Windows Server on a Dedicated Server: IIS Cutover Guide

Why this cutover plan matters
This tutorial walks you through moving a windows server dedicated server from first login to a working IIS deployment, with a clean cutover path along the way. It is written for real hosting migrations: a fresh dedicated machine, a public site, a second terminal for verification, and a rollback plan if anything breaks.
Use it when you need a production-ready baseline on Windows Server 2022 or Windows Server 2025, then want IIS, firewall rules, TLS-ready web settings, and reboot persistence checked before you hand the server over or switch DNS live. If you are still sizing hardware for a high-traffic deployment, the Hostperl dedicated server hosting page is a practical place to start.
For teams that also need operating-system tuning and broader server hardening around the same migration, Hostperl covers related production topics in Windows Server sizing and scaling and Windows Server checkout recovery. The steps below stay focused on IIS and server readiness.
What you need before you begin
- A Windows Server 2022 or Windows Server 2025 dedicated server.
- Administrator access through RDP or console.
- A domain name already pointed to the server, or ready to be switched.
- A maintenance window long enough to test before you change DNS.
- A second terminal on your local computer for verification.
This guide assumes a fresh server. If you are moving from another host, keep the old server running until the new IIS site passes every check.
Connect and confirm the platform
On your local computer, connect first. The IP below is only an example; replace 203.0.113.10 with the public IP assigned to your server.
ssh root@203.0.113.10If your Hostperl deployment uses a default non-root administrator instead, connect with that account using the same example IP. Keep the original session open until you verify the new login.
On Windows Server, you will use PowerShell in the console or over RDP. Start by confirming the edition and build.
Get-ComputerInfo | Select-Object WindowsProductName, WindowsVersion, OsHardwareAbstractionLayerYou should see Windows Server 2022 or 2025. If you are connected over SSH to a Windows build with OpenSSH enabled, use PowerShell there as well. If the output does not match a supported Windows Server release, stop and confirm the image before you continue.
Create a non-administrator account for day-to-day work
Run these commands in PowerShell on the server as Administrator. This gives you a safer daily account and keeps the built-in administrator for recovery only.
New-LocalUser -Name "deploy" -Password (Read-Host -AsSecureString "Set a strong password for deploy") -FullName "Deploy User" -Description "Primary administration account"
Add-LocalGroupMember -Group "Administrators" -Member "deploy"The first command creates deploy. The second adds it to the local Administrators group. When prompted, set a strong password and store it securely.
Now create a profile and prepare an SSH directory if you will use OpenSSH for remote administration. Run this in PowerShell as Administrator:
$userPath = "C:\Users\deploy"
New-Item -ItemType Directory -Path "$userPath\.ssh" -Force | Out-Null
Set-Content -Path "$userPath\.ssh\authorized_keys" -Value "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIExampleKeyForDocumentationOnly"Replace the example public key with your own. Then lock down permissions so only the account can read it:
icacls "C:\Users\deploy\.ssh" /inheritance:r
icacls "C:\Users\deploy\.ssh" /grant "deploy:(OI)(CI)F" /grant "Administrators:F"
icacls "C:\Users\deploy\.ssh\authorized_keys" /inheritance:r
icacls "C:\Users\deploy\.ssh\authorized_keys" /grant "deploy:R" /grant "Administrators:F"Open a second terminal and test the new account. Keep the original administrator session open.
ssh deploy@203.0.113.10After login, confirm you have administrative rights:
whoami /groupsYou should see the Administrators group. Do not disable the original admin access yet.
Install IIS and the management tools
On the VPS as Administrator, install IIS and the features needed for a typical hosted site. This uses the built-in Windows Server feature model rather than a Linux package manager.
Install-WindowsFeature Web-Server, Web-Mgmt-Tools, Web-Default-Doc, Web-Static-Content, Web-Http-Errors, Web-Filtering, Web-Health, Web-Performance, Web-App-Dev, Web-Asp-Net45, Web-Net-Ext45 -IncludeManagementToolsExpected result: the command finishes with Success. Then confirm the IIS service is present and running.
Get-Service W3SVCIf it is stopped, start it:
Start-Service W3SVCCheck that the server listens on port 80:
Get-NetTCPConnection -LocalPort 80 -State ListenBuild a simple IIS site directory
Create a dedicated web root instead of using the default folder. That keeps the deployment clean and makes rollback easier.
New-Item -ItemType Directory -Path "C:\Sites\example.com\public" -Force | Out-Null
Set-Content -Path "C:\Sites\example.com\public\index.html" -Value @'
IIS is live on Hostperl dedicated hardware
If you can read this, the cutover test passed.
'@Create a new application pool and site. Replace example.com only if you are using a different domain in your environment.
Import-Module WebAdministration
New-WebAppPool -Name "example.com-pool"
New-Website -Name "example.com" -Port 80 -HostHeader "example.com" -PhysicalPath "C:\Sites\example.com\public" -ApplicationPool "example.com-pool"Now confirm the site exists and is started:
Get-Website | Select-Object Name, State, PhysicalPath, BindingsExpect to see Started for the site. If the state is stopped, check the IIS event log before you move on.
Open the firewall safely before removing any old access path
On the VPS as Administrator, create the rules you need first. Only after that should you remove any old or temporary exposure.
New-NetFirewallRule -DisplayName "Allow HTTP 80" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 80
New-NetFirewallRule -DisplayName "Allow HTTPS 443" -Direction Inbound -Action Allow -Protocol TCP -LocalPort 443Confirm the rules are active:
Get-NetFirewallRule -DisplayName "Allow HTTP 80","Allow HTTPS 443" | Format-Table DisplayName, Enabled, Direction, ActionIf you use the Windows Remote Management service or RDP over a dedicated management network, keep that path open until web testing is complete. Do not close your only admin path too early.
