf 内容
-log_sync() {
+ensure_dns_in_resolv() {
-    local message="$1"
+    local dns_conf="$1"
-    local timestamp=$(date '+%Y-%m-%d %H:%M:%S')
+    local dns_ips
-    echo "[$timestamp] $message" >> "$LOG_FILE"
+    mapfile -t dns_ips < <(get_dns_ips "$dns_conf")
+    [[ ${#dns_ips[@]} -eq 0 ]] && return
+
+    for ip in "${dns_ips[@]}"; do
+        if ! grep -q "nameserver $ip" "$RESOLV_CONF" 2>/dev/null; then
+            log_warning "检测到 /etc/resolv.conf 缺少 $ip，执行兜底修复"
+            update_resolv_conf "$dns_conf"
+            return
+        fi
+    done
+    log_info "/etc/resolv.conf 已包含所有 DNS"
 }
 
-# 主函数
+log_sync() { echo "[$(date '+%F %T')] $1" >>"$LOG_FILE"; }
+
 main() {
     log_info "开始 DNS 同步检查..."
-    log_sync "DNS 同步检查开始"
+    mkdir -p /opt/argus-metric
-    
+
-    # 确保系统目录存在
-    mkdir -p "/opt/argus-metric"
-    
-    # 获取 FTP 配置
     get_ftp_config
-    
+    local remote_file
-    # 检查本地 DNS 配置文件是否存在
+    if ! remote_file=$(download_remote_dns_conf); then
+        log_error "下载失败"; log_sync "同步失败"; exit 1
+    fi
+
     if [[ ! -f "$LOCAL_DNS_CONF" ]]; then
-        log_warning "本地 DNS 配置文件不存在: $LOCAL_DNS_CONF"
+        log_info "本地 dns.conf 不存在，初始化..."
-        log_warning "将下载远程配置文件并更新系统 DNS 设置"
+        cp "$remote_file" "$LOCAL_DNS_CONF"
-        
+        update_resolv_conf "$LOCAL_DNS_CONF"
-        # 下载远程配置文件
+        log_sync "首次同步完成"
-        if remote_file=$(download_remote_dns_conf); then
-            # 复制到本地
-            cp "$remote_file" "$LOCAL_DNS_CONF"
-            log_success "远程 DNS 配置文件已保存到本地"
-            
-            # 更新 resolv.conf
-            update_resolv_conf "$LOCAL_DNS_CONF"
-            log_sync "首次同步完成，DNS 配置已更新"
-            
-            # 清理临时文件
-            rm -f "$remote_file"
-        else
-            log_error "无法下载远程 DNS 配置文件，同步失败"
-            log_sync "同步失败：无法下载远程配置文件"
-            exit 1
-        fi
     else
-        log_info "本地 DNS 配置文件存在: $LOCAL_DNS_CONF"
+        if compare_files "$LOCAL_DNS_CONF" "$remote_file"; then
-        
+            log_info "dns.conf 无变化"
-        # 下载远程配置文件进行比较
+            ensure_dns_in_resolv "$LOCAL_DNS_CONF"
-        if remote_file=$(download_remote_dns_conf); then
+            log_sync "dns.conf 无变化，执行兜底检查"
-            # 比较文件
-            if compare_files "$LOCAL_DNS_CONF" "$remote_file"; then
-                log_info "DNS 配置文件无变化，无需更新"
-                log_sync "DNS 配置文件无变化"
-            else
-                log_info "检测到 DNS 配置文件有变化，开始同步..."
-                log_sync "检测到 DNS 配置文件变化，开始同步"
-                
-                # 更新本地配置文件
-                cp "$remote_file" "$LOCAL_DNS_CONF"
-                log_success "本地 DNS 配置文件已更新"
-                
-                # 更新 resolv.conf
-                update_resolv_conf "$LOCAL_DNS_CONF"
-                log_sync "DNS 配置同步完成"
-            fi
-            
-            # 清理临时文件
-            rm -f "$remote_file"
         else
-            log_error "无法下载远程 DNS 配置文件，跳过本次同步"
+            log_info "检测到 DNS 配置更新"
-            log_sync "同步失败：无法下载远程配置文件"
+            cp "$remote_file" "$LOCAL_DNS_CONF"
-            exit 1
+            update_resolv_conf "$LOCAL_DNS_CONF"
+            log_sync "DNS 配置同步完成"
         fi
     fi
-    
+
-    log_success "DNS 同步检查完成"
+    rm -f "$remote_file"
-    log_sync "DNS 同步检查完成"
+    log_success "DNS 同步流程完成"
 }
 
-# 脚本入口
 if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
     main "$@"
 fi

src/metric/tests/client-test-gpu-node/build/Dockerfile

@@ -0,0 +1,39 @@
+# 使用NVIDIA官方CUDA基础镜像 
+FROM nvidia/cuda:12.2.2-runtime-ubuntu22.04
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+# 设置时区
+ENV TZ=Asia/Shanghai
+
+RUN apt-get update -qq && \
+    apt-get install -y -qq \
+    tzdata \
+    curl \
+    wget \
+    gnupg2 \
+    software-properties-common \
+    ca-certificates \
+    && rm -rf /var/lib/apt/lists/*
+
+# 配置时区
+RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone
+
+WORKDIR /app
+
+# 创建启动脚本，在运行时验证GPU
+COPY <<EOF /app/start.sh
+#!/bin/bash
+echo "检查GPU环境..."
+if command -v nvidia-smi &> /dev/null; then
+    nvidia-smi
+    echo "GPU环境正常"
+else
+    echo "警告: nvidia-smi 命令不可用，请确保容器运行时启用了GPU支持"
+fi
+exec "\$@"
+EOF
+
+RUN chmod +x /app/start.sh
+
+CMD ["/app/start.sh", "/bin/bash"]

src/metric/tests/docker-compose.yml

@@ -142,3 +142,44 @@ services:
         max-size: "10m"
         max-file: "3"
 
+  test-gpu-node:
+    build:
+      context: ./client-test-gpu-node/build
+      dockerfile: Dockerfile
+    image: argus-metric-test-gpu-node:latest
+    container_name: argus-metric-test-gpu-node
+    hostname: test-metric-gpu-node-001
+    restart: unless-stopped
+    privileged: true
+    runtime: nvidia
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: all
+              capabilities: 
+                - gpu
+    depends_on:
+      - ftp
+      - prometheus
+    environment:
+      - TZ=Asia/Shanghai
+      - DEBIAN_FRONTEND=noninteractive
+      - NVIDIA_VISIBLE_DEVICES=all
+      - NVIDIA_DRIVER_CAPABILITIES=compute,utility
+      - GPU_MODE=gpu
+    volumes:
+      - ${DATA_ROOT:-/private}/argus/agent:/private/argus/agent
+      - /etc/localtime:/etc/localtime:ro
+      - /etc/timezone:/etc/timezone:ro
+    command: sleep infinity
+    networks:
+      default:
+        ipv4_address: 172.30.0.51
+    logging:
+      driver: "json-file"
+      options:
+        max-size: "10m"
+        max-file: "3"
+

src/metric/tests/scripts/01_start_services.sh

@@ -21,6 +21,7 @@ docker ps | grep argus-ftp
 docker ps | grep argus-prometheus
 docker ps | grep argus-grafana
 docker ps | grep argus-metric-test-node
+docker ps | grep argus-metric-test-gpu-node
 
-echo "[01] 所有服务已启动"
+echo "[01] 基础服务已启动"
 

src/metric/tests/scripts/02_publish_artifact.sh

@@ -20,7 +20,7 @@ else
     echo "[02] 默认路径: $OUTPUT_DIR"
 fi
 
-OWNER="${FTP_UID:-2133}:${FTP_GID:-2015}"
+OWNER="${ARGUS_BUILD_UID:-2133}:${ARGUS_BUILD_GID:-2015}"
 
 cd "$PLUGIN_DIR"
 

src/metric/tests/scripts/04_test_gpu_node_install.sh

@@ -0,0 +1,33 @@
+#!/bin/bash
+set -e
+
+FTP_SERVER="${FTP_SERVER:-172.30.0.40}"
+FTP_USER="${FTP_USER:-ftpuser}"
+FTP_PASSWORD="${FTP_PASSWORD:-ZGClab1234!}"
+FTP_PORT="${FTP_PORT:-21}"
+
+FTP_HOST="${FTP_SERVER}"
+
+echo "[03] 进入测试节点执行安装..."
+echo "[03] 使用 FTP 地址: ${FTP_HOST}:${FTP_PORT}"
+
+docker exec argus-metric-test-gpu-node bash -c "
+set -e
+
+if ! command -v curl &>/dev/null; then
+    echo '[03] curl 未安装，正在安装...'
+    apt-get update && apt-get install -y curl
+fi
+
+cd /tmp
+echo '[03] 下载 setup.sh...'
+curl -u ${FTP_USER}:${FTP_PASSWORD} ftp://${FTP_HOST}:${FTP_PORT}/setup.sh -o setup.sh
+
+echo '[03] 执行安装...'
+chmod +x setup.sh
+bash setup.sh --server ${FTP_HOST} --user ${FTP_USER} --password '${FTP_PASSWORD}' --port ${FTP_PORT}
+
+echo '[03] 安装完成'
+"
+
+echo "[03] 完成"

src/metric/tests/scripts/common/start-all.sh

@@ -44,6 +44,7 @@ BUILD_DIRS=(
     "../prometheus/build"
     "../grafana/build"
     "client-test-node/build"
+    "client-test-gpu-node/build"
 )
 
 for dir in "${BUILD_DIRS[@]}"; do
@@ -87,7 +88,7 @@ echo "2. 准备 Docker 镜像..."
 
