dev_1.0.0_sundp_2 优化Argus-metric模块的e2e部署测试流程 (#27)

Co-authored-by: sundapeng.sdp <sundapeng@hashdata.cn> Reviewed-on: #27 Reviewed-by: yuyr <yuyr@zgclab.edu.cn> Reviewed-by: xuxt <xuxt@zgclab.edu.cn>
2025-10-17 17:15:55 +08:00
parent 31ccb0b1b8
commit 1a768bc837
36 changed files with 1805 additions and 695 deletions
--- a/src/metric/tests/.gitignore
+++ b/src/metric/tests/.gitignore
@@ -1,7 +1,7 @@
 .env
 data/
 images-cache/
+private-test-node/
 *.tar
 *.log
 .DS_Store
-
--- a/src/metric/tests/README.md
+++ b/src/metric/tests/README.md
@@ -1,171 +1,97 @@
 # E2E Test - Argus Metric 部署测试
-## 概述
+## 1. 概述

 本项目用于对 Argus Metric 模块进行端到端（E2E）部署测试。
 通过一键脚本可快速搭建 Prometheus、FTP、Grafana 等服务，验证 Metric 模块的完整部署与运行流程。

-## 拉取完整项目，进入 metric.tests 目录
+功能包括：

-``` bash
-git clone https://git.nasp.fit/NASP/argus.git
+- 自动启动所需服务和测试节点
+- 发布安装包到 FTP
+- CPU/GPU 节点客户端安装测试
+- 验证安装结果与服务可用性
+- 支持环境清理和分步调试

-cd {$PROJECT_ROOT}/argus/src/metric/tests
+## 2. 前置条件
+
+在开始部署和测试之前，请确保完成以下准备工作：
+
+### 2.1 检查 all-in-one-full 客户端安装包
+确认客户端安装包目录是否存在：
+```bash
+{$PROJECT_ROOT}/argus/src/metric/client-plugins/all-in-one-full
 ```
+本项目依赖完整的 all-in-one-full 安装包，其中包含大量二进制文件、依赖包和测试制品，由于体积较大，无法直接上传到 Git 仓库。**请联系项目管理员获取最新版本的完整框架。**

-## 一键构建与部署 Prometheus / FTP / Grafana
-### 1. 修改环境变量文件
-
-将示例配置文件复制为 .env 并根据实际情况修改：
-
-``` bash
+### 2.2 配置环境变量
+查看配置文件是否存在，如不存在，则复制示例配置文件并根据实际环境修改：
+```bash
+cd {$PROJECT_ROOT}/argus/src/metric/tests
 cp env.example .env
 ```
+.env 文件用于指定构建UID:GID、FTP 配置、版本号等信息，确保各脚本运行时可以正确访问资源。

-### 2. 一键启动服务
+### 2.3 离线镜像准备
+  - 步骤1：在**在线服务器**执行以下脚本，会拉取和构建所需的 Docker 镜像：
+      ``` bash 
+      cd {$PROJECT_ROOT}/argus/src/metric/tests
+      bash scripts/01_start_services.sh
+      bash scripts/save-images.sh
+      ```
+  - 步骤2：镜像将被保存到 metric.tests.images-cache 目录中，用于离线迁移和后续导入。
+  - 步骤3：若目标服务器无法联网，可将该目录拷贝到离线服务器，并执行：
+      ``` bash 
+      cd {$PROJECT_ROOT}/argus/src/metric/tests
+      bash scripts/load-images.sh
+      ```
+  - 即可导入镜像并执行下面的QuickStart或分步操作。

