Troubleshooting cluster node installations

项目
10/08/2009

Applies To: Windows Server 2003, Windows Server 2003 R2, Windows Server 2003 with SP1, Windows Server 2003 with SP2

Troubleshooting cluster node installations

What problem are you having?

During cluster creation, the New Server Cluster Wizard did not detect all of the potential cluster disks.
During cluster creation, the New Server Cluster Wizard configured all of the available storage devices to be managed by the cluster, including storage that you did not want to be managed by the cluster.
When trying to add a node to a cluster, you receive an error stating that a multinode cluster cannot be created because there is not a quorum-capable resource common to all nodes.
During cluster creation, either the New Server Cluster Wizard or the Add Nodes Wizard did not detect all of the potential network adapters.
During cluster creation, you are experiencing problems creating the first node of the cluster.
During cluster creation, you are experiencing problems creating subsequent nodes after creating the first node of the Cluster.
During cluster creation, the Physical Disk resources are not automatically created for cluster disks.
After setting up a new cluster, Cluster Administrator cannot connect to the cluster.

During cluster creation, the New Server Cluster Wizard did not detect all of the potential cluster disks.

Possible Cause: Your hardware, device driver, or software is not configured correctly.

Solution: Check Disk Management to see whether the disks you want to use for shared storage are available.

See also: Disk Management

Possible Cause: You are using incompatible disk hardware or drivers.

Solution: Check your disk cabling to make sure that the cables are connected correctly between the node in question and the shared storage.

Confirm that your hardware, including your cluster disks and other cluster hardware, is compatible with Windows Server 2003, Datacenter Edition, by clicking the appropriate link in Support resources.

Make sure that the host adapter is configured correctly according to the manufacturer's instructions, and that all disks are visible to the host adapter.

Make sure that you have the drivers you need in order to use the disks with the Cluster service. Drivers are available from your hardware manufacturer. Review the manufacturer's instructions carefully before you begin installing cluster hardware.

Note

If you make any changes to your disk hardware or drivers, you must restart the New Server Cluster Wizard for the changes to be detected.

Possible Cause: Your cluster storage disks are not formatted as NTFS.

Solution: Make sure all partitions on the cluster storage disks are formatted to use the NTFS file system. Any disk with a FAT or FAT32 partition will not be recognized during cluster creation.

See also: NTFS

Possible Cause: Your cluster storage disks are not configured as basic MBR disks.

Solution: Make sure your cluster storage disks are configured as basic MBR disks. Windows Server 2003, Datacenter Edition, offers two types of storage configurations: basic disk and dynamic disk. Basic disks use the same disk structures used by Microsoft Windows NT. Dynamic disks were introduced in Windows 2000.

Cluster servers can recognize only basic master boot record (MBR) disks. If a disk you want to use as shared storage is configured as a dynamic or GUID partition table (GPT) disk, change it back to a basic MBR disk.

During cluster creation, the New Server Cluster Wizard configured all of the available storage devices to be managed by the cluster, including storage that you did not want to be managed by the cluster.

Cause: With the Cluster Administrator New Server Cluster Wizard, you can choose between Typical (full) configuration and Advanced (minimum) configuration options. The Typical configuration configures all disks as clustered disks and creates cluster resources for them all. The only exception is the system disk, which contains the operating system and is the disk from which the computer boots from. The system disk will not be automatically clustered.

Solution: Use the Advanced (minimum) configuration option. With the Advanced configuration option, you can keep certain disks private because the configuration creates a cluster in which only the quorum disk is managed by the cluster (if you chose to use a physical disk as the quorum resource). After the cluster is created, you must then use Cluster Administrator to add any other disks that you want the cluster to manage.

When trying to add a node to a cluster, you receive an error stating that a multinode cluster cannot be created because there is not a quorum-capable resource common to all nodes.

