Upravit

Sdílet prostřednictvím


Cluster service fails to start troubleshooting guidance

Troubleshooting checklist

Check the ports that the cluster service uses

Make sure that the following ports are open to cluster traffic on any firewalls:

  • Port 135: Remote procedure call (RPC) endpoint mapper or distributed component object model (DCOM).

  • Port 135: RPC endpoint mapper over user datagram protocol (UDP).

  • Port 3343: Cluster network driver.

  • Port 445: Server message block (SMB).

  • Port 139: NetBIOS session service.

  • Ports in the range of 5000 to 5099: If Event ID 1721 is logged when you connect to a cluster as a cluster administrator, try opening the ports in this range (or other ports) to RPC traffic. The ports support communication through RPC unless you just type a period character (.).

    This issue can occur because the cluster service uses at least 100 ports for RPC communication. The number of ports available to the cluster service can become too small when other services use some of the necessary ports. These services may include Windows DNS service, Windows Internet Name service (WINS), or Microsoft SQL Server service.

  • Ports in the range of 8011 to 8031: If firewalls separate the cluster nodes, the ports in the range of 8011 to 8031 must be open to internode RPC traffic. Otherwise, errors in the cluster log indicate that a sponsor node isn't available. These errors occur because there aren't enough ports available for RPC communication between a node that tries to join the cluster and a node that can sponsor that node.

For more information about how to configure a network and network ports for a cluster, see the following articles:

After you change the port settings, try to bring the node online again before you proceed.

Run the cluster validation tool

  1. Open the Failover Cluster Manager snap-in (CluAdmin.msc).

  2. Select Failover Cluster Manager in the top left column.

  3. Select Validate Configuration.

  4. Type the name of each node in the cluster and select Add after each one.

  5. When all nodes have been added to the Selected servers: list, select Next.

  6. Select Run all tests (recommended) > Next > Next.

  7. Allow the test to finish. Once complete, select View Report.

  8. Review any tests results that are labeled as Failed or Warning. This information may help provide actionable steps to fix the issue.

  9. To get a downloadable file, navigate to the C:\Windows\Cluster\Reports folder and open the Validation Report (.MHT) file.

    Note

    In Windows Server 2016 and later versions, it's an .HTM file.

Check the security policies that might affect the cluster node

In the Group Policy Object Editor, these policy objects are located in Computer Configuration\Windows Settings\Security Settings\Local Policies\User Rights Assignment.

Note

To access the local security policy settings, select Start, type local security policy, and then select Local Security Policy.

After you change the policy settings, try to bring the node online again before you proceed.

Temporarily disable firewalls

Disable the firewall between the node and the rest of the cluster, and then try to bring the node online again. If the node still doesn't come online, the firewall can be the cause.

Important

Don't leave this change in place after you finish troubleshooting. After you use this change for testing, return these settings to the original configuration.

Check the network hardware and software for issues

  • Check the System event log for hardware or software errors that are related to the network adapters on this node.

  • Check the network adapter, cables, and network configuration for the networks that connect the nodes.

  • If you're teaming the network adapters, make sure that the teaming configuration is correct.

  • Check hubs, switches, or bridges in the networks that connect the nodes.

Review log files

To identify the source of the issue, review log information from multiple sources. For example:

  • In Event Viewer, navigate to Applications and Services Logs\Microsoft\Windows\FailoverClustering-Client\Diagnostic, and review the Cluster API Debug Tracing logs.

  • Generate a fresh cluster log for the node. On the server that is running the affected node, open an elevated PowerShell prompt and run the following cmdlet:

    Get-ClusterLog -Node 'Local Node Name' -Destination c:\temp -UseLocalTime

To generate a more detailed trace, follow these steps:

  1. At an elevated PowerShell prompt, run the following cmdlet to start the trace:

    logman create trace "base_cluster" -ow -o c:\base_cluster.etl -p "Microsoft-Windows-FailoverClustering-Client" 0xffffffffffffffff 0xff -nb 16 16 -bs 1024 -mode Circular -f bincirc -max 4096 -ets

  2. Reproduce the issue.

  3. To stop the trace, run the following cmdlet:

    Logman stop base_cluster.etl -ets

  4. To convert the trace, run the following cmdlet:

    Netsh trace convert base_cluster.etl

  5. To generate a cluster log from the data, run the following cmdlet:

    Get-ClusterLog -Node 'Local Node Name' -Destination c:\temp -UseLocalTime

For more information about tracing and other issues to look out for, see How to Troubleshoot Create Cluster Failures.