-执行以下命令完成环境初始化、镜像构建与服务启动：
+## 3. QuickStart

-``` bash
-sudo bash start-all.sh
+执行完整的端到端测试流程：
+
+```bash
+bash scripts/00_e2e_test.sh
 ```

-该脚本将自动完成：
- 初始化目录结构（如 /private/argus/metric）
- 构建各服务 Docker 镜像
- 启动 Prometheus、FTP、Grafana 容器
+该脚本将自动执行以下步骤：
+1. 启动所有服务（Prometheus、FTP、Grafana、测试节点）
+2. 发布安装包到 FTP 服务
+3. 在 CPU 测试节点上安装客户端
+4. 在 GPU 测试节点上安装客户端
+5. 验证安装结果
+6. 清理测试环境

-### 3. 检查容器日志
+## 4. 分步执行

-可手动验证容器运行状态：
+| 步骤         | 脚本                                      | 功能描述                                               |
+|--------------|-------------------------------------------|--------------------------------------------------------|
+| 启动基础服务 | bash scripts/01_start_services.sh         | 构建 Docker 镜像、创建持久化目录、启动容器服务          |
+| 发布安装包   | bash scripts/02_publish_artifact.sh       | 自动递增版本号、打包安装制品、发布到 FTP                |
+| CPU 节点安装 | bash scripts/03_test_node_install.sh      | 在 CPU 节点下载安装程序并执行安装                       |
+| GPU 节点安装 | bash scripts/04_test_gpu_node_install.sh  | 在 GPU 节点下载安装程序并执行安装                       |
+| 验证安装     | bash scripts/05_verify_install.sh         | 检查监控端口、端口连通性及服务可用性                    |
+| 清理环境     | bash scripts/06_cleanup.sh                | 停止并清理所有测试容器及环境                            |

-``` bash
-docker logs argus-metric-ftp
-docker logs argus-metric-grafana
-docker logs argus-metric-prometheus
-```
+## 5. 查看监控采集数据及展示面板

-如日志输出中无 ERROR 或 supervisor 报错信息，则表示服务启动正常。
-
-## 客户端安装包打包与分发
-
-> **前置说明**：完整的 `all-in-one` 安装包打包分发框架因包含大量二进制文件和依赖包，无法上传至 Git 仓库。请先联系项目管理员获取最新的 `all-in-one` 完整框架，再执行后续操作。
-
-打包后服务端会将安装包发布至 FTP 共享目录，默认路径为：
-
-``` bash
-$DATA_ROOT/ftp/share
-```
-
-发布后的文件权限与 FTP 目录账户保持一致。
-
-### 1. 递增版本号
-``` bash
-bash scripts/version-manager.sh bump minor
-```
-该脚本会自动更新版本号（如 1.101.0 → 1.102.0）。
-
-### 2. 打包安装制品
-``` bash
-bash scripts/package_artifact.sh
-```
-执行后会在输出目录中生成压缩包或安装脚本。
-
-### 3. 发布制品至 FTP
-``` bash
-sudo bash scripts/publish_artifact.sh $VERSION --output-dir $OUTPUT_DIR --owner $UID:$GID
-```
-
-参数说明：
-
-参数	说明
-$VERSION	发布版本号（如 1.102.0）
-$OUTPUT_DIR	输出目录（默认 /private/argus/ftp/share）
-$UID:$GID	文件属主（用户ID:组ID）
-
-示例：
-
-``` bash
-sudo bash scripts/publish_artifact.sh 1.102.0 --output-dir /private/argus/ftp/share --owner 2133:2015
-```
-
-更多详情可参考 client-plugins/all-in-one/README.md。
-
-## 客户端安装（通过 FTP）
-
-客户端下载与安装步骤如下：
-
-``` bash
-curl -u ${USER}:${PASSWD} ftp://${FTP_SERVER}/setup.sh -o setup.sh
-chmod +x setup.sh
-sudo bash setup.sh --server ${FTP_SERVER} --user ${USER} --password ${PASSWD} --port ${PORT}
-```
-
-参数说明：
-
-参数	说明
-$FTP_SERVER	服务器 IP 地址
-$USER	默认 ftpuser
-$PASSWD	默认 ZGClab1234!
-$PORT	FTP 服务端口（需与 .env 保持一致）
-
-示例：
-
-``` bash
-curl -u ftpuser:ZGClab1234! ftp://10.211.55.4/setup.sh -o setup.sh
-chmod +x setup.sh
-sudo bash setup.sh --server 10.211.55.4 --user ftpuser --password 'ZGClab1234!' --port 2122
-```
-
-更多细节可参考 client-plugins/all-in-one/README.md。
-
-## 模拟 Argus-Master 配置下发
-
-可通过手动写入 nodes.json 文件模拟 Argus-Master 对 Argus-Metric 的配置下发：
-
-``` json
-[
-  {
-    "node_id": "A1",
-    "user_id": "sundapeng",
-    "ip": "10.211.55.4",
-    "hostname": "dev-sundapeng-nsche-wohen-pod-0",
-    "labels": ["label-a", "label-b"]
-  }
-]
-```
-
-路径：
-
-``` bash
-${DATA_ROOT}/prometheus/nodes.json
-```
-
-Argus-Metric 中的 prometheus 模块会自动解析该文件，并将其拆分生成目标配置：
-
-``` bash
-${DATA_ROOT}/prometheus/targets/
-```
-
-## Grafana 手动配置（如未自动接入 Prometheus）
-
-如 Grafana 未自动导入 Prometheus 数据源，可手动执行以下操作：
-
-1. 添加数据源
- 进入 Grafana → Data sources
- 选择 Add data source → Prometheus
- URL 填写：http://prometheus:9090(Docker 内部 DNS 地址)
-
-2. 导入测试 Dashboard
- 打开 Grafana → Dashboards → Import
- 上传或粘贴 test_grafana_dashboard.json
-
-## 查看监控数据
 Prometheus 访问以下地址查看节点活性：
 ``` bash
-http://127.0.0.1:9091/targets
+http://127.0.0.1:9090/targets
 ```

 Grafana 访问以下地址查看监控大屏：
 ``` bash
 http://127.0.0.1:3000/d/node_gpu_metrics/node-and-gpu-metrics
-```
+```
+
+PS: 如果 Grafana 未自动导入 Prometheus 数据源，可手动执行以下操作：
+
+1. 添加数据源
+- 进入 Grafana → Data sources
+- 选择 Add data source → Prometheus
+- URL 填写：http://prom.metric.argus.com:9090
+
+2. 导入测试 Dashboard
+- 打开 Grafana → Dashboards → Import
+- 上传或粘贴 test_grafana_dashboard.json
--- a/src/metric/tests/client-test-gpu-node/build/Dockerfile
+++ b/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"]
--- a/src/metric/tests/client-test-node/build/Dockerfile
+++ b/src/metric/tests/client-test-node/build/Dockerfile
@@ -0,0 +1,6 @@
+FROM ubuntu:22.04
+RUN apt-get update -qq && \
+    DEBIAN_FRONTEND=noninteractive apt-get install -y -qq tzdata && \
+    rm -rf /var/lib/apt/lists/*
+ENV TZ=Asia/Shanghai
+
--- a/src/metric/tests/docker-compose.yml
+++ b/src/metric/tests/docker-compose.yml
@@ -1,29 +1,39 @@
+networks:
+  default:
+    name: argus-debug-net
+    external: true
+
 services:
  ftp:
    build:
      context: ../ftp/build
      dockerfile: Dockerfile
      args:
-        FTP_UID: ${FTP_UID:-2133}
-        FTP_GID: ${FTP_GID:-2015}
+        ARGUS_BUILD_UID: ${ARGUS_BUILD_UID:-2133}
+        ARGUS_BUILD_GID: ${ARGUS_BUILD_GID:-2015}
+        USE_INTRANET: ${USE_INTRANET:-false}
    image: argus-metric-ftp:latest
    container_name: argus-ftp
    restart: unless-stopped
    environment:
+      - TZ=Asia/Shanghai
      - FTP_BASE_PATH=/private/argus/ftp
      - FTP_PASSWORD=${FTP_PASSWORD:-ZGClab1234!}
-      - DOMAIN=${FTP_DOMAIN:-prom.ftp.argus.com}
-      - FTP_UID=${FTP_UID:-2133}
-      - FTP_GID=${FTP_GID:-2015}
+      - DOMAIN=${FTP_DOMAIN:-ftp.metric.argus.com}
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
    ports:
      - "${FTP_PORT:-21}:21"
      - "${FTP_DATA_PORT:-20}:20"
      - "21100-21110:21100-21110"
    volumes:
-      - ${DATA_ROOT:-./data}/ftp:/private/argus/ftp
-      - ${DATA_ROOT:-./data}/etc:/private/argus/etc
+      - ${DATA_ROOT:-/private}/argus/metric/ftp:/private/argus/ftp
+      - ${DATA_ROOT:-/private}/argus/etc:/private/argus/etc
+      - /etc/localtime:/etc/localtime:ro
+      - /etc/timezone:/etc/timezone:ro
    networks:
-      - argus-network
+      default:
+        ipv4_address: 172.30.0.40
    logging:
      driver: "json-file"
      options:
@@ -35,23 +45,27 @@ services:
      context: ../prometheus/build
      dockerfile: Dockerfile
      args:
-        PROMETHEUS_UID: ${PROMETHEUS_UID:-2133}
-        PROMETHEUS_GID: ${PROMETHEUS_GID:-2015}
+        ARGUS_BUILD_UID: ${ARGUS_BUILD_UID:-2133}
+        ARGUS_BUILD_GID: ${ARGUS_BUILD_GID:-2015}
        USE_INTRANET: ${USE_INTRANET:-false}
    image: argus-metric-prometheus:latest
    container_name: argus-prometheus
    restart: unless-stopped
    environment:
+      - TZ=Asia/Shanghai
      - PROMETHEUS_BASE_PATH=/private/argus/metric/prometheus
-      - PROMETHEUS_UID=${PROMETHEUS_UID:-2133}
-      - PROMETHEUS_GID=${PROMETHEUS_GID:-2015}
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
    ports:
      - "${PROMETHEUS_PORT:-9090}:9090"
    volumes:
-      - ${DATA_ROOT:-./data}/prometheus:/private/argus/metric/prometheus
-      - ${DATA_ROOT:-./data}/etc:/private/argus/etc
+      - ${DATA_ROOT:-/private}/argus/metric/prometheus:/private/argus/metric/prometheus
+      - ${DATA_ROOT:-/private}/argus/etc:/private/argus/etc
+      - /etc/localtime:/etc/localtime:ro
+      - /etc/timezone:/etc/timezone:ro
    networks:
-      - argus-network
+      default:
+        ipv4_address: 172.30.0.41
    logging:
      driver: "json-file"
      options:
@@ -63,25 +77,29 @@ services:
      context: ../grafana/build
      dockerfile: Dockerfile
      args:
-        GRAFANA_UID: ${GRAFANA_UID:-2133}
-        GRAFANA_GID: ${GRAFANA_GID:-2015}
+        ARGUS_BUILD_UID: ${ARGUS_BUILD_UID:-2133}
+        ARGUS_BUILD_GID: ${ARGUS_BUILD_GID:-2015}
    image: argus-metric-grafana:latest
    container_name: argus-grafana
    restart: unless-stopped
    environment:
+      - TZ=Asia/Shanghai
      - GRAFANA_BASE_PATH=/private/argus/metric/grafana
-      - GRAFANA_UID=${GRAFANA_UID:-2133}
-      - GRAFANA_GID=${GRAFANA_GID:-2015}
+      - ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+      - ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
      - GF_SERVER_HTTP_PORT=3000
      - GF_LOG_LEVEL=warn
      - GF_LOG_MODE=console
    ports:
      - "${GRAFANA_PORT:-3000}:3000"
    volumes:
-      - ${DATA_ROOT:-./data}/grafana:/private/argus/metric/grafana
-      - ${DATA_ROOT:-./data}/etc:/private/argus/etc
+      - ${DATA_ROOT:-/private}/argus/metric/grafana:/private/argus/metric/grafana
+      - ${DATA_ROOT:-/private}/argus/etc:/private/argus/etc
+      - /etc/localtime:/etc/localtime:ro
+      - /etc/timezone:/etc/timezone:ro
    networks:
-      - argus-network
+      default:
+        ipv4_address: 172.30.0.42
    depends_on:
      - prometheus
    logging:
@@ -90,16 +108,78 @@ services:
        max-size: "10m"
        max-file: "3"