 # 检查镜像是否存在
 IMAGE_CACHE_DIR="$TEST_DIR/images-cache"
-IMAGES=("argus-metric-ftp:latest" "argus-metric-prometheus:latest" "argus-metric-grafana:latest" "argus-metric-test-node:latest")
+IMAGES=("argus-metric-ftp:latest" "argus-metric-prometheus:latest" "argus-metric-grafana:latest" "argus-metric-test-node:latest" "argus-metric-test-gpu-node:latest")
 all_images_exist=true
 
 for image in "${IMAGES[@]}"; do
@@ -126,6 +127,9 @@ else
                     "argus-metric-test-node:latest")
                         cache_file="${IMAGE_CACHE_DIR}/argus-test-node.tar"
                         ;;
+                    "argus-metric-test-gpu-node:latest")
+                        cache_file="${IMAGE_CACHE_DIR}/argus-test-gpu-node.tar"
+                        ;;
                 esac
                 
                 if [ -f "$cache_file" ]; then
@@ -174,6 +178,9 @@ else
                     "argus-metric-test-node:latest")
                         docker save -o "${IMAGE_CACHE_DIR}/argus-test-node.tar" "$image" && echo "  已保存: argus-test-node.tar"
                         ;;
+                    "argus-metric-test-gpu-node:latest")
+                        docker save -o "${IMAGE_CACHE_DIR}/argus-test-gpu-node.tar" "$image" && echo "  已保存: argus-test-gpu-node.tar"
+                        ;;
                 esac
             done
             echo "镜像已保存到: $IMAGE_CACHE_DIR/"
@@ -185,40 +192,12 @@ else
 fi
 
 echo ""
-echo "3. 启动服务..."
+echo "3. 启动基础服务..."
 cd "$TEST_DIR"
-docker compose up -d
+# 启动除GPU节点外的所有服务
+docker compose up -d ftp prometheus grafana test-node test-gpu-node
 
 echo ""
 echo "4. 等待服务启动..."
 sleep 5
 
-echo ""
-echo "5. 检查服务状态..."
-cd "$TEST_DIR"
-docker compose ps
-
-echo ""
-echo "=========================================="
-echo "  服务启动完成！"
-echo "=========================================="
-echo ""
-echo "服务访问地址："
-echo "  - FTP:        ftp://localhost:${FTP_PORT:-21}"
-echo "    用户名: ftpuser"
-echo "    密码:   ${FTP_PASSWORD:-ZGClab1234!}"
-echo ""
-echo "  - Prometheus: http://localhost:${PROMETHEUS_PORT:-9090}"
-echo ""
-echo "  - Grafana:    http://localhost:${GRAFANA_PORT:-3000}"
-echo "    用户名: admin"
-echo "    密码:   admin"
-echo ""
-echo "常用命令："
-echo "  查看日志:     docker compose logs -f [service]"
-echo "  停止服务:     docker compose stop"
-echo "  重启服务:     docker compose restart"
-echo "  停止并删除:   docker compose down"
-echo "  停止并删除卷: docker compose down -v"
-echo ""
-

src/metric/tests/scripts/save-images.sh

@@ -24,7 +24,8 @@ declare -A IMAGES=(
     ["argus-metric-ftp:latest"]="argus-ftp.tar"
     ["argus-metric-prometheus:latest"]="argus-prometheus.tar"
     ["argus-metric-grafana:latest"]="argus-grafana.tar"
-    ["ubuntu:22.04"]="test-node.tar"
+    ["argus-metric-test-node:latest"]="argus-test-node.tar"
+    ["argus-metric-test-gpu-node:latest"]="argus-test-gpu-node.tar"
 )
 
 # 检查镜像是否存在并保存

src/sys/debug/.env.example

@@ -0,0 +1,12 @@
+# Generated by 01_bootstrap.sh
+SYS_DEBUG_PRIVATE_CORE=/absolute/path/to/private
+SYS_DEBUG_PRIVATE_NODEA=/absolute/path/to/private-nodea
+SYS_DEBUG_PRIVATE_NODEB=/absolute/path/to/private-nodeb
+SYS_DEBUG_TMP_DIR=/absolute/path/to/tmp
+SYS_DEBUG_NETWORK_NAME=argus-debug-net
+SYS_DEBUG_NETWORK_SUBNET=172.30.0.0/16
+SYS_DEBUG_NETWORK_GATEWAY=172.30.0.1
+SYS_DEBUG_PROJECT_NAME=argus-debug
+SYS_DEBUG_CONTAINER_PREFIX=argus-debug
+ARGUS_BUILD_UID=2133
+ARGUS_BUILD_GID=2015

src/sys/debug/README.md

@@ -0,0 +1,68 @@
+# ARGUS 系统调试部署模式
+
+该目录提供基于系统级 E2E 测试构建的调试部署流程，便于本地快速复现与排查问题。核心特性：
+
+- 独立 docker 网络 `argus-debug-net`（默认子网 `172.30.0.0/16`），避免与 `src/sys/tests` 冲突。
+- 私有数据目录可通过参数自定义，例如 `--private-root /tmp/argus-debug`。
+- 默认保留调试过程生成的文件，避免 `down`/`bootstrap` 自动删除。
+
+## 快速开始
+
+```bash
+cd src/sys/debug
+
+# 仅首次需要，创建 external 网络
+./scripts/network-create.sh
+
+# 初始化目录/构建 agent/写入 .env
+./scripts/01_bootstrap.sh --private-root /tmp/argus-debug
+
+# 启动调试栈
+./scripts/02_up.sh
+
+# 根据需要执行验证脚本（03～08）
+./scripts/03_wait_ready.sh
+...
+
+# 调试结束停止服务
+./scripts/09_down.sh
+
+# 若需移除网络或数据
+./scripts/network-destroy.sh
+./scripts/clean-data.sh
+```
+
+> **提示**：调试与测试栈不能同时运行，应保持 `src/sys/tests` 中的 `argus-sys` 栈已停止。
+
+## 参数与环境变量
+
+- `--private-root <path>`：同时指定核心服务与两个节点的私有目录根，脚本自动派生 `private`、`private-nodea`、`private-nodeb`。
+- `--private-core <path>`、`--private-nodea <path>`、`--private-nodeb <path>`：分别覆盖单独目录。
+- 环境变量可覆盖 `.env` 中写入的值，例如 `export SYS_DEBUG_NETWORK_NAME=my-debug-net`。
+- `.env` 文件字段：
+  - `SYS_DEBUG_PRIVATE_CORE`
+  - `SYS_DEBUG_PRIVATE_NODEA`
+  - `SYS_DEBUG_PRIVATE_NODEB`
+  - `SYS_DEBUG_TMP_DIR`
+  - `SYS_DEBUG_NETWORK_NAME`
+  - `SYS_DEBUG_NETWORK_SUBNET`
+  - `SYS_DEBUG_NETWORK_GATEWAY`
+  - `SYS_DEBUG_PROJECT_NAME`
+  - `SYS_DEBUG_CONTAINER_PREFIX`
+  - `ARGUS_BUILD_UID` / `ARGUS_BUILD_GID`
+
+## 脚本说明
+
+- `scripts/common.sh`：通用函数与环境加载。
+- `scripts/network-create.sh` / `network-destroy.sh`：管理 external 网络。
+- `scripts/00_debug_all.sh`：顺序执行 01～08（默认不执行 09）。
+- `scripts/clean-data.sh`：选择性清理宿主机私有数据。
+- `scripts/03_wait_ready.sh`：除了等待各服务就绪，还会在 Elasticsearch 就绪后自动将磁盘水位阈值放宽（97%/98%/99%），避免在磁盘紧张的调试环境中分片分配失败。
+- `scripts/08_restart_agent_reregister.sh`：将 node-b 切换到 `SYS_DEBUG_NODEB_FIXED_IP`（默认 `172.30.0.200`），如果目标地址与当前 IP 相同脚本会报错提醒重新选择地址。
+- 其它 `01～09` 与测试目录对应，但针对参数化路径及网络做了调整。
+
+## 注意事项
+
+- 若宿主机未安装 Docker，脚本将提示错误并退出。
+- 当指定的私有目录已存在数据时，脚本不会清理，请确认内容安全后再复用。
+- 与测试环境共用镜像：请提前执行仓库根目录的 `./build/build_images.sh`。

