共用方式為


In-Place Migration for a Two-Node Cluster

 

Applies To: Windows Server 2012

This topic provides an overview and steps for upgrading an existing failover cluster to Windows Server 2012 when you have only two servers - that is, for performing an in-place migration.

Important

Before you begin the migration, confirm that the clustered service or application that you want to migrate can be migrated by using the Migrate a Cluster Wizard, as described in Migration Paths for Migrating to a Failover Cluster Running Windows Server 2012, and note any preparation or follow-up steps that are required for the service type that is being migrated.

Note

For an alternative approach to failover cluster migration, see Migration Between Two Multi-Node Clusters.

Overview of an in-place migration for a two-node cluster

This migration uses the Migrate a Cluster Wizard, and it has four phases:

  1. Evict one node, install Windows Server 2012, and create a single-node failover cluster. For this phase, allow one existing server to continue running Windows Server 2008 R2 or Windows Server 2008 and the Cluster service while you begin the migration process. Evict the other server from the old cluster, and then perform a clean installation of Windows Server 2012 and the Failover Clustering feature on it. On that server, run all tests that the Validate a Configuration Wizard will run. The wizard will recognize that this is a single node without storage and limit the tests that it runs. Tests that require two nodes (for example, tests that compare the nodes or that simulate failover) will not run.

    Note that the tests that you run at this stage do not provide complete information about whether the storage will work in a cluster running Windows Server 2012. As described later in this section, you will run the Validate a Configuration Wizard later with all tests included.

    For steps to complete this phase, see Steps for evicting a node and creating a new single-node Windows Server 2012 failover cluster, later in this topic.

  2. Migrate clustered services and applications to the new single-node cluster. Run the Migrate a Cluster Wizard, but leave the migrated resources offline on the new cluster.

    For steps to complete this phase, see Steps for migrating clustered services and applications to the new cluster, later in this topic.

  3. Make existing data available to the new cluster, and bring the cluster online. Confirm that the settings for the migrated services and applications are correct. Next, take the migrated services and applications in the old cluster offline. If the new cluster will use new storage, copy the folders and data to appropriate LUNs or disks in the new storage, and make sure that those LUNs or disks are visible to the new cluster (and not visible to any other servers). If the new cluster will use the old storage, make the appropriate disks or LUNs accessible to the new cluster. Bring the services and applications in the new cluster online, and make sure that the resources are functioning and can access the storage.

    For steps to complete this phase, see Steps for making existing data available to the new cluster and bringing it online, later in this topic.

  4. Add the second node to the new cluster. Destroy the old cluster and, on that server, install Windows Server 2012 and the Failover Clustering feature. Connect that server to the networks and storage that are used by the new cluster. If the appropriate disks or LUNs are not already accessible to both servers, make them accessible. Run the Validate a Configuration Wizard, specifying both servers, and confirm that all tests pass. Finally, add the second server to the new cluster.

    For information about steps for this phase, see Steps for adding the second node to the new cluster, later in this topic.

Steps for evicting a node and creating a new single-node Windows Server 2012 failover cluster

You must complete the following steps to create a single-node Windows Server 2012 failover cluster:

Step 1: Evict one node from the old cluster, and perform a clean installation of Windows Server 2012

To begin, you must evict one node from the old cluster, and perform a clean installation of Windows Server 2012 on that node.

Before you evict a node from a cluster

  • For each clustered service or application that you plan to migrate, verify that there are no special requirements or procedures for removing or evicting a node from the cluster. You can evict a node from a clustered file server or a cluster with the Hyper-V role with no special preparation. However, you might need to uncluster some services or applications before you evict a node.

  • To prevent any loss of application data when the node is evicted, shut down all services and applications on the cluster before you evict the node.

To evict a node from a cluster

  1. From the Start screen, open Failover Cluster Manager.

  2. In the console tree, expand the cluster, expand Nodes, and then click the node that you want to evict to select it.

  3. Right-click the node, click More Actions, and then click Evict.

Step 2: Create a single-node cluster and install other needed software

For information about how to create a Windows Server 2012 failover cluster, see Create a Failover Cluster. To prepare to migrate a clustered service or application to the new failover cluster, make the following preparations.