-networks:
-  argus-network:
-    driver: bridge
-    name: argus-network
+  test-node:
+    build:
+      context: ./client-test-node/build
+      dockerfile: Dockerfile
+    image: argus-metric-test-node:latest
+    container_name: argus-metric-test-node
+    hostname: test-metric-node-001
+    restart: unless-stopped
+    privileged: true
+    depends_on:
+      - ftp
+      - prometheus
+    environment:
+      - TZ=Asia/Shanghai
+      - DEBIAN_FRONTEND=noninteractive
+      - FTP_DOMAIN=${FTP_DOMAIN:-ftp.metric.argus.com}
+      - FTP_SERVER=${FTP_SERVER:-172.30.0.40}
+      - FTP_USER=${FTP_USER:-ftpuser}
+      - FTP_PASSWORD=${FTP_PASSWORD:-ZGClab1234!}
+      - FTP_PORT=${FTP_PORT:-21}
+    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: 172.30.0.50
+    logging:
+      driver: "json-file"
+      options:
+        max-size: "10m"
+        max-file: "3"

-volumes:
-  ftp_data:
-    driver: local
-  prometheus_data:
-    driver: local
-  grafana_data:
-    driver: local
+  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"

--- a/src/metric/tests/env.example
+++ b/src/metric/tests/env.example
@@ -1,19 +1,15 @@
-# 用户和组配置
-FTP_UID=2133
-FTP_GID=2015
-PROMETHEUS_UID=2133
-PROMETHEUS_GID=2015
-GRAFANA_UID=2133
-GRAFANA_GID=2015
+# 统一用户和组配置
+ARGUS_BUILD_UID=1048
+ARGUS_BUILD_GID=1048

 # 数据根目录
-DATA_ROOT=/private/argus
+DATA_ROOT=/private

 # FTP 配置
-FTP_PORT=2122
-FTP_DATA_PORT=2022
+FTP_PORT=21
+FTP_DATA_PORT=20
 FTP_PASSWORD=ZGClab1234!
-FTP_DOMAIN=prom.ftp.argus.com
+FTP_DOMAIN=ftp.metric.argus.com