src/sys/debug/docker-compose.yml

@@ -0,0 +1,147 @@
+version: "3.8"
+
+networks:
+  argus-debug-net:
+    external: true
+    name: ${SYS_DEBUG_NETWORK_NAME:-argus-debug-net}
+
+services:
+  bind:
+    image: ${BIND_IMAGE_TAG:-argus-bind9:latest}
+    container_name: ${SYS_DEBUG_CONTAINER_PREFIX:-argus-debug}-bind
+    networks:
+      argus-debug-net:
+        ipv4_address: ${SYS_DEBUG_BIND_IP:-172.30.0.2}
+    volumes:
+      - ${SYS_DEBUG_PRIVATE_CORE}:/private
+    restart: unless-stopped
+
+  master:
+    image: ${MASTER_IMAGE_TAG:-argus-master:latest}
+    container_name: ${SYS_DEBUG_CONTAINER_PREFIX:-argus-debug}-master
+    depends_on:
+      - bind
+    environment:
+      - OFFLINE_THRESHOLD_SECONDS=6
+      - ONLINE_THRESHOLD_SECONDS=2
+      - SCHEDULER_INTERVAL_SECONDS=1
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
+    ports:
+      - "32300:3000"
+    volumes:
+      - ${SYS_DEBUG_PRIVATE_CORE}/argus/master:/private/argus/master
+      - ${SYS_DEBUG_PRIVATE_CORE}/argus/metric/prometheus:/private/argus/metric/prometheus
+      - ${SYS_DEBUG_PRIVATE_CORE}/argus/etc:/private/argus/etc
+    networks:
+      argus-debug-net:
+        ipv4_address: ${SYS_DEBUG_MASTER_IP:-172.30.0.10}
+    restart: unless-stopped
+
+  es:
+    image: ${ES_IMAGE_TAG:-argus-elasticsearch:latest}
+    container_name: ${SYS_DEBUG_CONTAINER_PREFIX:-argus-debug}-es
+    environment:
+      - discovery.type=single-node
+      - xpack.security.enabled=false
+      - ES_JAVA_OPTS=-Xms512m -Xmx512m
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
+    volumes:
+      - ${SYS_DEBUG_PRIVATE_CORE}/argus/log/elasticsearch:/private/argus/log/elasticsearch
+      - ${SYS_DEBUG_PRIVATE_CORE}/argus/etc:/private/argus/etc
+    ports:
+      - "9200:9200"
+    networks:
+      argus-debug-net:
+        ipv4_address: ${SYS_DEBUG_ES_IP:-172.30.0.20}
+    restart: unless-stopped
+
+  kibana:
+    image: ${KIBANA_IMAGE_TAG:-argus-kibana:latest}
+    container_name: ${SYS_DEBUG_CONTAINER_PREFIX:-argus-debug}-kibana
+    environment:
+      - ELASTICSEARCH_HOSTS=http://es.log.argus.com:9200
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
+    volumes:
+      - ${SYS_DEBUG_PRIVATE_CORE}/argus/log/kibana:/private/argus/log/kibana
+      - ${SYS_DEBUG_PRIVATE_CORE}/argus/etc:/private/argus/etc
+    depends_on:
+      - es
+    ports:
+      - "5601:5601"
+    networks:
+      argus-debug-net:
+        ipv4_address: ${SYS_DEBUG_KIBANA_IP:-172.30.0.30}
+    restart: unless-stopped
+
+  node-a:
+    image: ubuntu:22.04
+    container_name: ${SYS_DEBUG_CONTAINER_PREFIX:-argus-debug}-node-a
+    hostname: ${SYS_DEBUG_NODEA_HOST:-dev-yyrshare-nbnyx10-cp2f-pod-0}
+    depends_on:
+      - master
+      - bind
+      - es
+    environment:
+      - MASTER_ENDPOINT=http://master.argus.com:3000
+      - REPORT_INTERVAL_SECONDS=2
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
+      - ES_HOST=es
+      - ES_PORT=9200
+      - CLUSTER=local
+      - RACK=dev
+    volumes:
+      - ${SYS_DEBUG_PRIVATE_NODEA}/argus/agent/${SYS_DEBUG_NODEA_HOST:-dev-yyrshare-nbnyx10-cp2f-pod-0}:/private/argus/agent/${SYS_DEBUG_NODEA_HOST:-dev-yyrshare-nbnyx10-cp2f-pod-0}
+      - ../../agent/dist/argus-agent:/usr/local/bin/argus-agent:ro
+      - ../tests/scripts/node_entrypoint.sh:/usr/local/bin/node-entrypoint.sh:ro
+      - ../../log/fluent-bit/build/start-fluent-bit.sh:/assets/start-fluent-bit.sh:ro
+      - ../../log/fluent-bit/build/etc:/assets/fluent-bit/etc:ro
+      - ../../log/fluent-bit/build/packages:/assets/fluent-bit/packages:ro
+    entrypoint:
+      - /usr/local/bin/node-entrypoint.sh
+    dns:
+      - ${SYS_DEBUG_BIND_IP:-172.30.0.2}
+    ports:
+      - "2020:2020"
+    networks:
+      argus-debug-net:
+        ipv4_address: ${SYS_DEBUG_NODEA_IP:-172.30.0.101}
+    restart: unless-stopped
+
+  node-b:
+    image: ubuntu:22.04
+    container_name: ${SYS_DEBUG_CONTAINER_PREFIX:-argus-debug}-node-b
+    hostname: ${SYS_DEBUG_NODEB_HOST:-dev-yyrshare-uuuu10-ep2f-pod-0}
+    depends_on:
+      - master
+      - bind
+      - es
+    environment:
+      - MASTER_ENDPOINT=http://master.argus.com:3000
+      - REPORT_INTERVAL_SECONDS=2
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
+      - ES_HOST=es
+      - ES_PORT=9200
+      - CLUSTER=local
+      - RACK=dev
+    volumes:
+      - ${SYS_DEBUG_PRIVATE_NODEB}/argus/agent/${SYS_DEBUG_NODEB_HOST:-dev-yyrshare-uuuu10-ep2f-pod-0}:/private/argus/agent/${SYS_DEBUG_NODEB_HOST:-dev-yyrshare-uuuu10-ep2f-pod-0}
+      - ../../agent/dist/argus-agent:/usr/local/bin/argus-agent:ro
+      - ../tests/scripts/node_entrypoint.sh:/usr/local/bin/node-entrypoint.sh:ro
+      - ../../log/fluent-bit/build/start-fluent-bit.sh:/assets/start-fluent-bit.sh:ro
+      - ../../log/fluent-bit/build/etc:/assets/fluent-bit/etc:ro
+      - ../../log/fluent-bit/build/packages:/assets/fluent-bit/packages:ro
+    entrypoint:
+      - /usr/local/bin/node-entrypoint.sh
+    dns:
+      - ${SYS_DEBUG_BIND_IP:-172.30.0.2}
+    ports:
+      - "2021:2020"
+    networks:
+      argus-debug-net:
+        ipv4_address: ${SYS_DEBUG_NODEB_IP:-172.30.0.102}
+    restart: unless-stopped

src/sys/debug/scripts/00_debug_all.sh

@@ -0,0 +1,24 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+SCRIPTS=(
+  "01_bootstrap.sh"
+  "02_up.sh"
+  "03_wait_ready.sh"
+  "04_verify_dns_routing.sh"
+  "05_agent_register.sh"
+  "06_write_health_and_assert.sh"
+  "07_logs_send_and_assert.sh"
+  "08_restart_agent_reregister.sh"
+)
+
+for script in "${SCRIPTS[@]}"; do
+  echo "[SYS-DEBUG] Running $script"
+  "$SCRIPT_DIR/$script"
+  echo "[SYS-DEBUG] $script completed"
+  echo
+done
+
+echo "[SYS-DEBUG] Complete. Run scripts/09_down.sh when finished (data retained)."