Cause: The wizard was not able to validate your storage configuration. In some complex storage solutions that contain several switches, such as a fiber channel switched fabric, a particular storage unit might have a different identity on each computer in the cluster. Although this is a valid storage configuration, it violates the storage validation heuristics in the New Server Cluster Wizard and the Add Nodes Wizard.

Solution: Verify that your storage configuration is set up correctly. Then, if you are in the Add Nodes Wizard, disable the storage validation heuristics by using the Back button to return to the Select Computer page. On the Select Computer page, click Advanced and select the Advanced (minimum) configuration option. If you are not in the Add Nodes Wizard, restart the wizard and use the Advanced configuration option as previously described.

During cluster creation, either the New Server Cluster Wizard or the Add Nodes Wizard did not detect all of the potential network adapters.

Possible Cause: The cluster may not be able to detect a network composed of a crossover cable.

A two-node cluster may use a crossover cable to implement a private network. When the cluster is created on the first node, the second node should be turned on, but stopped in the BIOS or at the startup menu. In this state, the Media Sense feature of Windows may not recognize the network adapter as connected.

When this happens:

In Network Connections, the icon for the connection that represents the network adapter is shown with a red "X".
The taskbar displays an icon that, when you pause the mouse on it, displays Network Cable unplugged.

Solution: If you continue creating the cluster, the crossover cable will be detected when you start the second node. The network will be established in the default mode, which is mixed. You can then change the network mode to private using Cluster Administrator.

See also: Manage Networks

Possible Cause: Your system's network connections are not running. If the following error message appears in the New Server Cluster Wizard tree, your network connections may not be running:

"Warning: This system is not attached to all defined cluster networks."

Solution: Make sure that your network connections are enabled. Otherwise, the Cluster service cannot detect and use them.

To check your network connections:

Open Network Connections in Control Panel.
Double-click the connection that you want to check.
On the General tab, check the status of the connection. If the connection is not started, start it.

You must click Reanalyze in the New Server Cluster Wizard for the changes to be detected.

See also: TCP/IP

Possible Cause: Your network hardware is not installed correctly.

Solution: Verify that your network cables are connected correctly. Make sure that no cables are crossed by mistake (private-to-public, or public-to-private), and that each network adapter is connected as you planned. With more than two nodes, use a dedicated hub (not a switch) for the private network.

Set all multispeed network adapters on a private or mixed network to the same speed. Do not use the autodetect setting.

If fault-tolerant network adapters are used, create multiple private networks instead of a single fault-tolerant network.

Make sure all network adapters are correctly seated and the correct drivers are installed in accordance with the manufacturer's instructions. Confirm that your hardware is compatible with Windows Server 2003, Datacenter Edition, by clicking the appropriate link in Support resources.

Possible Cause: The Cluster networks are not on the correct subnet. All network adapters designated for cluster use on a single node need to be on different subnets. Otherwise, only the first network adapter on a subnet is detected by either the New Server Cluster Wizard or the Add Nodes Wizard, and is used for either cluster or client communication (depending on which role you chose).

However, in order for subsequent nodes to pick up the network roles defined on the first node in the cluster, network adapters in each node must be on the same subnets as the corresponding adapters on the first node.

Solution: The Cluster service can detect and use multiple network adapters on the same computer if the following requirements are met:

The network adapters are bound by TCP/IP.
You are using hardware that is compatible with Windows Server 2003, Datacenter Edition. To confirm this, click the appropriate link in Support resources.
Each network adapter is on a different subnet.

Do not use the same IP address for two network adapters, even though they are connected to two different networks.

After the network adapters are configured correctly, use the ping command to check each network adapter for connectivity with the loopback address (127.0.0.1), the adapter's own IP address, and the IP address of another computer.

Make sure that each adapter works correctly and can communicate on each network, and then restart the New Server Cluster Wizard or the Add Nodes Wizard.

Possible Cause: The cluster does not have network connectivity.

Solution: Use the ping command to test connectivity. Ping the node's fully qualified domain name.