 # Prometheus 配置
 PROMETHEUS_PORT=9090
--- a/src/metric/tests/init-directories.sh
+++ b/src/metric/tests/init-directories.sh
@@ -1,90 +0,0 @@
-#!/bin/bash
-
-# 初始化目录脚本
-# 用于创建所有必要的数据目录并设置正确的权限
-
-set -e
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-cd "$SCRIPT_DIR"
-
-# 加载 .env 文件（如果存在）
-if [ -f .env ]; then
-    echo "加载 .env 配置文件..."
-    source .env
-fi
-
-# 默认配置
-FTP_UID=${FTP_UID:-2133}
-FTP_GID=${FTP_GID:-2015}
-PROMETHEUS_UID=${PROMETHEUS_UID:-2133}
-PROMETHEUS_GID=${PROMETHEUS_GID:-2015}
-GRAFANA_UID=${GRAFANA_UID:-2133}
-GRAFANA_GID=${GRAFANA_GID:-2015}
-DATA_ROOT=${DATA_ROOT:-./data}
-
-echo "开始初始化目录结构..."
-echo "数据目录: ${DATA_ROOT}"
-echo ""
-
-# 创建 FTP 目录
-echo "创建 FTP 目录..."
-sudo mkdir -p ${DATA_ROOT}/ftp/share
-sudo chown -R ${FTP_UID}:${FTP_GID} ${DATA_ROOT}/ftp
-sudo chmod -R 755 ${DATA_ROOT}/ftp
-
-# 创建 Prometheus 目录
-echo "创建 Prometheus 目录..."
-sudo mkdir -p ${DATA_ROOT}/prometheus/{data,rules,targets}
-
-# 创建默认的 targets 文件（先创建文件再改权限）
-if [ ! -f "${DATA_ROOT}/prometheus/targets/node_exporter.json" ]; then
-    echo "创建默认 node_exporter targets..."
-    echo '[
-  {
-    "targets": [],
-    "labels": {
-      "job": "node"
-    }
-  }
-]' | sudo tee ${DATA_ROOT}/prometheus/targets/node_exporter.json > /dev/null
-fi
-
-if [ ! -f "${DATA_ROOT}/prometheus/targets/dcgm_exporter.json" ]; then
-    echo "创建默认 dcgm_exporter targets..."
-    echo '[
-  {
-    "targets": [],
-    "labels": {
-      "job": "dcgm"
-    }
-  }
-]' | sudo tee ${DATA_ROOT}/prometheus/targets/dcgm_exporter.json > /dev/null
-fi
-
-# 统一设置 Prometheus 目录权限
-sudo chown -R ${PROMETHEUS_UID}:${PROMETHEUS_GID} ${DATA_ROOT}/prometheus
-sudo chmod -R 755 ${DATA_ROOT}/prometheus
-
-# 创建 Grafana 目录
-echo "创建 Grafana 目录..."
-sudo mkdir -p ${DATA_ROOT}/grafana/{data,logs,plugins,provisioning/datasources,provisioning/dashboards,data/sessions,data/dashboards,config}
-sudo chown -R ${GRAFANA_UID}:${GRAFANA_GID} ${DATA_ROOT}/grafana
-sudo chmod -R 755 ${DATA_ROOT}/grafana
-
-# 创建公共配置目录
-sudo mkdir -p ${DATA_ROOT}/etc
-sudo chown -R ${FTP_UID}:${FTP_GID} ${DATA_ROOT}/etc
-sudo chmod -R 755 ${DATA_ROOT}/etc
-
-echo "目录初始化完成！"
-echo ""
-echo "目录结构："
-echo "  ${DATA_ROOT}/"
-echo "  ├── ftp/           (UID:${FTP_UID}, GID:${FTP_GID})"
-echo "  ├── prometheus/    (UID:${PROMETHEUS_UID}, GID:${PROMETHEUS_GID})"
-echo "  ├── grafana/       (UID:${GRAFANA_UID}, GID:${GRAFANA_GID})"
-echo "  └── etc/           (UID:${FTP_UID}, GID:${FTP_GID})"
-echo ""
-echo "您现在可以运行 'docker-compose up -d' 来启动所有服务"
-
--- a/src/metric/tests/scripts/00_e2e_test.sh
+++ b/src/metric/tests/scripts/00_e2e_test.sh
@@ -0,0 +1,20 @@
+#!/bin/bash
+set -e
+
+SCRIPT_DIR="$(dirname "$0")"
+
+echo "=========================================="
+echo "Argus Metric E2E Test"
+echo "=========================================="
+
+bash "$SCRIPT_DIR/01_start_services.sh"
+bash "$SCRIPT_DIR/02_publish_artifact.sh"
+bash "$SCRIPT_DIR/03_test_node_install.sh"
+bash "$SCRIPT_DIR/04_test_gpu_node_install.sh"
+bash "$SCRIPT_DIR/05_verify_install.sh"
+bash "$SCRIPT_DIR/06_cleanup.sh"
+
+echo "=========================================="
+echo "E2E 测试完成"
+echo "=========================================="
+
--- a/src/metric/tests/scripts/01_start_services.sh
+++ b/src/metric/tests/scripts/01_start_services.sh
@@ -0,0 +1,27 @@
+#!/bin/bash
+set -e
+
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+
+# 解析参数
+REBUILD_FLAG=""
+if [[ "$1" == "--rebuild" || "$1" == "-r" ]]; then
+    REBUILD_FLAG="--rebuild"
+    echo "[01] 启用强制重新构建模式"
+fi
+
+echo "[01] 启动所有服务..."
+bash "$SCRIPT_DIR/common/start-all.sh" $REBUILD_FLAG
+
+echo "[01] 等待服务就绪..."
+sleep 5
+
+echo "[01] 检查服务状态..."
+docker ps | grep argus-ftp
+docker ps | grep argus-prometheus
+docker ps | grep argus-grafana
+docker ps | grep argus-metric-test-node
+docker ps | grep argus-metric-test-gpu-node
+
+echo "[01] 基础服务已启动"
+
--- a/src/metric/tests/scripts/02_publish_artifact.sh
+++ b/src/metric/tests/scripts/02_publish_artifact.sh
@@ -0,0 +1,60 @@
+#!/bin/bash
+set -e
+
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+TEST_DIR="$(cd "$SCRIPT_DIR/.." && pwd)"
+PLUGIN_DIR="$(cd "$SCRIPT_DIR/../../client-plugins/all-in-one-full" && pwd)"
+
+# 加载 .env
+if [ -f "$TEST_DIR/.env" ]; then
+    source "$TEST_DIR/.env"
+fi
+
+# 检测容器挂载目录
+if docker ps --format '{{.Names}}' | grep -q '^argus-ftp$'; then
+    FTP_MOUNT=$(docker inspect argus-ftp --format '{{range .Mounts}}{{if eq .Destination "/private/argus/ftp"}}{{.Source}}{{end}}{{end}}')
+    OUTPUT_DIR="${FTP_MOUNT}/share"
+    echo "[02] 容器挂载: $OUTPUT_DIR"
+else
+    OUTPUT_DIR="${DATA_ROOT:-$TEST_DIR/data}/ftp/share"
+    echo "[02] 默认路径: $OUTPUT_DIR"
+fi
+
+OWNER="${ARGUS_BUILD_UID:-2133}:${ARGUS_BUILD_GID:-2015}"
+
+cd "$PLUGIN_DIR"
+
+echo "[02] 递增版本号..."