src/sys/debug/scripts/01_bootstrap.sh

@@ -0,0 +1,210 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+PRIVATE_ROOT=""
+PRIVATE_CORE="$SYS_DEBUG_PRIVATE_CORE"
+PRIVATE_NODEA="$SYS_DEBUG_PRIVATE_NODEA"
+PRIVATE_NODEB="$SYS_DEBUG_PRIVATE_NODEB"
+TMP_DIR_VAL="$SYS_DEBUG_TMP_DIR"
+NETWORK_NAME="$SYS_DEBUG_NETWORK_NAME"
+NETWORK_SUBNET="$SYS_DEBUG_NETWORK_SUBNET"
+NETWORK_GATEWAY="$SYS_DEBUG_NETWORK_GATEWAY"
+PROJECT_NAME="$SYS_DEBUG_PROJECT_NAME"
+CONTAINER_PREFIX="$SYS_DEBUG_CONTAINER_PREFIX"
+NODEB_FIXED_IP=${SYS_DEBUG_NODEB_FIXED_IP:-172.30.0.200}
+
+usage() {
+  cat <<EOF
+Usage: ${0##*/} [--private-root PATH] [--private-core PATH] \
+                 [--private-nodea PATH] [--private-nodeb PATH] \
+                 [--tmp-dir PATH] [--network-name NAME] \
+                 [--network-subnet CIDR] [--network-gateway IP]
+
+Prepare directories, build agent binary, and write .env for debug stack.
+EOF
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --private-root)
+      shift; [[ $# -gt 0 ]] || { echo "--private-root requires value" >&2; exit 1; }
+      PRIVATE_ROOT="$1"
+      ;;
+    --private-root=*)
+      PRIVATE_ROOT="${1#*=}"
+      ;;
+    --private-core)
+      shift; [[ $# -gt 0 ]] || { echo "--private-core requires value" >&2; exit 1; }
+      PRIVATE_CORE="$1"
+      ;;
+    --private-core=*)
+      PRIVATE_CORE="${1#*=}"
+      ;;
+    --private-nodea)
+      shift; [[ $# -gt 0 ]] || { echo "--private-nodea requires value" >&2; exit 1; }
+      PRIVATE_NODEA="$1"
+      ;;
+    --private-nodea=*)
+      PRIVATE_NODEA="${1#*=}"
+      ;;
+    --private-nodeb)
+      shift; [[ $# -gt 0 ]] || { echo "--private-nodeb requires value" >&2; exit 1; }
+      PRIVATE_NODEB="$1"
+      ;;
+    --private-nodeb=*)
+      PRIVATE_NODEB="${1#*=}"
+      ;;
+    --tmp-dir)
+      shift; [[ $# -gt 0 ]] || { echo "--tmp-dir requires value" >&2; exit 1; }
+      TMP_DIR_VAL="$1"
+      ;;
+    --tmp-dir=*)
+      TMP_DIR_VAL="${1#*=}"
+      ;;
+    --network-name)
+      shift; [[ $# -gt 0 ]] || { echo "--network-name requires value" >&2; exit 1; }
+      NETWORK_NAME="$1"
+      ;;
+    --network-name=*)
+      NETWORK_NAME="${1#*=}"
+      ;;
+    --network-subnet)
+      shift; [[ $# -gt 0 ]] || { echo "--network-subnet requires value" >&2; exit 1; }
+      NETWORK_SUBNET="$1"
+      ;;
+    --network-subnet=*)
+      NETWORK_SUBNET="${1#*=}"
+      ;;
+    --network-gateway)
+      shift; [[ $# -gt 0 ]] || { echo "--network-gateway requires value" >&2; exit 1; }
+      NETWORK_GATEWAY="$1"
+      ;;
+    --network-gateway=*)
+      NETWORK_GATEWAY="${1#*=}"
+      ;;
+    -h|--help)
+      usage
+      exit 0
+      ;;
+    *)
+      echo "Unknown argument: $1" >&2
+      usage >&2
+      exit 1
+      ;;
+  esac
+  shift
+done
+
+if [[ -n "$PRIVATE_ROOT" ]]; then
+  PRIVATE_CORE="$PRIVATE_ROOT/private"
+  PRIVATE_NODEA="$PRIVATE_ROOT/private-nodea"
+  PRIVATE_NODEB="$PRIVATE_ROOT/private-nodeb"
+fi
+
+PRIVATE_CORE=$(abs_path "$PRIVATE_CORE")
+PRIVATE_NODEA=$(abs_path "$PRIVATE_NODEA")
+PRIVATE_NODEB=$(abs_path "$PRIVATE_NODEB")
+TMP_DIR_VAL=$(abs_path "$TMP_DIR_VAL")
+
+log "Preparing directories under $PRIVATE_CORE"
+mkdir -p \
+  "$PRIVATE_CORE/argus/etc" \
+  "$PRIVATE_CORE/argus/bind" \
+  "$PRIVATE_CORE/argus/master" \
+  "$PRIVATE_CORE/argus/metric/prometheus" \
+  "$PRIVATE_CORE/argus/log/elasticsearch" \
+  "$PRIVATE_CORE/argus/log/kibana" \
+  "$PRIVATE_NODEA/argus/agent/$HOST_A/health" \
+  "$PRIVATE_NODEB/argus/agent/$HOST_B/health" \
+  "$TMP_DIR_VAL"
+
+log "Aligning ownership for core directories"
+chown -R "${ARGUS_BUILD_UID}:${ARGUS_BUILD_GID}" \
+  "$PRIVATE_CORE/argus/log/elasticsearch" \
+  "$PRIVATE_CORE/argus/log/kibana" \
+  "$PRIVATE_CORE/argus/etc" 2>/dev/null || true
+
+log "Distributing update-dns.sh"
+BIND_UPDATE_SRC="$REPO_ROOT/src/bind/build/update-dns.sh"
+BIND_UPDATE_DEST="$PRIVATE_CORE/argus/etc/update-dns.sh"
+if [[ -f "$BIND_UPDATE_SRC" ]]; then
+  cp "$BIND_UPDATE_SRC" "$BIND_UPDATE_DEST"
+  chmod +x "$BIND_UPDATE_DEST"
+else
+  echo "[WARN] Missing $BIND_UPDATE_SRC" >&2
+fi
+
+require_docker
+
+ensure_image() {
+  local image="$1"
+  if ! docker image inspect "$image" >/dev/null 2>&1; then
+    echo "[ERR] Missing image: $image. Run ./build/build_images.sh" >&2
+    exit 1
+  fi
+}
+
+log "Ensuring required images exist"
+ensure_image "${ES_IMAGE_TAG:-argus-elasticsearch:latest}"
+ensure_image "${KIBANA_IMAGE_TAG:-argus-kibana:latest}"
+ensure_image "${BIND_IMAGE_TAG:-argus-bind9:latest}"
+ensure_image "${MASTER_IMAGE_TAG:-argus-master:latest}"
+
+log "Building agent binary"
+pushd "$REPO_ROOT/src/agent" >/dev/null
+./scripts/build_binary.sh
+popd >/dev/null
+
+AGENT_BIN="$REPO_ROOT/src/agent/dist/argus-agent"
+if [[ ! -x "$AGENT_BIN" ]]; then
+  echo "[ERR] Agent binary not found at $AGENT_BIN" >&2
+  exit 1
+fi
+echo "$AGENT_BIN" > "$TMP_DIR_VAL/agent_binary_path"
+
+log "Preparing environment file contents"
+tmp_env="$(mktemp)"
+cat > "$tmp_env" <<EOF
+SYS_DEBUG_PRIVATE_CORE=$PRIVATE_CORE
+SYS_DEBUG_PRIVATE_NODEA=$PRIVATE_NODEA
+SYS_DEBUG_PRIVATE_NODEB=$PRIVATE_NODEB
+SYS_DEBUG_TMP_DIR=$TMP_DIR_VAL
+SYS_DEBUG_NETWORK_NAME=$NETWORK_NAME
+SYS_DEBUG_NETWORK_SUBNET=$NETWORK_SUBNET
+SYS_DEBUG_NETWORK_GATEWAY=$NETWORK_GATEWAY
+SYS_DEBUG_PROJECT_NAME=$PROJECT_NAME
+SYS_DEBUG_CONTAINER_PREFIX=$CONTAINER_PREFIX
+SYS_DEBUG_NODEA_HOST=$HOST_A
+SYS_DEBUG_NODEB_HOST=$HOST_B
+SYS_DEBUG_BIND_IP=${SYS_DEBUG_BIND_IP:-172.30.0.2}
+SYS_DEBUG_MASTER_IP=${SYS_DEBUG_MASTER_IP:-172.30.0.10}
+SYS_DEBUG_ES_IP=${SYS_DEBUG_ES_IP:-172.30.0.20}
+SYS_DEBUG_KIBANA_IP=${SYS_DEBUG_KIBANA_IP:-172.30.0.30}
+SYS_DEBUG_NODEA_IP=${SYS_DEBUG_NODEA_IP:-172.30.0.101}
+SYS_DEBUG_NODEB_IP=${SYS_DEBUG_NODEB_IP:-172.30.0.102}
+SYS_DEBUG_NODEB_FIXED_IP=$NODEB_FIXED_IP
+ARGUS_BUILD_UID=$ARGUS_BUILD_UID
+ARGUS_BUILD_GID=$ARGUS_BUILD_GID
+EOF
+
+if [[ -f "$ENV_FILE" ]]; then
+  if cmp -s "$tmp_env" "$ENV_FILE"; then
+    log ".env already up-to-date"
+    rm -f "$tmp_env"
+    if [[ ! -f "$DEBUG_ROOT/.env.lock" ]]; then
+      cp "$ENV_FILE" "$DEBUG_ROOT/.env.lock"
+    fi
+  else
+    mv "$ENV_FILE" "$ENV_FILE.bak"
+    mv "$tmp_env" "$ENV_FILE"
+    cp "$ENV_FILE" "$DEBUG_ROOT/.env.lock"
+    log "Bootstrap updated .env (previous saved at ${ENV_FILE}.bak)"
+  fi
+else
+  mv "$tmp_env" "$ENV_FILE"
+  cp "$ENV_FILE" "$DEBUG_ROOT/.env.lock"
+  log "Bootstrap created .env at $ENV_FILE"
+fi

