Windows 10 deployment with Linux - Initial Setup

7 minute read

I’ve always been under the impression that you need Windows Deployment Services (WDS) in order to deploy Windows with the Microsoft Deployment Toolkit (MDT). But it turns out you need nothing more than a Samba share, TFTP server, and a few tweaks to DHCP to get a pretty solid set up.

This procedure involves generating a boot image with MDT and copying that image to a TFTP server. A client then PXE boots, loads that image into memory through memdisk and runs the task sequence to install Windows. You will of course require one Windows machine to configure the task sequences.

This guide only contains the barebones installation of Windows through a Deployment Share. You can add software and automate the process to provide a zero touch installation. More on that in a future post.


Below are the requirements:

To keep things simple, I will configure Samba, DHCP, and TFTP on the same computer. I will be using CentOS in this guide but these steps should be easily portable to other flavours of Linux.

This guide is based on a virtual environment using KVM. The CentOS computer will be configured with an IP address of All other computers (Windows 10, deployment machine) are virtual and configured to use DHCP.

The CentOS computer has been installed with the CentOS minimal ISO. The latest updates have been installed.


Linux computer with a Samba share

Install samba and policycoreutils-python. The latter package will give us the semanage command for SELinux.

~]# yum install -y samba policycoreutils-python

Create two user accounts; mdtread and mdtwrite. The mdtread account will used to connect to the deployment share during deployment. The mdtwrite account is what the Windows 10 computer will use to connect and make changes to the deployment share. These accounts are only needed for authentication so we won’t create a home directory (-M) or give them a shell (-s /sbin/nologin).

~]# useradd -M -s /sbin/nologin mdtread
~]# useradd -M -s /sbin/nologin mdtwrite
~]# passwd mdtread
~]# passwd mdtread
~]# smbpasswd -a mdtread
~]# smbpasswd -a mdtwrite

Create and configure the MDT Share for Samba. I create this as a hidden share so it does not appear in an explorer window. I remove everything in the default smb.conf except the [global] section.

~]# mkdir -p /data/mdt
~]# chown -R mdtwrite:mdtwrite /data/mdt
~]# vi /etc/samba/smb.conf
    path = /data/mdt
    valid users = mdtread mdtwrite
    write list = mdtwrite

Verify everything is ok with the Samba configuration file

~]# testparm

Configure SELinux to allow Samba access to the mdt directory. Verify this with the ls commmand.

~]# semanage fcontext -a -t samba_share_t "/data/mdt(/.*)?"
~]# restorecon /data/mdt
~]# ls -alZ /data/mdt
~]# systemctl start smb
~]# systemctl enable smb
~]# firewall-cmd --add-service=samba
~]# firewall-cmd --add-service=samba --permanent


If you must enable NetBIOS, start and enable nmbd. This will allow you to resolve to the hostname of your Linux computer through NetBIOS, rather than DNS.

~]# systemctl start nmb
~]# systemctl enable nmb

Verify Share Access

On your Windows 10 computer, verify you can access the share. Click on the Start menu and type \\\mdt$

Windows 10 computer with MDT

Install MDT and ADK

You will need to download and install the following:

  • Microsoft Deployment Toolkit
  • Windows 10 Assessment and Deployment Kit (I just install Deployment Tools and Windows PE)

A simple Google search will bring up the download links for this software.

Add Deployment Share to MDT

Once MDT and ADK are installed, open the Deployment Workbench.

Right click Deployment Shares and select New Deployment Share.

New Deployment Share

If you try to add the deployment share through its UNC path, you will receieve an error. This is because you haven’t been authenticated yet. What you can do to work around this is open the Credential Manager from the Control Panel and add new Windows credentials for your share (, mdtwrite, <password>).

Add using UNC path

Credential Manager

Follow the rest of the prompts. These can always be changed at a later stage. The Deployment Workbench will now copy some files to the MDT share.

Obtain Windows 10 ISO and Upload to Deployment Share

You will need a Windows 10 ISO to upload to the MDT share. Thankfully this can be downloaded quite easily from the Microsoft website, unlike previous versions of Windows.

If using a Mac or Linux computer, you can go here and select the edition and language to download.

Downloading from a Windows machine requires a few more steps. It’s nothing hard but we just need to trick the web server into thinking that we are using another operating system, otherwise we are presented wtith the Windows 10 installation media download. In Google Chrome, go to the Windows 10 download site. Open the Developer tools (CTRL + SHIFT + I). Click the three dots at the top right hand corner > More Tools > Network Conditions. Uncheck Select Automatically and select a non-Windows OS, such as Chrome - Mac. Hit the refresh button and you will be able to select the edition and language to download.