+bash scripts/version-manager.sh bump minor
+
+VERSION_FILE="config/VERSION"
+if [ ! -f "$VERSION_FILE" ]; then
+    echo "[02] 错误: 未找到 $VERSION_FILE"
+    exit 1
+fi
+
+VERSION=$(cat "$VERSION_FILE" | tr -d '[:space:]')
+echo "[02] 新版本: $VERSION"
+
+echo "[02] 构建安装包..."
+bash scripts/package_artifact.sh --force
+
+echo "[02] 发布到 FTP: $OUTPUT_DIR"
+sudo bash scripts/publish_artifact.sh "$VERSION" --output-dir "$OUTPUT_DIR" --owner "$OWNER"
+
+echo "[02] 设置文件权限..."
+# 设置所有者
+sudo chown -R "$OWNER" "$OUTPUT_DIR"
+# 设置目录权限为 755 (rwxr-xr-x)
+sudo find "$OUTPUT_DIR" -type d -exec chmod 755 {} \;
+# 设置文件权限为 644 (rw-r--r--)
+sudo find "$OUTPUT_DIR" -type f -exec chmod 644 {} \;
+# 特别处理 .sh 文件，给予执行权限 755
+sudo find "$OUTPUT_DIR" -type f -name "*.sh" -exec chmod 755 {} \;
+echo "[02] 权限设置完成 (UID:GID=$OWNER, dirs=755, files=644, scripts=755)"
+
+echo "[02] 发布完成，验证文件..."
+ls -lh "$OUTPUT_DIR"
+
+echo "[02] 完成"
+
--- a/src/metric/tests/scripts/03_test_node_install.sh
+++ b/src/metric/tests/scripts/03_test_node_install.sh
@@ -0,0 +1,33 @@
+#!/bin/bash
+set -e
+
+FTP_SERVER="${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-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 ${FTP_PORT}
+
+echo '[03] 安装完成'
+"
+
+echo "[03] 完成"
--- a/src/metric/tests/scripts/04_test_gpu_node_install.sh
+++ b/src/metric/tests/scripts/04_test_gpu_node_install.sh
@@ -0,0 +1,33 @@
+#!/bin/bash
+set -e
+
+FTP_SERVER="${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 ${FTP_PORT}
+
+echo '[03] 安装完成'
+"
+
+echo "[03] 完成"
--- a/src/metric/tests/scripts/05_verify_install.sh
+++ b/src/metric/tests/scripts/05_verify_install.sh
@@ -0,0 +1,96 @@
+#!/bin/bash
+set -e
+
+echo "[04] 验证安装结果 - 检查监控端口..."
+echo "=========================================="
+
+# 检查容器是否运行
+if ! docker ps --format '{{.Names}}' | grep -q '^argus-metric-test-node$'; then
+    echo "错误: 容器 argus-metric-test-node 未运行"
+    exit 1
+fi
+
+ERRORS=0
+
+# ==================== 检查监听端口 ====================
+echo ""
+echo "[1] 检查监听端口..."
+echo "----------------------------------------"
+CHECK_RESULT=$(docker exec argus-metric-test-node bash -c '
+if command -v netstat >/dev/null 2>&1; then
+    echo "使用 netstat 检查端口:"
+    if netstat -tlnp 2>/dev/null | grep -E ":(9100|9400|2020)"; then
+        echo "✓ 找到监控端口"
+        exit 0
+    else
+        echo "✗ 未找到监控端口 (9100/9400/2020)"
+        exit 1
+    fi
+elif command -v ss >/dev/null 2>&1; then
+    echo "使用 ss 检查端口:"
+    if ss -tlnp 2>/dev/null | grep -E ":(9100|9400|2020)"; then
+        echo "✓ 找到监控端口"
+        exit 0
+    else
+        echo "✗ 未找到监控端口 (9100/9400/2020)"
+        exit 1
+    fi
+elif command -v lsof >/dev/null 2>&1; then
+    echo "使用 lsof 检查端口:"
+    if lsof -i :9100 -i :9400 -i :2020 2>/dev/null | grep LISTEN; then
+        echo "✓ 找到监控端口"
+        exit 0
+    else
+        echo "✗ 未找到监控端口 (9100/9400/2020)"
+        exit 1
+    fi
+else
+    echo "? 没有可用的端口检查工具 (netstat/ss/lsof)，跳过此检查"
+    exit 0
+fi
+')
+echo "$CHECK_RESULT"
+# 只有在明确失败时才计入错误（exit 1），没有工具（exit 0）不算错误
+if echo "$CHECK_RESULT" | grep -q "✗ 未找到监控端口"; then
+    ERRORS=$((ERRORS + 1))
+fi
+
+# ==================== 测试端口连通性 ====================
+echo ""
+echo "[2] 测试端口连通性..."
+echo "----------------------------------------"
+docker exec argus-metric-test-node bash -c '
+if command -v curl >/dev/null 2>&1; then
+    FAILED=0
+    for port in 9100 9400 2020; do
+        echo -n "端口 $port: "
+        if curl -s --connect-timeout 2 "http://localhost:$port/metrics" > /dev/null 2>&1; then
+            echo "✓ 可访问 (/metrics)"
+        elif curl -s --connect-timeout 2 "http://localhost:$port/" > /dev/null 2>&1; then
+            echo "✓ 可访问 (根路径)"
+        else
+            echo "✗ 不可访问"
+            FAILED=$((FAILED + 1))
+        fi
+    done
+    exit $FAILED
+else
+    echo "? curl 不可用，跳过连通性测试"
+    exit 0
+fi
+' || ERRORS=$((ERRORS + 1))
+
+echo ""
+echo "=========================================="
+if [ $ERRORS -eq 0 ]; then
+    echo "✓ [04] 验证完成 - 所有端口检查通过"
+else
+    echo "✗ [04] 验证失败 - 发现 $ERRORS 个问题"
+    echo ""
+    echo "调试建议："
+    echo "  1. 进入容器检查: docker exec -it argus-metric-test-node bash"
+    echo "  2. 查看进程: docker exec argus-metric-test-node ps aux"
+    echo "  3. 查看日志: docker exec argus-metric-test-node cat /tmp/argus_install.log"
+    exit 1
+fi
+echo "=========================================="
--- a/src/metric/tests/scripts/06_cleanup.sh
+++ b/src/metric/tests/scripts/06_cleanup.sh
@@ -0,0 +1,11 @@
+#!/bin/bash
+set -e
+
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+
+echo "[05] 清理环境..."
+
+bash "$SCRIPT_DIR/common/stop-all.sh" || true
+
+echo "[05] 清理完成"
+
--- a/src/metric/tests/scripts/common/check-paths.sh
+++ b/src/metric/tests/scripts/common/check-paths.sh
@@ -6,7 +6,8 @@
 set -e