src/sys/debug/scripts/02_up.sh

@@ -0,0 +1,19 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+ensure_paths_defined
+require_docker
+
+if ! docker network inspect "$SYS_DEBUG_NETWORK_NAME" >/dev/null 2>&1; then
+  echo "[ERR] Network $SYS_DEBUG_NETWORK_NAME not found. Run scripts/network-create.sh first." >&2
+  exit 1
+fi
+
+log "Starting debug stack on project $SYS_DEBUG_PROJECT_NAME"
+compose up -d
+
+log "Services started: master:32300 es:9200 kibana:5601 node-a:2020 node-b:2021"

src/sys/debug/scripts/03_wait_ready.sh

@@ -0,0 +1,84 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+ensure_paths_defined
+
+service_id() {
+  compose ps -q "$1"
+}
+
+wait_http() {
+  local url="$1"; local attempts="${2:-120}"; local i=1
+  while (( i <= attempts )); do
+    if curl -fsS "$url" >/dev/null 2>&1; then
+      return 0
+    fi
+    echo "[..] waiting $url ($i/$attempts)"
+    sleep 5
+    ((i++))
+  done
+  echo "[ERR] Timeout waiting for $url" >&2
+  return 1
+}
+
+log "Waiting for ES/Kibana/Master/Fluent Bit/Bind"
+
+attempt=1; max=120
+while (( attempt <= max )); do
+  if curl -fsS "http://localhost:9200/_cluster/health?wait_for_status=yellow&timeout=1s" >/dev/null 2>&1; then
+    break
+  fi
+  echo "[..] waiting ES ($attempt/$max)"
+  sleep 5
+  ((attempt++))
+done
+if (( attempt > max )); then
+  echo "[ERR] ES not ready" >&2
+  exit 1
+fi
+
+log "Applying relaxed ES disk watermarks for debug"
+curl -fsS -XPUT "http://localhost:9200/_cluster/settings" \
+  -H 'Content-Type: application/json' \
+  -d '{
+    "transient": {
+      "cluster.routing.allocation.disk.watermark.low": "99%",
+      "cluster.routing.allocation.disk.watermark.high": "99%",
+      "cluster.routing.allocation.disk.watermark.flood_stage": "99%"
+    }
+  }' >/dev/null || echo "[WARN] Failed to adjust ES watermarks"
+
+log "Waiting for Kibana to be available (HTTP 200)"
+kb_attempt=1; kb_max=180
+while (( kb_attempt <= kb_max )); do
+  body=$(curl -sS "http://localhost:5601/api/status" 2>/dev/null || true)
+  code=$(curl -s -o /dev/null -w "%{http_code}" "http://localhost:5601/api/status" || echo 000)
+  if [[ "$code" == "200" ]] && echo "$body" | grep -q '"level":"available"'; then
+    log "Kibana available"
+    break
+  fi
+  echo "[..] waiting kibana 200 ($kb_attempt/$kb_max), last_code=$code"
+  sleep 5
+  ((kb_attempt++))
+done
+if (( kb_attempt > kb_max )); then
+  echo "[ERR] Kibana did not reach HTTP 200" >&2
+  exit 1
+fi
+
+wait_http "http://localhost:32300/readyz" 120
+wait_http "http://localhost:2020/api/v2/metrics" 120
+wait_http "http://localhost:2021/api/v2/metrics" 120
+
+BIND_ID="$(service_id bind)"
+if [[ -n "$BIND_ID" ]]; then
+  docker exec "$BIND_ID" named-checkconf >/dev/null
+else
+  echo "[WARN] bind container id not found" >&2
+fi
+
+log "All services are ready"

src/sys/debug/scripts/04_verify_dns_routing.sh

@@ -0,0 +1,51 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+ensure_paths_defined
+
+service_id() {
+  compose ps -q "$1"
+}
+
+log "Verifying DNS routing via bind"
+
+MASTER_FILE="$SYS_DEBUG_PRIVATE_CORE/argus/etc/master.argus.com"
+if [[ ! -f "$MASTER_FILE" ]]; then
+  echo "[ERR] master.argus.com file missing at $MASTER_FILE" >&2
+  exit 1
+fi
+MASTER_IP_HOST="$(tr -d '\r\n' < "$MASTER_FILE" || true)"
+log "master.argus.com file content: $MASTER_IP_HOST"
+
+BIN_ID="$(service_id bind)"
+if [[ -n "$BIN_ID" ]]; then
+  DIG_IP="$(docker exec "$BIN_ID" dig +short master.argus.com A | tail -n1 || true)"
+  log "dig(master.argus.com) from bind container -> $DIG_IP"
+  if [[ -z "$DIG_IP" ]]; then
+    echo "[ERR] bind did not resolve master.argus.com" >&2
+    exit 1
+  fi
+else
+  echo "[WARN] bind container not found; skip dig" >&2
+fi
+
+for node in node-a node-b; do
+  CID="$(service_id "$node")"
+  if [[ -z "$CID" ]]; then
+    echo "[ERR] Container for $node not found" >&2
+    exit 1
+  fi
+  log "Checking resolution inside $node"
+  if ! docker exec "$CID" getent hosts master.argus.com >/dev/null 2>&1; then
+    echo "[ERR] $node cannot resolve master.argus.com" >&2
+    exit 1
+  fi
+  RES="$(docker exec "$CID" getent hosts master.argus.com | awk '{print $1}' | head -n1)"
+  log "$node resolved master.argus.com -> $RES"
+done
+
+log "DNS routing verified"