Network Conditions

Chrome Mac

I download the English version. I haven’t verified this but I believe the difference between English and English International is the English version is US English. The English International version is British English and contains additional keyboard layouts.

Once the download has completed, mount the ISO.

Mount OS

Back in the Deployment Share, right click Operating Systems and select Import Operating System.

Import OS

Make sure Full set of source files is checked and click next.

OS Type

Browse to the drive where the ISO was mounted and click next.

OS Source

Follow the rest of the prompts. The OS files will begin uploadeding to the Deployment Share. This may take a few minutes.

Now we want to create a task sequence. Right click Task Sequences and select New Task Sequence.

New Task Sequence

Give the Task Sequence an ID and Name and click next.

Task Sequence Name

Make sure Standard Client Task Sequence is selected and click next.

Task Sequence Template

Select the OS that you are deploying. Here I’m using Windows 10 Pro.

Task Sequence Select OS

Follow the rest of the prompts. I only have individual activation keys, so I don’t specify them during the install.

We are now ready to generate the ISO that we can boot for deployment. Right click on the Deployment Share and select Update Deployment Share. Since this is a new Deployment Share, just click next through the prompts. The ISO that needs to be uploaded to the TFTP Server will be generated. This may take a few minutes.

Update Deployment Share

TFTP server

Install TFTP Server.

~]# yum install -y tftp-server

Enable TFTP Server in its configuration file by changing the disable line from yes to no.

~]# vi /etc/xinetd.d/tftp
disable = no

Download syslinux. We need to use memdisk to boot the ISO. I just pull syslinux from the CentOS repos.

~]# yum install -y syslinux

Copy the necessary syslinux files to the tftpboot directory.

~]# cp /usr/share/syslinux/pxelinux.0 /var/lib/tftpboot/
~]# cp /usr/share/syslinux/memdisk /var/lib/tftpboot/
~]# cp /usr/share/syslinux/menu.c32 /var/lib/tftpboot/

Copy the LiteTouch ISO to the tftpboot directory.

~]# cp /data/mdt/Boot/LiteTouchPE_x64.iso /var/lib/tftpboot/

Create the boot menu.

~]# mkdir -p /var/lib/tftpboot/pxelinux.cfg
~]# vi /var/lib/tftpboot/pxelinux.cfg/default
default menu.c32
prompt 0
timeout 300

menu title ########## OS Deploy ##########

label 1
menu label ^1) Boot from local drive
localboot 0

label 2
menu label ^2) Windows 10
#menu default
kernel memdisk
append iso initrd=LiteTouchPE_x64.iso raw

Start the TFTP server and enable on startup.

~]# systemctl start tftp
~]# systemctl enable tftp

Configure firewall to allow TFTP through.

~]# firewall-cmd --add-service=tftp
~]# firewall-cmd --add-service=tftp --permanent

DHCP Server

Install the DHCP server

~]# yum install -y dhcp

Configure the DHCP server

~]# vi /etc/dhcp/dhcpd.conf
DHCPDARGS=eth0; #Listen on eth0
subnet netmask
        option routers        ;
        option subnet-mask    ;
        option domain-search            "lab.local";
        option domain-name-servers;
        option time-offset              -18000;     # Eastern Standard Time
        filename                        "pxelinux.0";
        next-server           ;
        range                 ;

Start the DHCP service and enable on startup

~]# systemctl start dhcpd
~]# systemctl enable dhcpd

Configure firewall to allow DHCP requests

~]# firewall-cmd --add-service=dhcp
~]# firewall-cmd --add-service=dhcp --permanent

Finishing Touches

When the Deployment Share was created, a few files and folders were created on it. We need to adjust some of the permissions so our mdtread user can access the share.

~]# chmod -R 755 /data/mdt


We are now ready to deploy Windows 10.

Network boot the computer and select Windows 10. PXE Windows 10

Select Run Deployment Wizard to install a new Operating System. Run Deployment Wizard

Enter the read only credentials of the Deployment Share. Deployment Credentials

Select the Windows 10 Deployment Task Sequence and follow the rest of the prompts. Follow the rest of the prompts. Windows 10 will now be installed.

Deployment Install


Samba - Creating Local Accounts

Samba Configuration File