Configuring Windows Failover Clustering

You can configure Windows Failover Clustering for Windows server nodes running as virtual machines within Oracle Linux Virtualization Manager.

The Windows Failover Clustering software requires a Windows domain controller running the Active Directory Domain Services (ADDS) and Domain Name System (DNS). The ADDS domain controller must run separate to the Windows Failover Clustering nodes.

Note:

Windows Server 2016 introduced Workgroup Cluster, an alternative configuration that doesn’t require a domain controller running ADDS; in a Workgroup Cluster, the nodes join a workgroup instead of a domain. A DNS server is, however, still required. A Workgroup Cluster can be less expensive and require fewer hardware resources than a traditional Failover Cluster without sacrificing high availability.

The procedure requires that the Windows Failover Clustering Node VMs are already created, are running the same Windows server version, are in the same time zone, and are joined to the same Windows domain. The Windows Failover Clustering tool must be installed on all the node VMs.

At least one shared LUN is required to attach to each of the cluster nodes. The LUN must not be configured as a Direct LUN disk.

KVM Host Configuration

Configure each KVM host using the following steps:

  1. On each KVM host, create /etc/multipath/conf.d/scsi3-reservation.conf with the following content:

    defaults {
       reservation_key file
    }
  2. Reload the multipathd.service:

    systemctl reload multipathd.service
  3. Confirm the reservation_key file option is enabled, by running the following command:

    multipath -t | grep reservation_key

Configuring the Engine

Configure the engine using the following steps:

  1. Enable disk error propagation for storage reliability:

    engine-config -s PropagateDiskErrors=true
  2. Restart the engine to apply changes:

    systemctl restart ovirt-engine.service
  3. Validate the engine configuration changes:

    engine-config -g PropagateDiskErrors

Windows Failover Cluster Node Configuration

  1. Configure the first virtual machine node in the cluster to add the direct LUN into the cluster.

    1. Edit the first virtual machine in the cluster and create a new disk by selecting the + button, then selecting the Create button.
    2. In the New Virtual Disk dialog, select the Direct LUN button:
      • Provide a name for the virtual disk.
      • In the Discover Targets section, enter the portal Address and select discover.
      • Login to the LUN and expand the Target Name section. Check the radio box for the LUN to use.
      • Check the Shareable and Enable SCSI-3 Reservation checkboxes.
    3. Select OK to confirm and close.

    Note:

    You must create the disk by editing the virtual machine and using the New Virtual Disk dialog. The required options that you need to enable are only available in this dialog. You can't enable these options if you add the disk by using the Storage Disks New dialog.

  2. Attach the disk to each of the remaining virtual machines in the cluster. For each virtual machine:
    1. Edit the virtual machine and create a new disk by selecting the + button, then selecting the Attach button.
    2. In the Attach Virtual Disk dialog, select the Direct LUN button:
      • Check the radio button for the LUN from the list, and select Ok.
      • Select Ok again to close the Edit Virtual Machine window and save the configuration.
      • Edit the virtual machine again, and edit the recently added Direct LUN.
      • Check the Enable SCSI-3 Reservation checkbox, and select Ok to close the Virtual Disk definition.
    3. Select OK to confirm and close.
  3. Use the Disk Management tool in one of the virtual machines, to bring the virtual disk online, and initialize it.

Validating Windows Failover Clustering Resources

Validate the Windows Failover Clustering resources before creating the Windows Failover Cluster.

You can perform validation by using either the Failover Cluster Manager tool, or by running a command in PowerShell on one of the Windows virtual machine guests:

  • Run the Failover Cluster Manager tool.
    1. Under the Management section, go to the Actions pane on the right side of the window and select Validate Configuration.
    2. On the Before You Begin page, select Next.
    3. On the Select Servers or a Cluster page in the Enter name field, enter the hostnames for each of the nodes in the cluster as a comma-separated list and select Next.
    4. On the Testing Options page, select Run all tests (recommended) and then select Next.
    5. On the Confirmation page, select Next to run the tests.
    6. Review the output on the Summary page.
  • Run PowerShell:
    1. Run the Test-Cluster command:
      Test-Cluster –Node node1, node2, node3

      Substitute node1, node2, and node3 with the resolvable hostnames of the nodes in the cluster.

    2. Review the output.

Create the Windows Failover Cluster

You can create the Windows Failover Cluster by using either the Failover Cluster Manager tool or by running a command in PowerShell on one of the Windows virtual machine guests:

  • Run the Failover Cluster Manager tool.
    1. Select the Create Cluster link under the Management section.
    2. On the Before You Begin page, select Next.
    3. On Select Servers page, in the Enter name field, enter the hostnames of each of the nodes in the cluster, separate the names by a comma or a semicolon. Select Next.
    4. On the Access Point for Administering the Cluster page, provide a name for the cluster in the Cluster Name field. Select Next.
    5. Review the final confirmation page, which provides details of the cluster. Create the cluster by selecting Next.
    6. Review the summary and select Finish.
  • Run PowerShell:
    1. Run the Test-Cluster command:
      New-Cluster -Name cluster1 -Node node1,node2,node3

      Substitute cluster1 with a suitable name for the cluster. Substitute node1, node2, and node3 with the resolvable hostnames of the nodes in the cluster.