src/sys/debug/scripts/05_agent_register.sh

@@ -0,0 +1,84 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+ensure_paths_defined
+
+TMP_DIR_LOCAL="$TMP_DIR"
+mkdir -p "$TMP_DIR_LOCAL"
+
+API_BASE="http://localhost:32300/api/v1/master"
+
+log "Waiting for agent nodes to register"
+
+extract_node() {
+  local name="$1"; local output="$2"; local json_file="$3"
+  python3 - "$name" "$output" "$json_file" <<'PY'
+import json, sys, pathlib
+name = sys.argv[1]
+out = pathlib.Path(sys.argv[2])
+json_file = sys.argv[3]
+with open(json_file, 'r') as fh:
+    data = json.load(fh)
+node = next((n for n in data if n.get("name") == name), None)
+if node:
+    out.write_text(node["id"])
+    print(node["id"])
+PY
+}
+
+ID_A=""; ID_B=""
+for _ in {1..60}; do
+  sleep 2
+  resp=$(curl -fsS "$API_BASE/nodes" 2>/dev/null || true)
+  [[ -z "$resp" ]] && continue
+  if ! echo "$resp" | head -c1 | grep -q '\['; then
+    continue
+  fi
+  echo "$resp" > "$TMP_DIR_LOCAL/nodes_list.json"
+  ID_A=$(extract_node "$HOST_A" "$TMP_DIR_LOCAL/node_id_a" "$TMP_DIR_LOCAL/nodes_list.json" 2>/dev/null || true)
+  ID_B=$(extract_node "$HOST_B" "$TMP_DIR_LOCAL/node_id_b" "$TMP_DIR_LOCAL/nodes_list.json" 2>/dev/null || true)
+  if [[ -s "$TMP_DIR_LOCAL/node_id_a" && -s "$TMP_DIR_LOCAL/node_id_b" ]]; then
+    break
+  fi
+done
+
+if [[ ! -s "$TMP_DIR_LOCAL/node_id_a" || ! -s "$TMP_DIR_LOCAL/node_id_b" ]]; then
+  echo "[ERR] Agents did not register in time" >&2
+  exit 1
+fi
+
+node_detail() {
+  local id="$1"; local out="$2"
+  curl -fsS "$API_BASE/nodes/$id" -o "$out"
+}
+
+node_detail "$(cat "$TMP_DIR_LOCAL/node_id_a")" "$TMP_DIR_LOCAL/detail_a.json"
+node_detail "$(cat "$TMP_DIR_LOCAL/node_id_b")" "$TMP_DIR_LOCAL/detail_b.json"
+
+python3 - "$TMP_DIR_LOCAL/detail_a.json" "$TMP_DIR_LOCAL/initial_ip_a" <<'PY'
+import json, sys, pathlib
+node=json.load(open(sys.argv[1]))
+ip=node.get("meta_data",{}).get("ip")
+assert ip, "missing ip"
+pathlib.Path(sys.argv[2]).write_text(ip)
+PY
+
+python3 - "$TMP_DIR_LOCAL/detail_b.json" "$TMP_DIR_LOCAL/initial_ip_b" <<'PY'
+import json, sys, pathlib
+node=json.load(open(sys.argv[1]))
+ip=node.get("meta_data",{}).get("ip")
+assert ip, "missing ip"
+pathlib.Path(sys.argv[2]).write_text(ip)
+PY
+
+NODE_JSON_A="$SYS_DEBUG_PRIVATE_NODEA/argus/agent/$HOST_A/node.json"
+NODE_JSON_B="$SYS_DEBUG_PRIVATE_NODEB/argus/agent/$HOST_B/node.json"
+
+[[ -f "$NODE_JSON_A" ]] || { echo "[ERR] node.json missing for $HOST_A" >&2; exit 1; }
+[[ -f "$NODE_JSON_B" ]] || { echo "[ERR] node.json missing for $HOST_B" >&2; exit 1; }
+
+log "Agents registered: $(cat "$TMP_DIR_LOCAL/node_id_a") , $(cat "$TMP_DIR_LOCAL/node_id_b")"

src/sys/debug/scripts/06_write_health_and_assert.sh

@@ -0,0 +1,78 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+ensure_paths_defined
+
+API_BASE="http://localhost:32300/api/v1/master"
+
+HEALTH_A="$SYS_DEBUG_PRIVATE_NODEA/argus/agent/$HOST_A/health"
+HEALTH_B="$SYS_DEBUG_PRIVATE_NODEB/argus/agent/$HOST_B/health"
+
+write_health() {
+  local dir="$1"; mkdir -p "$dir"
+  cat > "$dir/log-fluentbit.json" <<JSON
+{ "status": "healthy", "timestamp": "2025-10-13T12:05:00Z" }
+JSON
+  cat > "$dir/metric-node-exporter.json" <<JSON
+{ "status": "healthy", "timestamp": "2025-10-13T12:05:00Z" }
+JSON
+}
+
+log "Writing health files for both nodes"
+write_health "$HEALTH_A"
+write_health "$HEALTH_B"
+
+ID_A="$TMP_DIR/node_id_a"
+ID_B="$TMP_DIR/node_id_b"
+
+[[ -f "$ID_A" && -f "$ID_B" ]] || { echo "[ERR] node id files missing in $TMP_DIR" >&2; exit 1; }
+
+ID_A_VAL="$(cat "$ID_A")"
+ID_B_VAL="$(cat "$ID_B")"
+
+check_health() {
+  local id="$1"; local tries=40
+  for _ in $(seq 1 $tries); do
+    sleep 2
+    resp=$(curl -fsS "$API_BASE/nodes/$id" 2>/dev/null || true)
+    [[ -z "$resp" ]] && continue
+    echo "$resp" > "$TMP_DIR/node_${id}_detail.json"
+    if python3 - "$TMP_DIR/node_${id}_detail.json" <<'PY'
+import json,sys
+node=json.load(open(sys.argv[1]))
+h=node.get("health",{})
+if "log-fluentbit" in h and "metric-node-exporter" in h:
+    sys.exit(0)
+sys.exit(1)
+PY
+    then
+      return 0
+    fi
+  done
+  return 1
+}
+
+check_health "$ID_A_VAL" || { echo "[ERR] health keys not reported for node A" >&2; exit 1; }
+check_health "$ID_B_VAL" || { echo "[ERR] health keys not reported for node B" >&2; exit 1; }
+
+NODES_JSON="$SYS_DEBUG_PRIVATE_CORE/argus/metric/prometheus/nodes.json"
+if [[ ! -f "$NODES_JSON" ]]; then
+  echo "[ERR] nodes.json missing at $NODES_JSON" >&2
+  exit 1
+fi
+
+python3 - "$NODES_JSON" <<'PY'
+import json,sys
+with open(sys.argv[1]) as h:
+  nodes=json.load(h)
+if not isinstance(nodes, list):
+    raise SystemExit("nodes.json expected list")
+if len(nodes) != 2:
+    raise SystemExit(f"expected 2 nodes online, got {len(nodes)}")
+PY
+
+log "Health reported and nodes.json has 2 online nodes"

src/sys/debug/scripts/07_logs_send_and_assert.sh

@@ -0,0 +1,73 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+ensure_paths_defined
+
+log "Sending logs and asserting ES counts"
+
+get_count() {
+  local idx="$1"
+  curl -s "http://localhost:9200/${idx}/_count?ignore_unavailable=true&allow_no_indices=true" | sed -E 's/.*"count":([0-9]+).*/\1/' | awk 'NF{print $0;exit} END{if(NR==0)print 0}'
+}
+
+train0=$(get_count "train-*")
+infer0=$(get_count "infer-*")
+base=$((train0 + infer0))
+log "initial counts: train=${train0} infer=${infer0} total=${base}"
+
+service_id() {
+  compose ps -q "$1"
+}
+
+send_logs() {
+  local sid="$1"; local hosttag="$2"
+  docker exec "$sid" sh -lc 'mkdir -p /logs/train /logs/infer'
+  docker exec "$sid" sh -lc "ts=\
+\$(date '+%F %T'); echo \"\$ts INFO [$hosttag] training step=1 loss=1.23 model=bert\" >> /logs/train/train-demo.log"
+  docker exec "$sid" sh -lc "ts=\
+\$(date '+%F %T'); echo \"\$ts INFO [$hosttag] training step=2 loss=1.10 model=bert\" >> /logs/train/train-demo.log"
+  docker exec "$sid" sh -lc "ts=\
+\$(date '+%F %T'); echo \"\$ts WARN [$hosttag] inference slow on batch=2 latency=1.9s\" >> /logs/infer/infer-demo.log"
+}
+
+CID_A="$(service_id node-a)"
+CID_B="$(service_id node-b)"
+
+[[ -n "$CID_A" && -n "$CID_B" ]] || { echo "[ERR] node containers not found" >&2; exit 1; }
+
+send_logs "$CID_A" "host01"
+send_logs "$CID_B" "host02"
+
+log "Waiting for ES to ingest"
+sleep 10
+
+train1=$(get_count "train-*")
+infer1=$(get_count "infer-*")
+final=$((train1 + infer1))
+log "final counts: train=${train1} infer=${infer1} total=${final}"
+
+if (( final <= base )); then
+  echo "[ERR] ES total did not increase (${base} -> ${final})" >&2
+  exit 1
+fi
+
+if (( final < 4 )); then
+  echo "[ERR] ES total below expected threshold: ${final} < 4" >&2
+  exit 1
+fi
+
+es_health=$(curl -s "http://localhost:9200/_cluster/health" | grep -o '"status":"[^"]*"' | cut -d'"' -f4)
+if [[ "$es_health" != "green" && "$es_health" != "yellow" ]]; then
+  echo "[ERR] ES health not green/yellow: $es_health" >&2
+  exit 1
+fi
+
+if ! curl -fs "http://localhost:5601/api/status" >/dev/null 2>&1; then
+  echo "[WARN] Kibana status endpoint not available"
+fi
+
+log "ES counts increased and services healthy"

