Service startup type and crash recovery: Set-Service plus sc.exe failure/qfailure

Goal

Check and change a service's startup type, and configure what happens automatically the first, second, and later times it crashes, on a host reachable only over a PowerShell session.

Prerequisites

Administrator rights on the target; sc.exe is present on every supported Windows Server edition, including Server Core.

Steps

1. Inventory: Get-Service -Name Spooler | Select-Object Name,Status,StartType, or find everything not running that should be: Get-Service | Where-Object { $_.StartType -eq 'Automatic' -and $_.Status -ne 'Running' }. Get-Service exposes both Status and StartType directly.
2. Change startup type with the built-in cmdlet: Set-Service -Name Spooler -StartupType Automatic.
3. PowerShell's Set-Service does not expose crash-recovery actions, so use sc.exe failure: sc.exe failure Spooler reset= 86400 actions= restart/5000/restart/10000/run/0. The documented syntax is failure [ServiceName] [reset= ErrorFreePeriod] [reboot= BroadcastMessage] [command= CommandLine] [actions= FailureActionsAndDelayTime] — note the space after each key=, exactly as shown in the syntax itself; omitting it causes sc.exe to misparse the argument. reset= (seconds with no failures before the failure count clears) must be used together with actions= (a /-separated list of run, restart, or reboot, each followed by its delay in milliseconds).
4. Read back the configured behaviour non-interactively: sc.exe qfailure Spooler, which reports SERVICE_NAME, RESET_PERIOD, REBOOT_MESSAGE, COMMAND_LINE, and FAILURE_ACTIONS.
5. Check what a service depends on, or what depends on it, before changing its startup type: Get-Service -Name Spooler -DependentServices; Get-Service -Name Spooler -RequiredServices.
6. For a service account shared by identical service instances across a farm, use a group Managed Service Account instead of a hand-typed password: gMSAs let "the Windows operating system manage the password" and provide "a single identity solution for services running on a server farm or on systems behind Network Load Balancer."

Expected result

sc.exe qfailure echoes back the reset period and actions just set; recovery actions fire only when the service terminates unexpectedly, not on a clean Stop-Service. To test on a non-critical service, end its process (Stop-Process -Name spoolsv -Force) and watch the System event log for event 7031 and the restart.

Limits and test basis

"Not all services allow changes to their failure options" — some run as part of a shared service host process. To undo, save the output of sc.exe qfailure Spooler before the change and re-apply those values with sc.exe failure, and Set-Service -Name Spooler -StartupType <PreviousValue> to restore the prior startup type recorded in step 1. None of these steps requires a reboot; a stopped-and-restarted service is enough to test recovery actions.