 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-cd "$SCRIPT_DIR"
+TEST_DIR="$(cd "$SCRIPT_DIR/../.." && pwd)"
+cd "$TEST_DIR"

 echo "=========================================="
 echo "  路径检查脚本"
@@ -18,15 +19,15 @@ echo ""

 # 检查配置文件
 echo "检查配置文件..."
-if [ -f "$SCRIPT_DIR/docker-compose.yml" ]; then
+if [ -f "$TEST_DIR/docker-compose.yml" ]; then
    echo "  ✓ docker-compose.yml 存在"
 else
    echo "  ✗ docker-compose.yml 不存在"
 fi

-if [ -f "$SCRIPT_DIR/.env" ]; then
+if [ -f "$TEST_DIR/.env" ]; then
    echo "  ✓ .env 存在"
-elif [ -f "$SCRIPT_DIR/env.example" ]; then
+elif [ -f "$TEST_DIR/env.example" ]; then
    echo "  ⚠ .env 不存在，但 env.example 存在"
 else
    echo "  ✗ .env 和 env.example 都不存在"
--- a/src/metric/tests/scripts/common/init-directories.sh
+++ b/src/metric/tests/scripts/common/init-directories.sh
@@ -0,0 +1,61 @@
+#!/bin/bash
+
+# 初始化目录脚本
+
+set -e
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+TEST_DIR="$(cd "$SCRIPT_DIR/../.." && pwd)"
+cd "$TEST_DIR"
+
+# 加载 .env 文件（如果存在）
+if [ -f .env ]; then
+    echo "加载 .env 配置文件..."
+    source .env
+fi
+
+# 默认配置
+ARGUS_BUILD_UID=${ARGUS_BUILD_UID:-2133}
+ARGUS_BUILD_GID=${ARGUS_BUILD_GID:-2015}
+DATA_ROOT=${DATA_ROOT:-/private}
+
+echo "开始初始化目录结构..."
+echo "数据根目录: ${DATA_ROOT}"
+echo "统一 UID: ${ARGUS_BUILD_UID}"
+echo "统一 GID: ${ARGUS_BUILD_GID}"
+
+# 创建基础目录结构
+echo "创建基础目录结构..."
+sudo mkdir -p ${DATA_ROOT}/argus/metric
+sudo mkdir -p ${DATA_ROOT}/argus/etc
+sudo mkdir -p ${DATA_ROOT}/argus/agent
+
+# 创建 FTP 目录
+echo "创建 FTP 目录..."
+sudo mkdir -p ${DATA_ROOT}/argus/metric/ftp/share
+
+# 创建 Prometheus 目录
+echo "创建 Prometheus 目录..."
+sudo mkdir -p ${DATA_ROOT}/argus/metric/prometheus/{data,rules,targets}
+
+# 创建 Grafana 目录
+echo "创建 Grafana 目录..."
+sudo mkdir -p ${DATA_ROOT}/argus/metric/grafana/{data,logs,plugins,provisioning/datasources,provisioning/dashboards,data/sessions,data/dashboards,config}
+
+# 统一设置所有目录权限
+echo "设置目录权限..."
+sudo chown -R ${ARGUS_BUILD_UID}:${ARGUS_BUILD_GID} ${DATA_ROOT}/argus/metric
+sudo chmod -R 755 ${DATA_ROOT}/argus/metric
+
+echo "目录初始化完成！"
+echo ""
+echo "目录结构："
+echo "  ${DATA_ROOT}/"
+echo "  ├── argus/                 (UID:${ARGUS_BUILD_UID}, GID:${ARGUS_BUILD_GID})"
+echo "  │   ├── metric/"
+echo "  │   │   ├── ftp/"
+echo "  │   │   ├── prometheus/"
+echo "  │   │   └── grafana/"
+echo ""
+echo "您现在可以运行 'docker-compose up -d' 来启动所有服务"
+
--- a/src/metric/tests/scripts/common/init-environment.sh
+++ b/src/metric/tests/scripts/common/init-environment.sh
--- a/src/metric/tests/scripts/common/manage-images.sh
+++ b/src/metric/tests/scripts/common/manage-images.sh
@@ -6,7 +6,8 @@
 set -e

 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-cd "$SCRIPT_DIR"
+TEST_DIR="$(cd "$SCRIPT_DIR/../.." && pwd)"
+cd "$TEST_DIR"

 # 检测 docker-compose 命令
 if command -v docker-compose &> /dev/null; then
@@ -19,7 +20,7 @@ else
 fi

 # 镜像缓存目录
-IMAGE_CACHE_DIR="./images-cache"
+IMAGE_CACHE_DIR="$TEST_DIR/images-cache"
 mkdir -p "$IMAGE_CACHE_DIR"

 # 定义镜像列表
--- a/src/metric/tests/scripts/common/start-all.sh
+++ b/src/metric/tests/scripts/common/start-all.sh
@@ -6,13 +6,20 @@
 set -e

 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-cd "$SCRIPT_DIR"
+TEST_DIR="$(cd "$SCRIPT_DIR/../.." && pwd)"
+cd "$TEST_DIR"
+
+# 解析参数
+FORCE_REBUILD=false
+if [[ "$1" == "--rebuild" ]]; then
+    FORCE_REBUILD=true
+fi

 echo "=========================================="
 echo "  Argus Metrics 一键启动脚本"
 echo "=========================================="
 echo ""
-echo "当前工作目录: $SCRIPT_DIR"
+echo "当前工作目录: $TEST_DIR"
 echo ""

 # 检查 Docker 和 Docker Compose
@@ -21,19 +28,13 @@ if ! command -v docker &> /dev/null; then
    exit 1
 fi

-# 检测 docker-compose 命令（兼容新旧版本）
-COMPOSE_FILE="$SCRIPT_DIR/docker-compose.yml"
-if command -v docker-compose &> /dev/null; then
-    DOCKER_COMPOSE="docker-compose -f $COMPOSE_FILE"
-    echo "使用: docker-compose"
-elif docker compose version &> /dev/null 2>&1; then
-    DOCKER_COMPOSE="docker compose -f $COMPOSE_FILE"
-    echo "使用: docker compose"
-else
-    echo "错误: 未找到 docker-compose 或 docker compose 命令"
+# 检查 docker compose 命令
+if ! docker compose version &> /dev/null 2>&1; then
+    echo "错误: 未找到 docker compose 命令，请确保 Docker Compose V2 已安装"
    exit 1
 fi
-echo "Compose 文件: $COMPOSE_FILE"
+echo "使用: docker compose"
+echo "Compose 文件: $TEST_DIR/docker-compose.yml"
 echo ""

