Great — below is a clear, step-by-step **cleanup + rebuild** guide for a 2-node
Windows Failover Cluster when nodes were renamed / removed from AD and DNS and the
cluster is now broken. Follow carefully (especially the warnings) — I’ve included
the exact PowerShell commands you can run (replace placeholders with your real
names).
---
## ⚠️ Before you start (must-do)
1. **Take VM snapshots** / backups of both nodes and (if applicable) the domain
controllers.
2. You need **Domain Admin** (or equivalent) rights and local Administrator on both
cluster nodes.
3. Work during a maintenance window; this will remove cluster objects and may
affect services.
4. Note current names/IPs: `Node1`, `Node2`, `Domain = testvm.com`,
`DesiredClusterName = MyCluster`, `ClusterIP = 192.168.1.100`.
---
## 0) Collect current state (info)
Run from an admin PowerShell on a management host (or node) that has the
FailoverClusters module:
```powershell
Import-Module FailoverClusters
# See if any cluster exists / is discoverable
Get-Cluster # shows cluster(s) if reachable
# Check local service
Get-Service ClusSvc
# Show computer name & domain
$env:COMPUTERNAME
(Get-WmiObject -Class Win32_ComputerSystem).Domain
```
---
## 1) If the old cluster is still reachable (best case)
If `Get-Cluster` returns the cluster object and it is contactable, remove it
cleanly:
```powershell
Import-Module FailoverClusters
# Optional: test cluster
Test-Cluster -Node Node1,Node2
# Remove whole cluster (run as admin on a management node that can reach the
cluster)
Remove-Cluster -Name "MyCluster" -Force
```
After `Remove-Cluster`, check AD & DNS (step 3) and proceed to re-create.
---
## 2) If the cluster is broken / not contactable (most likely your situation)
Do this **on each cluster node (Node1 and Node2)**, logged in as local
Administrator or domain admin (Run PowerShell as Admin):
### 2.1 Stop Cluster Service (ignore errors)
```powershell
Stop-Service ClusSvc -Force -ErrorAction SilentlyContinue
```
### 2.2 Clear local cluster metadata
```powershell
Import-Module FailoverClusters
Clear-ClusterNode -Force
```
`Clear-ClusterNode` removes cluster configuration cached on that node and
unregisters it locally. If it errors, note messages — you can try again after
reboot.
> **Only if Clear-ClusterNode fails repeatedly** (and you understand the risk):
export and remove the registry key:
```powershell
reg export "HKLM\Cluster" C:\temp\Cluster_HKLM_Backup.reg
# then (ONLY if you backed up and are prepared)
Remove-Item -Path "HKLM:\Cluster" -Recurse -Force
```
(Manual registry delete is dangerous — backup first.)
### 2.3 Reboot the node
```powershell
Restart-Computer -Force
```
Repeat 2.1–2.3 on the other node.
---
## 3) Clean up AD & DNS (domain controller or machine with RSAT)
You must remove the old **Cluster Name Object (CNO)** and any **Virtual Computer
Objects (VCOs)** that belonged to the old cluster; also remove DNS A/PTR records.
### 3.1 Check & delete CNO and VCOs in ADUC
* Open **Active Directory Users and Computers** → enable **Advanced Features** →
search for the cluster computer object (`MyCluster`) and any VCOs (names of cluster
roles).
* **Right-click → Delete**. If protected from accidental deletion, uncheck that
protection first.
Or use PowerShell (on a host with the ActiveDirectory module):
```powershell
Import-Module ActiveDirectory
# Example: delete CNO (cluster computer object)
Remove-ADComputer -Identity "MyCluster$" -Confirm:$false
# Delete any VCO (example)
Remove-ADComputer -Identity "MyRoleVCO$" -Confirm:$false
```
> If you get replication/permission issues, wait for AD replication or fix
permissions first.
### 3.2 Remove DNS records (on your DNS server)
Use DNS Manager GUI or PowerShell on the DNS server:
```powershell
# Remove A record for cluster name
Remove-DnsServerResourceRecord -ZoneName "testvm.com" -Name "MyCluster" -RRType "A"
-Force
# If PTR exists, remove it too (adjust zone)
Remove-DnsServerResourceRecord -ZoneName "1.168.192.in-addr.arpa" -Name "100" -
RRType "PTR" -Force
```
(If you don't have the DNS Server module, use DNS Manager.)
---
## 4) Ensure nodes are healthy in AD & DNS
On each node:
* If they were removed from AD, **rejoin the domain**:
```powershell
Add-Computer -DomainName "testvm.com" -Credential (Get-Credential) -Restart
```
* If already joined but renamed, ensure the DNS records are correct:
```powershell
ipconfig /registerdns
nltest /dsgetdc:testvm.com
# check resolution
Resolve-DnsName Node1.testvm.com
```
* Restart Netlogon if you see DNS/AD issues:
```powershell
Restart-Service Netlogon
```
---
## 5) Prepare nodes for new cluster
On BOTH nodes:
### 5.1 Install Failover Clustering feature (if not present)
```powershell
Install-WindowsFeature -Name Failover-Clustering -IncludeManagementTools
```
### 5.2 Validate hardware & network & storage (always run)
```powershell
Test-Cluster -Node Node1,Node2 -Verbose
```
Fix any issues reported by the tests.
---
## 6) Create the new cluster
From one of the management hosts or Node1 (with admin privileges):
```powershell
Import-Module FailoverClusters
# Example: create cluster without adding storage
New-Cluster -Name "MyCluster" -Node Node1,Node2 -StaticAddress 192.168.1.100 -
NoStorage
```
* If you want cluster shared storage, omit `-NoStorage` and ensure storage is
visible & passed to nodes.
* If `New-Cluster` fails due to permission to create CNO in AD, either:
* Pre-create the computer object `MyCluster` in AD (ADUC) and delegate
permissions, **or**
* Give the account you use the rights to create computer accounts in that OU
(delegation).
After creation, the cluster will create CNO in AD and DNS A record for `MyCluster`.
---
## 7) Post-creation checks
* Verify cluster nodes:
```powershell
Get-Cluster
Get-ClusterNode
Get-ClusterResource
```
* Check AD & DNS to confirm CNO and VCOs exist.
* If you plan to host roles (FS, SQL, etc.) create and validate them.
---
## 8) Common troubleshooting notes
* **Error 5043** typically resolves after `Clear-ClusterNode` + AD/DNS cleanup +
reboot.
* If cluster creation fails with AD permission errors, pre-create the CNO or adjust
OU permissions.
* AD replication delays: allow time for deleted objects to replicate across DCs.
* If something still shows as “joined to a cluster”: collect cluster logs:
```powershell
Get-ClusterLog -Destination C:\temp\ClusterLogs
```
Share logs if you want me to help analyze.
---
## Quick Summary (one-liner checklist)
1. Snapshot VMs.
2. On both nodes: `Stop-Service ClusSvc`, `Clear-ClusterNode`, `Restart-Computer`.
3. In AD: delete old CNO and VCOs.
4. In DNS: delete old cluster A/PTR records.
5. Rejoin nodes to domain if needed; `ipconfig /registerdns`.
6. Install Failover-Clustering, run `Test-Cluster`.
7. `New-Cluster -Name MyCluster -Node Node1,Node2 -StaticAddress 192.168.1.100`.
8. Verify and create roles.
---
Would you like a single PowerShell script (one to run on each node and another to
run on a management/DC host) that automates the node cleanup + AD/DNS removal tasks
(I’ll include safety prompts and backups)? If yes, tell me the exact node names,
desired cluster name and cluster IP and I’ll generate the scripts.
