Fix 'The Cluster Service Cannot Be Started (Error 2)' in Exchange Server DAG & Windows Failover Clustering

When managing an Exchange Server Database Availability Group (DAG) or a Windows Server Failover Cluster, one of the most critical errors an administrator can encounter after an unclean shutdown, failed DAG member removal, or registry hive rollback is:

The Cluster service cannot be started. An attempt to read configuration data from the Windows registry failed with error '2'. Please use the Failover Cluster Management snap-in to ensure that this machine is a member of a cluster. If you intend to add this machine to an existing cluster use the Add Node Wizard. Alternatively, if this machine has been configured as a member of a cluster, it will be necessary to restore the missing configuration data that is necessary for the Cluster Service to identify that it is a member of a cluster. Perform a System State Restore of this machine in order to restore the configuration data.

Despite the alarming recommendation in the error message to "Perform a System State Restore", you almost never need to restore the OS from backup on an Exchange DAG member. In this guide, we break down why Error 2 (ERROR_FILE_NOT_FOUND on the HKLM\Cluster registry hive) happens and how to recover the node cleanly in minutes.


Why Cluster Service Error 2 Happens & Recovery Flow

The Windows Failover Cluster service (ClusSvc) stores its local cluster database hive in %SystemRoot%\Cluster\CLUSDB and mounts it into the Windows Registry at HKEY_LOCAL_MACHINE\Cluster.

When a node is partially evicted from a DAG while offline, experiences disk corruption on CLUSDB, or loses quorum state during a failed Remove-DatabaseAvailabilityGroupServer operation, the ClusSvc service remains set to Automatic, but the HKLM\Cluster registry hive is missing or unloaded—triggering Win32 Error 2 (The system cannot find the file specified).

Exchange DAG Cluster Service Error 2 Diagnostic & Recovery Workflow
Symptom on DAG Node
ClusSvc Fails with Error 2
  • HKLM\Cluster hive missing or corrupted
  • Databases on node stuck in Disconnected / ServiceDown
Step 1: Evict Local State
Clear Stale Cluster Metadata
  • Run Clear-ClusterNode -Force
  • Or cluster.exe node <Server> /forcecleanup
Step 2: Re-Join DAG Cluster
Rebuild CLUSDB from Healthy DAG
  • Datacenter Activation Mode: Start-DatabaseAvailabilityGroup
  • Standard DAG: Re-add node or Add-ClusterNode

Step 1: Verify Current Cluster & DAG Membership Status

Before running cleanup commands, open an elevated Exchange Management Shell on a healthy DAG member to check how Exchange and Windows Failover Clustering currently see the broken server (e.g., SITEA-MBX01):

POWERSHELL
# 1. Check Exchange DAG stopped vs. started mailbox servers
Get-DatabaseAvailabilityGroup -Identity "DAG01" -Status |
    Select-Object Name, PrimaryActiveManager, Servers, StartedMailboxServers, StoppedMailboxServers

# 2. Check Windows Failover Cluster node states
Import-Module FailoverClusters
Get-ClusterNode -Cluster "DAG01" | Select-Object Name, Id, State, NodeWeight
ScenarioWhat You SeeWhich Recovery Procedure to Use
Scenario A: Node was stopped via Stop-DatabaseAvailabilityGroupServer is listed in StoppedMailboxServersProcedure A (Force cleanup local node + Start-DatabaseAvailabilityGroup)
Scenario B: Node crashed / lost HKLM\Cluster during normal operationsServer is in Servers, ClusSvc won't start with Error 2Procedure B (Force cleanup local node + re-join cluster)
Scenario C: Majority of DAG nodes are down (Lost Quorum)Cluster service stopped across all nodes (Event ID 1135 / 1177)Procedure C (Start-ClusterNode -FixQuorum)

Step 2: Recovery Procedures

Procedure A: Node Was Stopped or Partially Evicted (Standard DAG Recovery)

If the broken node (SITEA-MBX01) throws Error 2 when starting ClusSvc:

  1. Log in directly to the affected Exchange Server (SITEA-MBX01) and open PowerShell (or Command Prompt) as Administrator.
  2. Force-clean the corrupted local cluster configuration so Windows resets the local ClusSvc state:
POWERSHELL
# Modern PowerShell equivalent (run on the affected server SITEA-MBX01)
Import-Module FailoverClusters
Clear-ClusterNode -Name $env:COMPUTERNAME -Force