 # 检查必要的构建目录
@@ -42,6 +43,8 @@ BUILD_DIRS=(
    "../ftp/build"
    "../prometheus/build"
    "../grafana/build"
+    "client-test-node/build"
+    "client-test-gpu-node/build"
 )

 for dir in "${BUILD_DIRS[@]}"; do
@@ -65,6 +68,18 @@ fi
 # 加载环境变量
 source .env

+# 检查并创建 Docker 网络
+echo "检查 Docker 网络..."
+NETWORK_NAME="argus-debug-net"
+if docker network inspect "$NETWORK_NAME" >/dev/null 2>&1; then
+    echo "网络 $NETWORK_NAME 已存在"
+else
+    echo "创建网络 $NETWORK_NAME..."
+    docker network create --driver bridge --subnet 172.30.0.0/16 "$NETWORK_NAME"
+    echo "网络创建成功"
+fi
+echo ""
+
 echo "1. 初始化目录结构..."
 bash "$SCRIPT_DIR/init-directories.sh"

@@ -72,8 +87,8 @@ echo ""
 echo "2. 准备 Docker 镜像..."

 # 检查镜像是否存在
-IMAGE_CACHE_DIR="./images-cache"
-IMAGES=("argus-metric-ftp:latest" "argus-metric-prometheus:latest" "argus-metric-grafana:latest")
+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" "argus-metric-test-gpu-node:latest")
 all_images_exist=true

 for image in "${IMAGES[@]}"; do
@@ -83,7 +98,12 @@ for image in "${IMAGES[@]}"; do
    fi
 done

-if $all_images_exist; then
+if $FORCE_REBUILD; then
+    echo "强制重新构建镜像（--rebuild 模式）..."
+    cd "$TEST_DIR"
+    docker compose build --no-cache
+    echo "镜像重新构建完成"
+elif $all_images_exist; then
    echo "所有镜像已存在，跳过构建"
 else
    echo "检测到缺失镜像，尝试从缓存加载..."
@@ -104,6 +124,12 @@ else
                    "argus-metric-grafana:latest")
                        cache_file="${IMAGE_CACHE_DIR}/argus-grafana.tar"
                        ;;
+                    "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
@@ -128,8 +154,8 @@ else
        echo ""
        echo "部分镜像缺失，开始构建..."
        echo "工作目录: $(pwd)"
-        cd "$SCRIPT_DIR"
-        $DOCKER_COMPOSE build
+        cd "$TEST_DIR"
+        docker compose build --no-cache
        
        # 询问是否保存镜像
        echo ""
@@ -149,6 +175,12 @@ else
                    "argus-metric-grafana:latest")
                        docker save -o "${IMAGE_CACHE_DIR}/argus-grafana.tar" "$image" && echo "  已保存: argus-grafana.tar"
                        ;;
+                    "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/"
@@ -160,40 +192,12 @@ else
 fi