If you are unable to ping the node's domain controller by name, consult the Hardware Troubleshooter. You can also learn more about the problem by pinging your domain controller by IP address, and by pinging other network servers by name and IP address, including the node's DNS or WINS server.

See also: Manage Networks

During cluster creation, you are experiencing problems creating the first node of the cluster.

Possible Cause: You are not using a valid IP address for the network interface.

Solution: The addresses for each linked pair of network adapters must be on the same subnet.

To check for the IP address, Click Start, then Run and type ipconfig /all, and then press ENTER. This displays your IP configuration, including IP address, subnet mask, default gateway, DNS server address, and DHCP availability. Confirm that these settings are correct. If you are using DHCP, change to static IP addresses.

If no IP address or other IP configuration is displayed, and you see Media State = Cable Disconnected, you may not have end-to-end network connectivity from the network adapter to the connecting network device. Make sure your physical connections are secure and valid. If you see an Automatic Private IP Addressing (APIPA) address, change it to a static IP.

Do not use an IP address that is on the same IP range as the public network.

See also: TCP/IP

Possible Cause: The cluster does not have network connectivity.

Solution: Use the ping command to test connectivity. Ping the node's fully qualified domain name.

If you are unable to ping the node's domain controller by name, open Network Connections and check the status of your network adapter. You can also attempt to map a drive to the other node in your cluster to verify communication and authentication. If these solutions do not work, consult the Hardware Troubleshooter. You can also learn more about the problem by pinging your domain controller by IP address, and by pinging other network servers by name and IP address, including the node's DNS or WINS server.

See also: Manage Networks

Possible Cause: You are not using a unique cluster name.

Solution: The cluster name must be different from the domain name, from all computer names on the network, from other cluster names on the network, and from all virtual server names on the network.

Use the net view command to determine whether your cluster name is unique.

To use the net view command, click Start, then Run and type the following and then press ENTER:

net view clustername

Clustername is the name of the cluster.

If your cluster name is not unique, assign a unique name. When the New Server Cluster Wizard asks you for the cluster name, type an appropriate new cluster name.

Possible Cause: The cluster may not be able to detect a network composed of a crossover cable.

When this happens:

In Network Connections, the icon for the connection that represents the network adapter is shown with a red "X".
The taskbar displays an icon that, when you pause the mouse on it, displays Network Cable unplugged.

See also: Manage Networks

Solution: The Cluster service can detect and use multiple network adapters on the same computer if the following requirements are met:

The network adapters are bound by TCP/IP.
You are using hardware that is compatible with Windows Server 2003, Datacenter Edition. To confirm this, click the appropriate link in Support resources.
Each network adapter is on a different subnet.

Do not use the same IP address for two network adapters, even though they are connected to two different networks.

Make sure that each adapter works correctly and can communicate on each network, and then restart the New Server Cluster Wizard or the Add Nodes Wizard.

Possible Cause: Your network hardware is not installed correctly.

Set all multispeed network adapters on a private or mixed network to the same speed. Do not use the autodetect setting.

If fault-tolerant network adapters are used, create multiple private networks instead of a single fault-tolerant network.

Possible Cause: The WINS client is not configured correctly on the node.

Solution: If you use WINS, all cluster resource names that have associated IP addresses are registered with WINS. Use the following instructions to check your WINS configuration.

First, check to see whether the client interconnect has at least one WINS server specified in the IP configuration by clicking Start, then Run and typing ipconfig /all. Confirm that the IP address returned for primary and (if listed) secondary WINS servers are the same IP addresses being used successfully on other computers on the same subnet that are looking for the same resources.

Next determine whether the name is resolved by clicking Start, then Run and typing ping server, where server is the network name of the server, and then press ENTER.

If you are using static IP addressing for your cluster nodes, you need to configure WINS on the IP addresses made available on the public network.

If clients have problems connecting to virtual servers, disable WINS on the private cluster network.

Important

Do not disable this for the network adapter used for client communication.

After you configure WINS, if you have already created the cluster on at least one node, ping the cluster name to test for connectivity.

