Disaster Recovery  «Prev  Next»

Lesson 6 PXE-Initiated Operating-System Deployment
Objective Explain how to boot a target server into Windows PE and deploy or restore a Windows Server 2025 image using UEFI PXE, a customized WinPE boot image, and a Configuration Manager task sequence

PXE-Initiated Operating-System Deployment on Windows Server 2025

Enterprise OS deployment on Windows Server 2025 begins with a network boot. A physical server or virtual machine selects UEFI PXE from the firmware boot menu, contacts the PXE infrastructure, downloads a customized Windows PE boot image, and then runs an automated task sequence that handles every step from disk partitioning through final validation. The same infrastructure and the same WinPE environment serve both fresh deployments and disaster-recovery restores — a key advantage of the modern model over its predecessor.

The Deployment Sequence

The Windows Server 2025 deployment sequence replaces the Windows 2000 RIS Client Installation Wizard workflow entirely:
Power on physical server or VM
→ Select UEFI network/PXE boot from firmware boot menu
→ Contact PXE infrastructure (DHCP + PXE service)
→ Download customized WinPE boot image
→ Start Configuration Manager task sequence
→ Partition disks, apply Windows Server 2025 image
→ Install drivers, apply answer file, configure server role
→ Install roles, applications, and scripts
→ Restart and validate

Windows Server 2025 deployment workflow showing ten steps from power-on through
       UEFI PXE boot, WinPE download, task sequence execution, disk preparation, image
       application, driver installation, role and application setup, and final validation
Figure 6-1: PXE-initiated OS deployment on Windows Server 2025. A target server boots into a customized WinPE environment and runs a task sequence that deploys or restores the Windows Server 2025 image.


The Ten-Step Deployment Process

Step 1: Power On the Target Server or VM

The target is a physical server undergoing bare-metal deployment or recovery, or a virtual machine being provisioned in Hyper-V, VMware, or a cloud environment. No OS is present on the target disk at this stage, or the existing OS is being replaced.

Step 2: Select UEFI Network Boot

At the firmware boot menu — accessed during POST, typically via F11 or F12 depending on the hardware vendor — the administrator selects UEFI network boot or the specific network adapter. Most modern server firmware supports UEFI PXE boot on integrated network adapters without additional configuration beyond enabling the option in UEFI settings.

For automated deployments where no administrator is present at the console, servers can be configured to attempt network boot first in the firmware boot order, so the deployment begins without manual interaction at the hardware level.

Step 3: Contact PXE Infrastructure

The server broadcasts a DHCP discover request. The DHCP server returns an IP address along with PXE boot server information — DHCP Option 66 (boot server hostname) and Option 67 (boot filename). The server contacts the PXE service and begins the boot file transfer.

In a Configuration Manager environment, a PXE-enabled distribution point serves the boot file. The native PXE responder built into Configuration Manager handles this without requiring a separate WDS installation. WDS can also serve as the PXE transport provider when a custom boot image is used — the WDS installation-media workflow that is blocked for Windows Server 2025 does not affect the PXE transport role.


Step 4: Download Custom WinPE Boot Image

The PXE service transfers the customized WinPE boot image to the target server over the network using TFTP or HTTP, depending on the PXE service configuration. The server loads WinPE into RAM and boots from it.

The WinPE image must be customized with drivers for the target server's storage controllers and network adapters before this step can succeed. A WinPE image that cannot see the server's disks or network interfaces will stall here. Lesson 5 covered the process for building and customizing a WinPE image with the ADK.



Step 5: Start Deployment Task Sequence

Once WinPE has loaded, the Configuration Manager task sequence environment starts. The task sequence is the automation engine that drives all subsequent steps. It was assigned to the device collection containing the target server (Lesson 4), and the task sequence engine retrieves the appropriate sequence for this device.