 echo ""
-echo "3. 启动服务..."
-cd "$SCRIPT_DIR"
-$DOCKER_COMPOSE up -d
+echo "3. 启动基础服务..."
+cd "$TEST_DIR"
+# 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 "$SCRIPT_DIR"
-$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: $DOCKER_COMPOSE down -v"
-echo ""
-
--- a/src/metric/tests/scripts/common/stop-all.sh
+++ b/src/metric/tests/scripts/common/stop-all.sh
@@ -0,0 +1,50 @@
+    #!/bin/bash
+
+    # 停止所有服务脚本
+
+    set -e
+
+    SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+    TEST_DIR="$(cd "$SCRIPT_DIR/../.." && pwd)"
+    cd "$TEST_DIR"
+
+    # 检查 docker compose 命令
+    if ! docker compose version &> /dev/null 2>&1; then
+        echo "错误: 未找到 docker compose 命令，请确保 Docker Compose V2 已安装"
+        exit 1
+    fi
+
+    echo "=========================================="
+    echo "  停止 Argus Metrics 服务"
+    echo "=========================================="
+    echo ""
+    echo "使用: docker compose"
+    echo "Compose 文件: $TEST_DIR/docker-compose.yml"
+    echo ""
+
+    # 检查是否有运行的容器
+    if [ "$(docker compose ps -q)" ]; then
+        echo "停止所有服务..."
+        docker compose stop
+        
+        echo ""
+        read -p "是否要删除容器? (y/N): " -n 1 -r
+        echo
+        if [[ $REPLY =~ ^[Yy]$ ]]; then
+            docker compose down
+            echo "容器已删除"
+            
+            read -p "是否要删除数据卷? (y/N): " -n 1 -r
+            echo
+            if [[ $REPLY =~ ^[Yy]$ ]]; then
+                docker compose down -v
+                echo "数据卷已删除"
+            fi
+        fi
+    else
+        echo "没有运行的服务"
+    fi
+
+    echo ""
+    echo "完成！"
+
--- a/src/metric/tests/scripts/load-images.sh
+++ b/src/metric/tests/scripts/load-images.sh
@@ -0,0 +1,85 @@
+#!/bin/bash
+
+# 镜像加载脚本
+# 用于从 tar 文件加载 Docker 镜像
+
+set -e
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+TEST_DIR="$(cd "$SCRIPT_DIR/.." && pwd)"
+INPUT_DIR="${1:-$TEST_DIR/images-cache}"
+
+echo "=========================================="
+echo "  Docker 镜像加载脚本"
+echo "=========================================="
+echo ""
+echo "输入目录: $INPUT_DIR"
+echo ""
+
+# 检查输入目录是否存在
+if [ ! -d "$INPUT_DIR" ]; then
+    echo "错误: 目录不存在: $INPUT_DIR"
+    exit 1
+fi
+
+# 查找所有tar文件并加载
+total=0
+success=0
+failed=0
+
+# 查找目录下所有.tar文件
+tar_files=($(find "$INPUT_DIR" -name "*.tar" -type f 2>/dev/null | sort))
+
+if [ ${#tar_files[@]} -eq 0 ]; then
+    echo "错误: 在目录 $INPUT_DIR 中未找到任何 .tar 文件"
+    exit 1
+fi
+
+echo "找到 ${#tar_files[@]} 个镜像文件："
+for tar_file in "${tar_files[@]}"; do
+    echo "  - $(basename "$tar_file")"
+done
+echo ""
+
+for tar_file in "${tar_files[@]}"; do
+    total=$((total + 1))
+    tar_filename=$(basename "$tar_file")
+    
+    echo "[$total] 处理: $tar_filename"
+    
+    # 强制加载，不检查镜像是否已存在
+    echo "    加载镜像..."
+    if docker load -i "$tar_file"; then
+        echo "    加载成功: $tar_filename"
+        success=$((success + 1))
+    else
+        echo "    加载失败: $tar_filename"
+        failed=$((failed + 1))
+    fi
+    echo ""
+done
+
+echo "=========================================="
+echo "  加载完成"
+echo "=========================================="
+echo ""
+echo "统计："
+echo "  总计: $total"
+echo "  成功: $success"
+echo "  失败: $failed"
+echo ""
+
+# 显示当前所有镜像
+echo "当前所有镜像："
+docker images
+echo ""
+
+if [ $failed -gt 0 ]; then
+    echo "部分镜像加载失败，请检查！"
+    exit 1
+fi
+
+if [ $success -gt 0 ]; then
+    echo "镜像加载成功！"
+fi
+
--- a/src/metric/tests/scripts/save-images.sh
+++ b/src/metric/tests/scripts/save-images.sh
@@ -0,0 +1,94 @@
+#!/bin/bash
+
+# 镜像保存脚本
+# 用于保存 Docker 镜像到 tar 文件，便于离线部署
+
+set -e
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+TEST_DIR="$(cd "$SCRIPT_DIR/.." && pwd)"
+OUTPUT_DIR="${1:-$TEST_DIR/images-cache}"
+
+echo "=========================================="
+echo "  Docker 镜像保存脚本"
+echo "=========================================="
+echo ""
+echo "输出目录: $OUTPUT_DIR"
+echo ""
+
+# 创建输出目录
+mkdir -p "$OUTPUT_DIR"
+
+# 定义镜像名称（与 docker-compose.yml 保持一致）
+declare -A IMAGES=(
+    ["argus-metric-ftp:latest"]="argus-ftp.tar"
+    ["argus-metric-prometheus:latest"]="argus-prometheus.tar"
+    ["argus-metric-grafana:latest"]="argus-grafana.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
+total=0
+success=0
+failed=0
+
+for image in "${!IMAGES[@]}"; do
+    total=$((total + 1))
+    output_file="${OUTPUT_DIR}/${IMAGES[$image]}"
+    
+    echo "[$total] 检查镜像: $image"
+    
+    if docker images --format "{{.Repository}}:{{.Tag}}" | grep -q "^${image}$"; then
+        echo "    ✓ 镜像存在，开始保存..."
+        if docker save -o "$output_file" "$image"; then
+            file_size=$(ls -lh "$output_file" | awk '{print $5}')
+            echo "    ✓ 保存成功: ${IMAGES[$image]} ($file_size)"
+            success=$((success + 1))
+        else
+            echo "    ✗ 保存失败: $image"
+            failed=$((failed + 1))
+        fi
+    else
+        echo "    ✗ 镜像不存在，请先构建镜像"
+        failed=$((failed + 1))
+    fi
+    echo ""
+done
+
+echo "=========================================="
+echo "  保存完成"
+echo "=========================================="
+echo ""
+echo "统计："
+echo "  总计: $total"
+echo "  成功: $success"
+echo "  失败: $failed"
+echo ""
+echo "输出目录: $OUTPUT_DIR"
+echo ""
+
+if [ $success -gt 0 ]; then
+    echo "已保存的文件："
+    ls -lh "$OUTPUT_DIR"/*.tar 2>/dev/null || true
+    echo ""
+    echo "文件列表："
+    for image in "${!IMAGES[@]}"; do
+        output_file="${OUTPUT_DIR}/${IMAGES[$image]}"
+        if [ -f "$output_file" ]; then
+            file_size=$(ls -lh "$output_file" | awk '{print $5}')
+            echo "  - ${IMAGES[$image]} ($file_size)"
+        fi
+    done
+fi
+
+echo ""
+echo "使用说明："
+echo "1. 将 images-cache 目录复制到目标服务器的 ~/argus/src/metric/tests/ 下"
+echo "2. 在目标服务器运行: bash scripts/common/start-all.sh"
+echo ""
+
+if [ $failed -gt 0 ]; then
+    exit 1
+fi
+
--- a/src/metric/tests/stop-all.sh
+++ b/src/metric/tests/stop-all.sh
@@ -1,51 +0,0 @@
-#!/bin/bash
-
-# 停止所有服务脚本
-
-set -e
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-cd "$SCRIPT_DIR"
-
-# 检测 docker-compose 命令（兼容新旧版本）
-COMPOSE_FILE="$SCRIPT_DIR/docker-compose.yml"
-if command -v docker-compose &> /dev/null; then
-    DOCKER_COMPOSE="docker-compose -f $COMPOSE_FILE"
-elif docker compose version &> /dev/null 2>&1; then
-    DOCKER_COMPOSE="docker compose -f $COMPOSE_FILE"
-else
-    echo "错误: 未找到 docker-compose 或 docker compose 命令"
-    exit 1
-fi
-
-echo "=========================================="
-echo "  停止 Argus Metrics 服务"
-echo "=========================================="
-echo ""
-
-# 检查是否有运行的容器
-if [ "$($DOCKER_COMPOSE ps -q)" ]; then
-    echo "停止所有服务..."
-    $DOCKER_COMPOSE stop
-    
-    echo ""
-    read -p "是否要删除容器? (y/N): " -n 1 -r
-    echo
-    if [[ $REPLY =~ ^[Yy]$ ]]; then
-        $DOCKER_COMPOSE down
-        echo "容器已删除"
-        
-        read -p "是否要删除数据卷? (y/N): " -n 1 -r
-        echo
-        if [[ $REPLY =~ ^[Yy]$ ]]; then
-            $DOCKER_COMPOSE down -v
-            echo "数据卷已删除"
-        fi
-    fi
-else
-    echo "没有运行的服务"
-fi
-
-echo ""
-echo "完成！"
-