See also: Configure TCP/IP to use WINS

Possible Cause: TCP/IP is not configured correctly.

Solution: If any of the following ping tests fail, there is a TCP/IP configuration problem:

Pinging the loopback address
Pinging the local Internet protocol (IP) address (public and private interconnects)
Pinging the IP address of another node (public and private interconnects)

Note

To test name resolution through the public interconnect, disconnect the private interconnect cable.

If pinging another node by name fails, but you can successfully ping that node's IP address, there is a name-resolution problem. Consult your network administrator for assistance.

TCP/IP errors caused by an incorrect IP address or subnet mask are usually recorded in the system log. Use Event Viewer to check the system log for these errors.

Even if there are no TCP/IP errors in the system log, still check your TCP/IP settings to make sure that they are correct by clicking Start, then Run and typing ipconfig /all. Verify that the listed IP address and subnet mask are the correct values for your computer. If you do not know the correct IP address and subnet mask values for your computer, ask your network administrator.

See also: TCP/IP

Possible Cause: Your system's network connections are not running. If the following error message appears in the New Server Cluster Wizard tree, your network connections may not be running:

"Warning: This system is not attached to all defined cluster networks."

Solution: Make sure that your network connections are enabled. Otherwise, the Cluster service cannot detect and use them.

To check your network connections:

Open Network Connections in Control Panel.
Double-click the connection that you want to check.
On the General tab, check the status of the connection. If the connection is not started, start it.

You must click Reanalyze in the New Server Cluster Wizard for the changes to be detected.

See also: TCP/IP

Possible Cause: You are not using a valid IP address.

Solution: Obtain a static IP address for the cluster itself. You cannot use DHCP to assign this address.

Each network adapter on each cluster node also needs an IP address. Assign a static IP address to each adapter. The addresses for each linked pair of network adapters (linked node-to-node) must be on the same subnet.

If your cluster nodes use DHCP to obtain noncluster IP addresses, at the command prompt, use the ipconfig command to verify that you have a valid primary IP address for the networks in question.

Use the ping command to confirm that you have IP connectivity to the cluster address and that the IP address is assigned to the correct network. If the ping test fails, you do not have basic network connectivity.

If you made any changes, restart the New Server Cluster Wizard. You must restart this wizard for the changes to be detected.

See also: Manage Networks

Possible Cause: You do not have the correct domain account and appropriate user rights.

Solution: Ensure that you are adhering to the following requirements:

To create a cluster, you must log on with an account that is a member of the local Administrators group on each node. It is recommended that you use a domain account. Each node must also be a member of that domain, or must be a member of a domain that is trusted by the domain that contains the account you are logging on to.
You can administer the Cluster service by using either a domain or a local account; however, other nodes can only be joined to a cluster by using a domain account.

If you make any account changes or changes to user rights, you must click Reanalyze in the New Server Cluster Wizard for the changes to be detected. You might also have to log off and back on again for the security changes to take effect.

See also: Security

During cluster creation, you are experiencing problems creating subsequent nodes after creating the first node of the Cluster.

Possible Cause: You are not using a valid IP address for the network interface.

Solution: The addresses for each linked pair of network adapters must be on the same subnet.

To check for the IP address, click Start, then Run and type ipconfig /all, and then press ENTER. This displays your IP configuration, including IP address, subnet mask, default gateway, DNS server address, and DHCP availability. Confirm that these settings are correct. If you are using DHCP, change to a static IP address.

If no IP address or other IP configuration is displayed, and you see Media State = Cable Disconnected, you might not have end-to-end network connectivity from the network adapter to the connecting network device. Make sure your physical connections are secure and valid.

See also:TCP/IP

Possible Cause: The cluster does not have network connectivity.

Solution: Use the ping command to test connectivity. Ping the node's fully qualified domain name.

See also:Manage Networks

Possible Cause: The cluster might not be able to detect a network composed of a crossover cable.