The task sequence type determines the outcome:
  • Deployment task sequence — installs Windows Server 2025 fresh on the target disk
  • Recovery task sequence — restores a bare-metal backup or applies a previously captured WIM image to return the server to a known-good state
  • Capture task sequence — runs on a reference server to capture a sysprepped WIM for later deployment

Step 6: Prepare Disks

The task sequence executes disk preparation using DiskPart commands or the Configuration Manager Format and Partition Disk step. For a UEFI system, the standard partition layout creates:
  • An EFI System Partition (ESP) — 100 to 500 MB, FAT32, required for UEFI boot
  • A Microsoft Reserved Partition (MSR) — 16 MB
  • The Windows OS partition — remainder of the disk, NTFS
  • An optional WinRE recovery partition — 500 MB to 1 GB, NTFS
For a disaster-recovery restore to replacement hardware, the partition layout must be compatible with the backup that will be restored. The task sequence handles this partitioning automatically when the backup platform supports it.

Step 7: Apply Windows Server 2025 Image

The task sequence applies the OS image using one of the following methods depending on whether this is a deployment or a recovery operation:
  • DISM Apply-Image — applies a WIM file directly to the prepared partition
  • Configuration Manager Apply Operating System Image step — wraps DISM with task-sequence variable support and answer-file injection
  • Backup vendor restore agent — restores a bare-metal backup directly to the partition for recovery scenarios
The WIM referenced can be the original install.wim from Windows Server 2025 media, a captured WIM from a reference server, or a serviced WIM with updates and drivers already integrated. A single base WIM is reused across all task sequences; the answer file and subsequent steps differentiate the resulting server role.

Step 8: Install Drivers and Settings

After the image is applied, the task sequence injects storage and network drivers using the Configuration Manager Apply Driver Package step or DISM. It also processes the unattended Setup answer file (unattend.xml) for the specialize and oobeSystem passes, applying machine-specific settings — computer name, domain join, time zone, network configuration, and security baseline settings.

Step 9: Install Roles, Applications, and Scripts

With the base OS and drivers in place, the task sequence installs Windows Server roles and features, required applications, and organization-specific configuration scripts. This is where deployments diverge by server role — the Domain Controller task sequence installs ADDS, DNS, and group policy here, while the IIS task sequence installs web server components, TLS certificates, and application pools.

Step 10: Restart and Validate

The server restarts into the fully configured Windows Server 2025 OS. Post-deployment validation confirms the deployment succeeded:
  • Server joined the correct Active Directory domain and organizational unit
  • Required services are running and healthy
  • Connectivity to dependent systems is confirmed
  • Monitoring agents are registered with the management platform
  • Deployment result is logged to Configuration Manager for compliance reporting

Disaster-Recovery Application

The same 10-step workflow applies directly to server restoration. In a recovery scenario the steps map as follows:
  • Step 1 — the failed server is powered on, or a replacement server is brought online
  • Steps 2–4 — the server boots into the WinPE recovery environment via PXE, USB, or mounted ISO (covered in Lesson 5)
  • Step 5 — a recovery task sequence starts, or an administrator runs a recovery script manually from the WinPE command prompt
  • Step 6 — disks are repartitioned to match the original layout required by the backup
  • Step 7 — the bare-metal backup is restored or a captured WIM is applied to the prepared partition
  • Steps 8–10 — drivers, settings, and server roles are validated; the server restarts and is confirmed healthy
The same WinPE image and the same PXE, USB, and ISO delivery infrastructure serve both deployment and recovery purposes. The legacy RIS model was deployment-only; the WinPE model handles both in the same boot environment.

Deployment Method Comparison

Windows 2000 RIS Windows Server 2025
RIS client computer Target physical server or virtual machine
F12 network-service startup UEFI PXE or network-boot firmware selection
RIS Startup disk Customized WinPE USB or ISO media
Client Installation Wizard Configuration Manager task-sequence environment
RIS installation image Windows Server 2025 WIM, generalized VHDX, or bare-metal backup
RIS server PXE-enabled Configuration Manager distribution point
Manual image selection by user Collection-targeted or policy-driven task sequence
Deployment only Deployment and disaster recovery from the same WinPE environment


The next lesson covers RIPrep concepts and their modern equivalents in Windows Server 2025 image capture and reference server preparation.

SEMrush Software 6 SEMrush Banner 6