src/sys/debug/scripts/08_restart_agent_reregister.sh

@@ -0,0 +1,110 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+ensure_paths_defined
+
+API_BASE="http://localhost:32300/api/v1/master"
+NODE_ENTRYPOINT="$DEBUG_ROOT/../tests/scripts/node_entrypoint.sh"
+[[ -f "$NODE_ENTRYPOINT" ]] || { echo "[ERR] node entrypoint script missing at $NODE_ENTRYPOINT" >&2; exit 1; }
+
+TARGET_FIXED_IP="${SYS_DEBUG_NODEB_FIXED_IP:-172.30.0.200}"
+
+ID_B_FILE="$TMP_DIR/node_id_b"
+IP_INIT_FILE="$TMP_DIR/initial_ip_b"
+[[ -f "$ID_B_FILE" && -f "$IP_INIT_FILE" ]] || { echo "[ERR] Required node id/ip files missing in $TMP_DIR" >&2; exit 1; }
+
+ID_B="$(cat "$ID_B_FILE")"
+IP0_B="$(cat "$IP_INIT_FILE")"
+
+DETAIL_BEFORE="$TMP_DIR/node_b_before.json"
+curl -fsS "$API_BASE/nodes/$ID_B" -o "$DETAIL_BEFORE"
+LAST0=$(python3 - "$DETAIL_BEFORE" <<'PY'
+import json,sys
+node=json.load(open(sys.argv[1]))
+print(node.get("last_updated",""))
+PY
+)
+IP_BEFORE=$(python3 - "$DETAIL_BEFORE" <<'PY'
+import json,sys
+node=json.load(open(sys.argv[1]))
+print(node.get("meta_data",{}).get("ip",""))
+PY
+)
+
+if [[ "$IP_BEFORE" != "$IP0_B" ]]; then
+  echo "[ERR] Expected initial IP $IP0_B for node-b, got $IP_BEFORE" >&2
+  exit 1
+fi
+
+if [[ "$IP_BEFORE" == "$TARGET_FIXED_IP" ]]; then
+  echo "[ERR] node-b current IP $IP_BEFORE already matches target $TARGET_FIXED_IP. Configure SYS_DEBUG_NODEB_FIXED_IP to a different address before rerun." >&2
+  exit 1
+fi
+
+service_id() {
+  compose ps -q "$1"
+}
+
+log "Recreating node-b (old IP $IP_BEFORE) with static IP $TARGET_FIXED_IP"
+compose rm -sf node-b >/dev/null 2>&1 || true
+
+CONTAINER_NAME="${SYS_DEBUG_CONTAINER_PREFIX:-argus-debug}-node-b"
+docker rm -f "$CONTAINER_NAME" >/dev/null 2>&1 || true
+
+AGENT_BIN_PATH="$(cat "$TMP_DIR/agent_binary_path")"
+[[ -f "$AGENT_BIN_PATH" ]] || { echo "[ERR] Agent binary path missing in $TMP_DIR" >&2; exit 1; }
+
+require_docker
+
+docker run -d \
+  --name "$CONTAINER_NAME" \
+  --hostname "$HOST_B" \
+  --network "$SYS_DEBUG_NETWORK_NAME" \
+  --ip "$TARGET_FIXED_IP" \
+  --dns "${SYS_DEBUG_BIND_IP:-172.30.0.2}" \
+  -e MASTER_ENDPOINT=http://master.argus.com:3000 \
+  -e REPORT_INTERVAL_SECONDS=2 \
+  -e ARGUS_BUILD_UID=$ARGUS_BUILD_UID \
+  -e ARGUS_BUILD_GID=$ARGUS_BUILD_GID \
+  -e ES_HOST=es \
+  -e ES_PORT=9200 \
+  -e CLUSTER=local \
+  -e RACK=dev \
+  -p 2021:2020 \
+  -v "$SYS_DEBUG_PRIVATE_NODEB/argus/agent/$HOST_B:/private/argus/agent/$HOST_B" \
+  -v "$AGENT_BIN_PATH:/usr/local/bin/argus-agent:ro" \
+  -v "$NODE_ENTRYPOINT:/usr/local/bin/node-entrypoint.sh:ro" \
+  -v "$REPO_ROOT/src/log/fluent-bit/build/start-fluent-bit.sh:/assets/start-fluent-bit.sh:ro" \
+  -v "$REPO_ROOT/src/log/fluent-bit/build/etc:/assets/fluent-bit/etc:ro" \
+  -v "$REPO_ROOT/src/log/fluent-bit/build/packages:/assets/fluent-bit/packages:ro" \
+  --entrypoint /usr/local/bin/node-entrypoint.sh \
+  ubuntu:22.04 >/dev/null
+
+log "Waiting for node-b to re-register with new IP"
+for _ in {1..40}; do
+  sleep 3
+  if curl -fsS "$API_BASE/nodes/$ID_B" -o "$TMP_DIR/node_b_after.json"; then
+    if python3 - "$TMP_DIR/node_b_after.json" "$LAST0" "$TARGET_FIXED_IP" <<'PY'
+import json,sys
+node=json.load(open(sys.argv[1]))
+last0=sys.argv[2]
+expected_ip=sys.argv[3]
+ip=node.get("meta_data",{}).get("ip")
+lu=node.get("last_updated")
+if ip == expected_ip and lu and lu != last0:
+    sys.exit(0)
+sys.exit(1)
+PY
+    then
+      log "node-b IP updated: $IP_BEFORE -> $TARGET_FIXED_IP"
+      exit 0
+    fi
+  fi
+done
+
+echo "[ERR] node-b did not update to IP $TARGET_FIXED_IP in time" >&2
+exit 1

src/sys/debug/scripts/09_down.sh

@@ -0,0 +1,13 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+require_docker
+
+log "Stopping debug stack (project $SYS_DEBUG_PROJECT_NAME)"
+compose down --remove-orphans >/dev/null 2>&1 || true
+
+log "Containers stopped. No host directories were removed."

src/sys/debug/scripts/clean-data.sh

@@ -0,0 +1,66 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+ensure_env_file
+ensure_paths_defined
+
+FORCE=false
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    -y|--yes)
+      FORCE=true
+      ;;
+    -h|--help)
+      cat <<USAGE
+Usage: ${0##*/} [--yes]
+
+Safely remove debug private directories after adjusting ownership.
+USAGE
+      exit 0
+      ;;
+    *)
+      echo "Unknown argument: $1" >&2
+      exit 1
+      ;;
+  esac
+  shift
+done
+
+if [[ $FORCE == false ]]; then
+  read -r -p "This will delete debug private directories. Continue? [y/N] " reply
+  case "$reply" in
+    y|Y|yes|YES)
+      ;;
+    *)
+      echo "Aborted"
+      exit 0
+      ;;
+  esac
+fi
+
+paths=(
+  "$SYS_DEBUG_PRIVATE_CORE"
+  "$SYS_DEBUG_PRIVATE_NODEA"
+  "$SYS_DEBUG_PRIVATE_NODEB"
+  "$SYS_DEBUG_TMP_DIR"
+)
+
+require_docker
+
+image="ubuntu:22.04"
+
+for dir in "${paths[@]}"; do
+  [[ -d "$dir" ]] || continue
+  log "Fixing ownership for $dir"
+  if ! docker run --rm -v "$dir:/target" "$image" chown -R "$(id -u):$(id -g)" /target >/dev/null 2>&1; then
+    echo "[WARN] Failed to adjust ownership via $image, attempting local chown" >&2
+    chown -R "$(id -u):$(id -g)" "$dir" >/dev/null 2>&1 || true
+  fi
+  log "Removing $dir"
+  rm -rf "$dir"
+done
+
+log "Clean data completed"

