启动AducIotAgent报错：无法初始化诊断工作流数据求解决

操作背景 #

在Ubuntu 20.04环境中部署AducIotAgent，执行了以下操作：

1. 克隆iot-hub-device-update仓库并安装依赖：

./scripts/install-deps.sh -a
sudo cmake --build out --target install

2. 克隆do-client仓库并安装相关deb包：

python3 build/build.py --project agent --package-for deb
cd /tmp/build-deliveryoptimization-agent/linux-debug/
sudo apt-get install ./deliveryoptimization-agent*.deb
cd <project root>
python3 build/build.py --project plugin-apt --package-for deb

cd /tmp/build-deliveryoptimization-sdk/linux-debug/
sudo apt-get install ./libdeliveryoptimization*.deb
cd /tmp/build-deliveryoptimization-agent/linux-debug/
sudo apt-get install ./deliveryoptimization-agent*.deb
cd /tmp/build-deliveryoptimization-plugin-apt/linux-debug/
sudo apt-get install ./deliveryoptimization-plugin-apt*.deb

3. 在IoT Hub创建设备后，将连接字符串配置到/etc/adu/du-config.json：

{
  "schemaVersion": "1.0",
  "aduShellTrustedUsers": [
     "adu",
     "do"
  ],
  "manufacturer": "fabrikam",
  "model": "vacuum",
  "agents": [
     {
     "name": "main",
     "runas": "adu",
     "connectionSource": {
        "connectionType": "string",
        "connectionData": "<created device on azure iothub connection String>"
     },
     "manufacturer": "fabrikam",
     "model": "vacuum"
     }
  ]}

启动报错 #

执行启动命令：

sudo AducIotAgent --enable-iothub-tracing --log-level 1

出现如下错误日志：

2023-09-07T07:57:11.8021Z 3717069[3717069] [I] Agent (linux; 1.0.2) starting. [main]
2023-09-07T07:57:11.8021Z 3717069[3717069] [I] Git Info: main:f9b180de [main]
2023-09-07T07:57:11.8021Z 3717069[3717069] [I] Supported Update Manifest version: min: 4, max: 5 [main]
2023-09-07T07:57:11.8022Z 3717069[3717069] [I] Health check passed. [HealthCheck]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] Initializing PnP components. [ADUC_PnP_Components_Create]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] ADUC agent started. Using IoT Hub Client SDK 1.7.0 [AzureDeviceUpdateCoreInterface_Create]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] Calling ADUC_RegisterPlatformLayer [ADUC_MethodCall_Register]
2023-09-07T07:57:11.8023Z 3717069[3717069] [E] Unable to initialize the diagnostic workflow data. [DiagnosticsInterface_Create]
2023-09-07T07:57:11.8023Z 3717069[3717069] [E] Failed to initialize PnP component 'diagnosticInformation'. [ADUC_PnP_Components_Create]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] ADUC agent stopping [AzureDeviceUpdateCoreInterface_Destroy]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] Calling ADUC_Unregister [ADUC_MethodCall_Unregister]
2023-09-07T07:57:11.8023Z 3717069[3717069] [E] ADUC_PnP_Components_Create failed [StartupAgent]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] Agent exited with code 1 [main]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] Agent is shutting down with signal 0. [ShutdownAgent]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] De-initializing command listener thread [UninitializeCommandListenerThread]
2023-09-07T07:57:11.8023Z 3717069[3717069] [I] ADUC agent stopping [AzureDeviceUpdateCoreInterface_Destroy]

排查思路 #

• 检查诊断目录权限：AducIotAgent需要读写诊断数据目录，默认路径为/var/lib/adu/diagnostics，确保adu用户拥有读写权限：

  sudo chown -R adu:adu /var/lib/adu/diagnostics
  sudo chmod -R 755 /var/lib/adu/diagnostics
  

• 验证诊断组件完整性：检查诊断相关组件是否存在，比如libaduc-diagnostics.so是否在/usr/lib/x86_64-linux-gnu/adu/目录下，若缺失，重新编译时确保开启诊断组件编译选项。
• 补充诊断配置项：在du-config.json根节点添加诊断相关配置：

  "diagnostics": {
    "enabled": true,
    "storagePath": "/var/lib/adu/diagnostics",
    "maxLogSizeBytes": 10485760
  }
  

• 查看系统日志细节：通过journalctl -u adu-agent.service（服务启动方式）或/var/log/syslog获取更多初始化失败的上下文信息。
• 重新编译并开启诊断选项：重新编译iot-hub-device-update仓库时显式开启诊断组件：

  cmake -S . -B out -DADUC_BUILD_DIAGNOSTICS=ON
  sudo cmake --build out --target install
  

• 确认adu用户权限：执行id adu查看用户组信息，确保其拥有访问IoT Hub资源、诊断目录的必要权限。

内容的提问来源于stack exchange，提问作者Gustavo Nóbrega