A two-node cluster might use a crossover cable to implement a private network. When the cluster is created on the first node the second node should be turned on but stopped in the BIOS or at the startup menu. In this state, the Media Sense feature of Windows might not recognize the network adapter as connected.

When this happens:

In Network Connections, the icon for the connection that represents the network adapter is shown with a red "X".
The taskbar displays an icon that, when you pause the mouse on it, displays Network Cable unplugged.

See also:Manage Networks

Possible Cause: The Cluster networks are not on the correct subnet.

All network adapters designated for cluster use on a single node need to be on different subnets. Otherwise, only the first network adapter on a subnet is detected by either the New Server Cluster Wizard or the Add Nodes Wizard and is used for either cluster or client communication (depending on which role you chose).

Solution: The Cluster service can detect and use multiple network adapters on the same computer if the following requirements are met:

The network adapters are bound by TCP/IP.
You are using hardware that is compatible with Windows Server 2003, Datacenter Edition. To confirm this, click the appropriate link in Support resources.
Each network adapter is on a different subnet. Make sure to verify your subnet entry for each node.

Do not use the same IP address for two network adapters, even though they are connected to two different networks.

After the network adapters are configured correctly, use the ping command to check each network adapter for connectivity with the loopback address (127.0.0.1), the adapter's own IP address, and the IP address of another computer.

Make sure that each adapter works correctly and can communicate on each network, and then restart either the New Server Cluster Wizard or the Add Nodes Wizard

Possible Cause: Your network hardware is not installed correctly.

Check Device Manager for resource conflicts.

Set all multispeed network adapters on a private or mixed network to the same speed. Do not use the autodetect setting.

If fault-tolerant network adapters are used, create multiple private networks instead of a single fault-tolerant network.

See also:Device Manager

Possible Cause: The WINS client is not configured correctly on the node.

Solution: If you use WINS, all cluster resource names that have associated IP addresses are registered with WINS. Use the following instructions to check your WINS configuration.

Next determine whether the name is resolved by clicking Start, then Run and typing ping server, where server is the network name of the server, and then press ENTER.

If you are using static IP addressing for your cluster nodes, you need to configure WINS on the IP addresses made available on the public network.

If clients have problems connecting to virtual servers, disable NetBios over TCP/IP on the private cluster network.

Important

Do not disable this for the network adapter used for client communication.

After you configure WINS, if you have already created the cluster on at least one node, ping the cluster name to test for connectivity.

See also:Configure TCP/IP to use WINS

Possible Cause: TCP/IP is not configured correctly.

Solution: If any of the following ping tests fail, there is a TCP/IP configuration problem:

Pinging the loopback address
Pinging the local Internet protocol (IP) address (public and private interconnects)
Pinging the IP address of another node (public and private interconnects)

Note

To test name resolution through the public interconnect, disconnect the private interconnect cable.

If pinging another node by name fails, but you can successfully ping that node's IP address, there is a name-resolution problem. Consult your network administrator for assistance.

TCP/IP errors caused by an incorrect IP address or subnet mask are usually recorded in the system log. Use Event Viewer to check the system log for these errors.

Verify the network adapter's bindings and priority. The public network adapter or adapters should be first, then any private network adapters.

See also:TCP/IP

Possible Cause: Your system's network connections are not running. If the following error message appears in the New Server Cluster Wizard tree, your network connections may not be running:

"Warning: This system is not attached to all defined cluster networks."

Solution: Make sure that your network connections are enabled. Otherwise, the Cluster service cannot detect and use them.

To check your network connections:

Open Network Connections in Control Panel.
Double-click the connection that you want to check.
On the General tab, check the status of the connection. If the connection is not started, start it.

You must click Reanalyze in the New Server Cluster Wizard for the changes to be detected.

See also:TCP/IP

Possible Cause: You are not using a valid IP address.

Solution: Obtain a static IP address for the cluster itself. You cannot use DHCP to assign this address.

If your cluster nodes use DHCP to obtain noncluster IP addresses, at the command prompt, use the ipconfig command to verify that you have a valid primary IP address for the networks in question.