src/sys/debug/scripts/common.sh

@@ -0,0 +1,96 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+DEBUG_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+REPO_ROOT="$(cd "$DEBUG_ROOT/../../.." && pwd)"
+ENV_FILE="$DEBUG_ROOT/.env"
+
+source "$REPO_ROOT/scripts/common/build_user.sh"
+load_build_user
+
+if [[ -f "$ENV_FILE" ]]; then
+  set -a
+  # shellcheck disable=SC1090
+  source "$ENV_FILE"
+  set +a
+fi
+
+SYS_DEBUG_NETWORK_NAME=${SYS_DEBUG_NETWORK_NAME:-argus-debug-net}
+SYS_DEBUG_NETWORK_SUBNET=${SYS_DEBUG_NETWORK_SUBNET:-172.30.0.0/16}
+SYS_DEBUG_NETWORK_GATEWAY=${SYS_DEBUG_NETWORK_GATEWAY:-172.30.0.1}
+SYS_DEBUG_PROJECT_NAME=${SYS_DEBUG_PROJECT_NAME:-argus-debug}
+SYS_DEBUG_CONTAINER_PREFIX=${SYS_DEBUG_CONTAINER_PREFIX:-argus-debug}
+SYS_DEBUG_PRIVATE_CORE=${SYS_DEBUG_PRIVATE_CORE:-$DEBUG_ROOT/private}
+SYS_DEBUG_PRIVATE_NODEA=${SYS_DEBUG_PRIVATE_NODEA:-$DEBUG_ROOT/private-nodea}
+SYS_DEBUG_PRIVATE_NODEB=${SYS_DEBUG_PRIVATE_NODEB:-$DEBUG_ROOT/private-nodeb}
+SYS_DEBUG_TMP_DIR=${SYS_DEBUG_TMP_DIR:-$DEBUG_ROOT/tmp}
+ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
+
+SYS_DEBUG_NODEA_HOST=${SYS_DEBUG_NODEA_HOST:-dev-yyrshare-nbnyx10-cp2f-pod-0}
+SYS_DEBUG_NODEB_HOST=${SYS_DEBUG_NODEB_HOST:-dev-yyrshare-uuuu10-ep2f-pod-0}
+
+HOST_A="$SYS_DEBUG_NODEA_HOST"
+HOST_B="$SYS_DEBUG_NODEB_HOST"
+
+COMPOSE_FILE="$DEBUG_ROOT/docker-compose.yml"
+
+abs_path() {
+  python3 - "$1" <<'PY'
+import os, sys
+path = sys.argv[1]
+print(os.path.abspath(path))
+PY
+}
+
+ensure_command() {
+  local cmd="$1"
+  if ! command -v "$cmd" >/dev/null 2>&1; then
+    echo "[ERR] Required command '$cmd' not found" >&2
+    exit 1
+  fi
+}
+
+require_docker() {
+  ensure_command docker
+}
+
+compose() {
+  require_docker
+  local bin
+  if docker compose version >/dev/null 2>&1; then
+    bin=(docker compose)
+  else
+    bin=(docker-compose)
+  fi
+  "${bin[@]}" -p "$SYS_DEBUG_PROJECT_NAME" -f "$COMPOSE_FILE" "$@"
+}
+
+ensure_paths_defined() {
+  local missing=()
+  for name in SYS_DEBUG_PRIVATE_CORE SYS_DEBUG_PRIVATE_NODEA SYS_DEBUG_PRIVATE_NODEB SYS_DEBUG_TMP_DIR; do
+    if [[ -z "${!name:-}" ]]; then
+      missing+=("$name")
+    fi
+  done
+  if (( ${#missing[@]} > 0 )); then
+    echo "[ERR] Missing required environment variables: ${missing[*]}" >&2
+    echo "      Run 01_bootstrap.sh first." >&2
+    exit 1
+  fi
+}
+
+ensure_env_file() {
+  if [[ ! -f "$ENV_FILE" ]]; then
+    echo "[ERR] Missing .env at $ENV_FILE. Run 01_bootstrap.sh first." >&2
+    exit 1
+  fi
+}
+
+log() {
+  echo "[INFO] $*"
+}
+
+TMP_DIR="$SYS_DEBUG_TMP_DIR"
+mkdir -p "$TMP_DIR"

src/sys/debug/scripts/network-create.sh

@@ -0,0 +1,76 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+NAME="$SYS_DEBUG_NETWORK_NAME"
+SUBNET="$SYS_DEBUG_NETWORK_SUBNET"
+GATEWAY="$SYS_DEBUG_NETWORK_GATEWAY"
+
+usage() {
+  cat <<EOF
+Usage: ${0##*/} [--name NAME] [--subnet CIDR] [--gateway IP]
+
+Create (if missing) the external debug docker network.
+
+Defaults derived from .env or:
+  name    = $NAME
+  subnet  = $SUBNET
+  gateway = $GATEWAY
+EOF
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --name)
+      shift; [[ $# -gt 0 ]] || { echo "--name requires value" >&2; exit 1; }
+      NAME="$1"
+      ;;
+    --name=*)
+      NAME="${1#*=}"
+      ;;
+    --subnet)
+      shift; [[ $# -gt 0 ]] || { echo "--subnet requires value" >&2; exit 1; }
+      SUBNET="$1"
+      ;;
+    --subnet=*)
+      SUBNET="${1#*=}"
+      ;;
+    --gateway)
+      shift; [[ $# -gt 0 ]] || { echo "--gateway requires value" >&2; exit 1; }
+      GATEWAY="$1"
+      ;;
+    --gateway=*)
+      GATEWAY="${1#*=}"
+      ;;
+    -h|--help)
+      usage
+      exit 0
+      ;;
+    *)
+      echo "Unknown argument: $1" >&2
+      usage >&2
+      exit 1
+      ;;
+  esac
+  shift
+done
+
+require_docker
+
+if docker network inspect "$NAME" >/dev/null 2>&1; then
+  log "Network $NAME already exists"
+  exit 0
+fi
+
+log "Creating network $NAME (subnet=$SUBNET gateway=$GATEWAY)"
+docker network create \
+  --driver bridge \
+  --subnet "$SUBNET" \
+  --gateway "$GATEWAY" \
+  "$NAME"
+
+mkdir -p "$TMP_DIR"
+echo "$NAME" > "$TMP_DIR/network.created"
+log "Network $NAME created"

src/sys/debug/scripts/network-destroy.sh

@@ -0,0 +1,55 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# shellcheck source=common.sh
+source "$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/common.sh"
+
+NAME="$SYS_DEBUG_NETWORK_NAME"
+
+usage() {
+  cat <<EOF
+Usage: ${0##*/} [--name NAME]
+
+Destroy the debug docker network if no containers are attached.
+EOF
+}
+
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --name)
+      shift; [[ $# -gt 0 ]] || { echo "--name requires value" >&2; exit 1; }
+      NAME="$1"
+      ;;
+    --name=*)
+      NAME="${1#*=}"
+      ;;
+    -h|--help)
+      usage
+      exit 0
+      ;;
+    *)
+      echo "Unknown argument: $1" >&2
+      usage >&2
+      exit 1
+      ;;
+  esac
+  shift
+done
+
+require_docker
+
+if ! docker network inspect "$NAME" >/dev/null 2>&1; then
+  log "Network $NAME not found; nothing to do"
+  exit 0
+fi
+
+attached=$(docker network inspect -f '{{range $id, $conf := .Containers}}{{printf "%s " $conf.Name}}{{end}}' "$NAME")
+if [[ -n "${attached// }" ]]; then
+  echo "[ERR] Cannot remove network $NAME: still connected containers -> $attached" >&2
+  exit 1
+fi
+
+log "Deleting network $NAME"
+docker network rm "$NAME" >/dev/null
+rm -f "$TMP_DIR/network.created"
+log "Network $NAME removed"

src/sys/tests/scripts/08_restart_agent_reregister.sh

@@ -8,6 +8,16 @@ REPO_ROOT="$(cd "$TEST_ROOT/../../.." && pwd)"
 
 API_BASE="http://localhost:32300/api/v1/master"
 
+if [[ -f "$TEST_ROOT/.env" ]]; then
+  set -a
+  # shellcheck disable=SC1090
+  source "$TEST_ROOT/.env"
+  set +a
+else
+  source "$REPO_ROOT/scripts/common/build_user.sh"
+  load_build_user
+fi
+
 ID_B="$(cat "$TMP_DIR/node_id_b")"
 IP0_B="$(cat "$TMP_DIR/initial_ip_b")"