Azure VM中启动Topshelf Windows服务失败求助

问题背景 #

尝试在Azure VM中通过Azure DevOps流水线安装并启动Topshelf Windows服务，但启动步骤执行失败。

初始流水线任务配置 #

使用第三方Topshelf安装和启动任务：

- task: jabbera.windows-service-release-tasks.install-topshelf-service-task.InstallTopshelfService@8
  displayName: 'Install a Topshelf Service'
  inputs:
    TopshelfExePaths: '${{ parameters.svcDir}}\${{ parameters.svcAssemblyName }}.exe'
    specialUser: localsystem
    ServiceName: '${{ parameters.svcName}}'
    StartupType: autostart
    KillMmcTaskManager: true
    deploymentGroup: true


 - task: jabbera.windows-service-release-tasks.start-windows-service-task.StartWindowsService@8
   displayName: 'Start Windows Service'
   inputs:
       ServiceNames: '${{ parameters.svcName}}'
       deploymentGroup: true

第一次启动错误信息 #

##[error]Unable to find type [VstsTaskSdk.TerminationException].
2024-05-08T09:54:34.2331237Z ##[error]Exit code 1 returned from process: file name 'C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe', arguments '-NoLogo -Sta -NoProfile -NonInteractive -ExecutionPolicy Unrestricted -Command ". ([scriptblock]::Create('if ([Console]::InputEncoding -is [Text.UTF8Encoding] -and [Console]::InputEncoding.GetPreamble().Length -ne 0) { [Console]::InputEncoding = New-Object Text.UTF8Encoding $false } if (!$PSHOME) { $null = Get-Item -LiteralPath ''variable:PSHOME'' } else { Import-Module -Name ([System.IO.Path]::Combine($PSHOME, ''Modules\Microsoft.PowerShell.Management\Microsoft.PowerShell.Management.psd1'')) ; Import-Module -Name ([System.IO.Path]::Combine($PSHOME, ''Modules\Microsoft.PowerShell.Utility\Microsoft.PowerShell.Utility.psd1'')) }')) 2>&1 | ForEach-Object { Write-Verbose $_.Exception.Message -Verbose } ; Import-Module -Name 'C:\azagent\A1_work_tasks\StartWindowsService_bf65dd9d-18cb-40d9-9254-f3ac527428ef\8.0.0\ps_modules\VstsTaskSdk\VstsTaskSdk.psd1' -ArgumentList @{ NonInteractive = $true } -ErrorAction Stop ; $VerbosePreference = 'SilentlyContinue' ; $DebugPreference = 'SilentlyContinue' ; Invoke-VstsTaskScript -ScriptBlock ([scriptblock]::Create('. ''C:\azagent\A1_work_tasks\StartWindowsService_bf65dd9d-18cb-40d9-9254-f3ac527428ef\8.0.0\StartWindowsService.ps1'''))".

替换为PowerShell任务后的配置 #

改用原生PowerShell任务尝试启动服务：

- powershell: |
    $serviceName = '${{ parameters.serviceName }}'
    $serviceStatus = Get-Service -Name $serviceName
    if ($serviceStatus.Status -eq 'Running') {
        Write-Host "Service '$serviceName' is already running."
    } elseif ($serviceStatus.Status -eq 'Stopped') {
        Start-Service -Name $serviceName
        Write-Host "Service '$serviceName' started successfully."
    } else {
        Write-Host "Service '$serviceName' is in an unexpected state: $($serviceStatus.Status)"
    }
  condition: ${{ parameters.startService }}
  displayName: 'Start Windows Service'

PowerShell任务错误信息 #

Start-Service : Service 'MyService' cannot be started due to the following error: Cannot
2024-05-08T14:25:47.7248406Z start service MyService on computer '.'.
2024-05-08T14:25:47.7250339Z At C:\azagent\A1_work_temp\5d2155f7-3687-4ce8-ab8e-fbf42272de9d.ps1:9 char:5
2024-05-08T14:25:47.7252117Z + Start-Service -Name $serviceName
2024-05-08T14:25:47.7253589Z + ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2024-05-08T14:25:47.7255675Z + CategoryInfo : OpenError: (System.ServiceProcess.ServiceController:ServiceController) [Start-Service],
2024-05-08T14:25:47.7257458Z ServiceCommandException
2024-05-08T14:25:47.7259190Z + FullyQualifiedErrorId : CouldNotStartService,Microsoft.PowerShell.Commands.StartServiceCommand
2024-05-08T14:25:47.7260730Z
2024-05-08T14:25:47.7300363Z ##[error]PowerShell exited with code '1'.
2024-05-08T14:25:47.7904871Z ##[section]Finishing: Start Windows Service: MyService

排查与解决方案 #

1. 修复第三方任务依赖问题 #

第一个错误Unable to find type [VstsTaskSdk.TerminationException]指向第三方任务的PowerShell模块加载失败：

• 检查Azure Agent目录下的VstsTaskSdk模块完整性，路径为C:\azagent\A1_work_tasks\StartWindowsService_bf65dd9d-18cb-40d9-9254-f3ac527428ef\8.0.0\ps_modules\VstsTaskSdk\
• 降级第三方任务到稳定版本，v8可能存在兼容性问题
• 确认Azure Agent运行的PowerShell v5.1与任务要求兼容

2. 定位服务自身启动失败原因 #

PowerShell任务的错误说明服务本身无法启动，需从以下方向排查：

• 查看系统事件日志：登录Azure VM，打开事件查看器，在Windows日志 -> 系统和应用程序中查找服务启动失败的具体错误（如依赖缺失、权限不足、配置错误）
• 手动启动验证：远程登录VM后手动启动服务，观察弹出的错误提示，这是最直接的排查方式
• 检查服务权限：当前服务使用localsystem账户，确认该账户拥有服务运行所需的权限（文件读写、数据库连接、网络访问等）
• 验证程序完整性：检查${{ parameters.svcDir }}下的exe及依赖DLL是否完整，直接运行exeName run命令测试程序能否正常执行
• 添加启动日志：修改Topshelf服务代码，增加启动环节的日志输出，或配置Topshelf自带的日志功能，定位启动失败的具体环节

3. 优化PowerShell脚本获取详细错误 #

修改脚本以捕获更多错误信息：

- powershell: |
    $serviceName = '${{ parameters.serviceName }}'
    try {
        $serviceStatus = Get-Service -Name $serviceName -ErrorAction Stop
        Write-Host "Current service state: $($serviceStatus.Status)"
        
        if ($serviceStatus.Status -eq 'Running') {
            Write-Host "Service '$serviceName' is already running."
        } elseif ($serviceStatus.Status -eq 'Stopped') {
            Start-Service -Name $serviceName -ErrorAction Stop
            Write-Host "Service '$serviceName' started successfully."
        } else {
            Write-Host "Service '$serviceName' is in an unexpected state: $($serviceStatus.Status)"
        }
    } catch {
        Write-Host "Error occurred: $_"
        # 获取服务启动失败的详细错误代码
        $service = Get-WmiObject -Class Win32_Service -Filter "Name='$serviceName'"
        if ($service) {
            Write-Host "Service exit code: $($service.ExitCode)"
            Write-Host "Service last error: $($service.LastErrorCode)"
        }
        throw $_
    }
  condition: ${{ parameters.startService }}
  displayName: 'Start Windows Service with error details'

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