If you made any changes, restart the New Server Cluster Wizard. You must restart this wizard for the changes to be detected.

See also:Manage Networks

Possible Cause: You do not have the correct domain account and appropriate user rights.

Solution: Ensure that you are adhering to the following requirements:

To create a cluster, you must log on with a domain account that is a member of the local Administrators group on each node, and each node must be a member of that domain.
You can administer the Cluster service by using either a domain or a local account; however, other nodes can only be joined to a cluster by using a domain account.

If you make any account changes or changes to user rights, you must click Reanalyze in the New Server Cluster Wizard for the changes to be detected.

See also:Security

Possible Cause: Your system does not meet the minimum requirements for a cluster.

Solution: Verify that all requirements, including hardware, operating system software configuration, network and disk requirements are met.

Possible Cause: Your cluster nodes are in separate domains or are not in a domain.

Solution: All cluster nodes must be members or domain controllers of the same domain, and must use the same account for the Cluster service.

To check computer name and domain name for a node right-click My Computer, click Properties, and then click the Computer Name tab. Make sure each cluster node belongs to the same domain. Use the same account and password for service on each node. This account must have the following user rights:

Back up files and directories
Increase scheduling priority
Log on as a Service

If you made any changes, restart the New Server Cluster Wizard. You must restart this wizard for the changes to be detected.

Possible Cause: The cluster name is not online.

Solution: To make sure that the cluster name is online:

On the first node, open Cluster Administrator by clicking Start, then Run, then typing CluAdmin.
Open Groups, click Cluster Group and check the status of Cluster Name.
If the status is not Online, right-click Cluster Name, and then click Bring Online.
Restart the New Server Cluster Wizard. You must restart this wizard for the changes to be detected.

Note

If the Internet protocol (IP) address is not online, the cluster name will not be online. Make sure that the IP address is online before you try to bring the cluster name online.

Possible Cause: The Cluster service is not running on the first node.

Solution: Confirm that the Cluster service is running on the first node and that all resources within the cluster group are online before you try to install another node.

To confirm that the Cluster service is started:

On the first node, open Computer Management.
Double-click Services and Applications.
Click Services.
In the details pane, find the Cluster service entry and check its Status column. If the service is not started, right-click it, and then click Start.
Restart the New Server Cluster Wizard. You must restart this wizard for the changes to be detected.

If the Cluster service was not started, review the records in the event log to see why not.

During cluster creation, the Physical Disk resources are not automatically created for cluster disks.

Cause: Typically when you create a cluster, the Physical Disk resources are automatically created for cluster disks that use drive letters. However, if your cluster disks use mount points, Physical Disk resources will not be automatically created for those disks.

Solution: You must manually create disk resources for cluster disks that use mount points.

For more information, see Checklist: Installing a Physical Disk resource.

After setting up a new cluster, Cluster Administrator cannot connect to the cluster.

Cause: The new cluster was given a name that corresponds to an existing computer object in Active Directory directory service.

Solution: Determine if you want the Network Name resource for the cluster to use Kerberos authentication. If you do, perform the following steps:

Using the command line, take the Network Name resource for the cluster offline.
Using the command line, enable Kerberos authentication for the Network Name resource.
Inspect the permissions on the corresponding computer object in Active Directory and ensure that the Cluster service user account has access to that object.
Using the command line, bring the Network Name resource of the cluster online.

As an alternative, if you choose not to use Kerberos authentication for the Network Name resource, you must instead delete the corresponding computer object in Active Directory. Before you do so, first ensure that the computer object is not required elsewhere by your installation.

For more information, see article Q307532, "How to Troubleshoot the Cluster Service Account When It Modifies Objects," and article Q302389, "Description of the Properties of the Cluster Network Name Resource in Windows Server 2003," in the Microsoft Knowledge Base. (https://www.microsoft.com/) Also see Cluster resource and Finding directory information.

通过

Troubleshooting cluster node installations