Test the site from the server and from your client
First test locally on the server. That catches IIS and file permission issues before DNS changes.
Invoke-WebRequest -Uri http://localhost -UseBasicParsing | Select-Object StatusCode, StatusDescriptionYou should get 200. Then test from your local computer:
curl -I http://203.0.113.10Replace 203.0.113.10 with your server’s public IP. A successful response returns HTTP/1.1 200 OK or similar. If you see a timeout, check the firewall and IIS bindings.
For domain testing before DNS is switched, you can also add a temporary hosts entry on your local machine. That lets you validate the hostname without affecting production traffic.
Add TLS when the site is ready
If the site will serve customers, add a certificate before the final DNS cutover. For a public domain, use Let’s Encrypt through a Windows-compatible client such as win-acme. Install it in a maintenance directory and point it at IIS, then bind the certificate to the site. The exact client can vary, but the sequence stays the same: obtain the certificate, confirm the bind, then reload the site.
After the certificate is installed, check bindings in IIS Manager or with PowerShell:
Get-WebBinding -Name "example.com"You should see both HTTP and HTTPS bindings if the certificate is in place. Then test HTTPS from your local computer:
curl -I https://203.0.113.10If you are using a real certificate and proper hostname binding, the browser should show a valid chain once DNS points to the server.
Review logs before you switch traffic
IIS logs are usually the fastest way to spot host header or permission problems. Check the default log location and the Windows event log.
Get-ChildItem "C:\inetpub\logs\LogFiles" -Recurse | Select-Object FullName, LastWriteTime -First 5
Get-WinEvent -LogName Application -MaxEvents 20 | Select-Object TimeCreated, Id, ProviderName, MessageLook for repeated 500-level responses, authentication failures, or binding errors. If you see a blank page with a 500 status, the application pool identity or web.config is usually the first place to check.
Safe rollback if the deployment fails
Rollback should be quick and boring. Keep the old server or the previous site content available until the new server proves itself.
If the new IIS site misbehaves, stop it and restore the previous content:
Stop-Website -Name "example.com"
Copy-Item "C:\Sites\example.com\public\index.html" "C:\Sites\example.com\public\index.html.bak" -ForceIf the issue is certificate-related, remove the HTTPS binding before you retry:
Get-WebBinding -Name "example.com" -Protocol https | Remove-WebBindingThen re-test only the HTTP site. Once that works, return to TLS. If the problem is broader, change DNS back to the old host and leave the Windows server in place for debugging.
Reboot and persistence checks
Rebooting confirms that IIS, firewall rules, and site bindings survive a real restart. Run this only after you have a working site and a second admin path.
Restart-Computer -ForceAfter the server returns, reconnect and confirm the service status:
Get-Service W3SVC
Get-Website | Select-Object Name, StateThen test the site again from your local computer:
curl -I http://203.0.113.10A healthy server should still return a 200 status after reboot.
Common failures and how to diagnose them
1. The site returns 403 or 404. Check the IIS binding and default document.
Get-WebBinding -Name "example.com"
Get-ChildItem "C:\Sites\example.com\public"If index.html is missing or the host header is wrong, correct the file path or binding and retest.
2. Port 80 is closed from the client side. Confirm the firewall rule and listening socket.
Get-NetFirewallRule -DisplayName "Allow HTTP 80" | Format-List DisplayName, Enabled, Direction, Action
Get-NetTCPConnection -LocalPort 80 -State ListenIf the rule is disabled, enable it. If nothing is listening, start IIS.
3. HTTPS fails after certificate install. Check the binding and the certificate store.
Get-ChildItem Cert:\LocalMachine\My | Select-Object Subject, Thumbprint, NotAfter
Get-WebBinding -Name "example.com" -Protocol httpsBind the right certificate to the right hostname and retry the curl test.
4. The admin account cannot log in over SSH. Confirm the profile and key permissions.
icacls "C:\Users\deploy\.ssh"
icacls "C:\Users\deploy\.ssh\authorized_keys"Fix the ACLs if they do not grant the local user read access. Keep the original administrator login available until this is corrected.
Where Hostperl fits into this workflow
A windows server dedicated server is the right fit when you need predictable hardware, isolated resources, and a clean public web tier your team can manage without sharing capacity. Hostperl’s dedicated servers and enterprise dedicated hosting options suit Windows workloads that need stable uptime, migration planning, and support you can escalate when a cutover needs eyes on it.
If you are moving a customer site, internal portal, or line-of-business app onto Windows Server, plan the cutover on hardware that leaves room for IIS, logs, and future growth. Hostperl can help you choose between dedicated server hosting and enterprise dedicated hosting based on the workload and support level you actually need.
Keep the old system online until the new one passes HTTP, HTTPS, login, and reboot checks. That is the difference between a controlled migration and an avoidable outage.
FAQ
Can I use this guide on Windows Server 2025?
Yes. The IIS, firewall, service, and verification steps are valid for Windows Server 2025, with the same production checks.
Do I need PowerShell or IIS Manager?
Use PowerShell for repeatable setup and IIS Manager for visual checks. PowerShell gives you clearer rollback commands.
Should I disable the built-in Administrator right away?
No. Keep it until the deploy account works, the site responds, and the reboot test passes.
What should I test before changing DNS?
Test localhost, the public IP, the hostname through a temporary hosts entry, TLS binding, and reboot persistence.
What if my app needs PHP or another runtime?
Install the runtime after IIS is stable, then repeat the same binding, log, and smoke-test checks for that stack.