# Or using classic cluster.exe CLI:
cluster.exe node $env:COMPUTERNAME /forcecleanup
  1. Now open the Exchange Management Shell and run Start-DatabaseAvailabilityGroup to pull a fresh copy of the cluster database (CLUSDB) from the active DAG members and re-incorporate the server:
POWERSHELL
# Re-incorporate the Mailbox Server into DAG01 and start ClusSvc
Start-DatabaseAvailabilityGroup -Identity "DAG01" -MailboxServer "SITEA-MBX01"

Procedure B: Node Was Removed from Cluster Membership While Offline

If Start-DatabaseAvailabilityGroup reports that the server is not listed in StoppedMailboxServers, clean up the node's DAG membership metadata with -ConfigurationOnly and re-add it to the DAG—this does not delete your passive mailbox database .edb files on disk:

POWERSHELL
# 1. On the affected server (SITEA-MBX01), ensure local cluster state is cleaned:
Clear-ClusterNode -Name "SITEA-MBX01" -Force

# 2. From a healthy Exchange Server, clear the stale DAG configuration entry:
Remove-DatabaseAvailabilityGroupServer -Identity "DAG01" -MailboxServer "SITEA-MBX01" -ConfigurationOnly -Confirm:$false

# 3. Re-add the server back into the DAG (automatically installs/configures ClusSvc and syncs CLUSDB):
Add-DatabaseAvailabilityGroupServer -Identity "DAG01" -MailboxServer "SITEA-MBX01"

# 4. Verify all database copy replication health
Test-ReplicationHealth -Identity "SITEA-MBX01"

Procedure C: Entire DAG Cluster Lost Quorum (Start-ClusterNode -FixQuorum)

If a power outage took down the majority of DAG nodes and ClusSvc refuses to start because it cannot achieve quorum:

POWERSHELL
Import-Module FailoverClusters

# 1. Force-start the cluster service on the primary surviving Exchange Server without quorum
Start-ClusterNode -Name "SITEA-MBX01" -FixQuorum

# 2. Once ClusSvc starts on the first node, start ClusSvc normally on the remaining nodes
Get-Service ClusSvc -ComputerName "SITEA-MBX02" | Start-Service

# 3. Verify Cluster Group (PAM) and Quorum state
Get-ClusterGroup -Cluster "DAG01"

Step 3: Post-Recovery Health Verification Script

After ClusSvc starts cleanly, run this diagnostic script in the Exchange Management Shell to confirm the HKLM\Cluster registry hive is mounted, ClusSvc and MSExchangeRepl are running, and all database copies are healthy:

POWERSHELL
function Test-ExchangeDagNodeClusterHealth {
    [CmdletBinding()]
    param (
        [Parameter(Mandatory = $false)]
        [string]$ServerName = $env:COMPUTERNAME
    )

    $ClusSvc  = Get-Service -Name "ClusSvc" -ComputerName $ServerName -ErrorAction SilentlyContinue
    $ReplSvc  = Get-Service -Name "MSExchangeRepl" -ComputerName $ServerName -ErrorAction SilentlyContinue
    $HivePath = Invoke-Command -ComputerName $ServerName -ScriptBlock { Test-Path "HKLM:\Cluster" } -ErrorAction SilentlyContinue

    [PSCustomObject]@{
        ServerName            = $ServerName
        ClusterServiceStatus  = if ($ClusSvc) { $ClusSvc.Status } else { "Not Installed" }
        ClusterServiceStart   = if ($ClusSvc) { $ClusSvc.StartType } else { "N/A" }
        HKLMClusterHiveLoaded = [bool]$HivePath
        ExchangeReplStatus    = if ($ReplSvc) { $ReplSvc.Status } else { "Not Found" }
    }

    Write-Host "`n[*] Running Test-ReplicationHealth on $ServerName..." -ForegroundColor Cyan
    Test-ReplicationHealth -Identity $ServerName | Format-Table Server, Check, Result, Error -AutoSize
}

# Example Usage:
# Test-ExchangeDagNodeClusterHealth -ServerName "SITEA-MBX01"