Preparation

Before you create the failover cluster, prepare storage, and install all required services, applications, and server roles.

  1. Prepare storage:

    1. Make an appropriate number of LUNs or disks accessible to the server, and do not make those LUNs or disks accessible to any other servers. If the new cluster will use old storage, for testing purposes, you can limit the number of LUNs or disks to one or two. If the new cluster will use new storage, make as many disks or LUNs accessible to the new server as you think the cluster will need.

      Note

      We recommend that you keep a small disk or LUN available (unused by clustered services and applications) throughout the life of the cluster, so that you can always run storage validation tests without taking your services and applications offline.

    2. On the server, open Computer Management from the Start screen, and then click Disk Management in the console tree. In Disk Management, confirm that the intended cluster disks are visible.

    3. Check the format of any exposed volume or LUN. We recommend that you use NTFS for the format. (For a disk witness, you must use NTFS.)

    4. If you are using new storage and your disk configuration uses mount points, review Cluster Migrations Involving New Storage: Mount Points to identify any additional steps you will need to perform.

  2. Install services, applications, and server roles:

    • After you install the Failover Clustering feature on the server, install any needed services, applications, and server roles. For example, if you plan to migrate clustered Windows Internet Name Service (WINS) to the new cluster, install the WINS Server feature by using Server Manager.

    • If you plan to migrate highly available virtual machines, add the Hyper-V role and install the latest Hyper-V integration components. You also must merge or discard all shadow copies on the volumes that contain the virtual machines. For step-by-step instructions for migrating highly available virtual machines, see Migration of Highly Available Virtual Machines Using the Migrate a Cluster Wizard.

    • If you are migrating a Generic Application, Generic Script, or Generic Service resource, you must confirm that any associated application is compatible with Windows Server 2012. You also must confirm that any associated service exists in Windows Server 2012 and has the same name that it had in the old cluster. Test the application or service (separately, not as part of a cluster) to confirm that it runs as expected.

After you create the failover cluster

After you create the cluster, ensure that your firewall is configured appropriately. For example, if you are using Windows Firewall, and you will be sharing folders and files, use your preferred Windows Firewall interface to allow the exception for Remote Volume Management.

Steps for migrating clustered services and applications to the new cluster

Use the following instructions to migrate clustered services and applications from your old one-node cluster to your new one-node cluster. The Migrate a Cluster Wizard leaves most of the migrated resources offline so that you can perform additional steps before you bring them online.

Note

To migrate a clustered service or application by using the Migrate a Cluster Wizard, you must be a local administrator on the destination failover cluster and on the cluster or cluster node from which you are migrating.

To migrate clustered services and applications from the old cluster to the new cluster

  1. If you want to migrate to new storage, before you run the Migrate a Cluster Wizard, ensure that the storage is available to the new cluster – that is, that the volumes have been added to the new cluster and that they are online.

  2. From the Start screen or Server Manager (Tools), open Failover Cluster Manager.

  3. In the console tree, if the cluster that you created is not displayed, right-click Failover Cluster Manager, click Connect to Cluster, and then select the cluster that you want to configure.

  4. In the console tree, expand the cluster that you created to see the items underneath it.

  5. If the clustered server is connected to a network that is not to be used for cluster communications (for example, a network intended only for iSCSI), then under Networks, right-click that network, click Properties, and then click Do not allow cluster network communication on this network. Click OK.

  6. In the console tree, select the cluster.

  7. Under Configure, click Migrate services and applications.

  8. Read the first page of the Migrate a Cluster Wizard, and then click Next.

  9. Specify the name or IP address of the cluster or cluster node from which you want to migrate services and applications, and then click Next.

  10. The Select Services and Applications page lists the clustered services and applications that can be migrated from the old cluster. The list does not contain any service or application that is not eligible for migration. Click View Report for details. Then select each service and application that you want to migrate to the new cluster, and click Next.

    Important

    We recommend that you read the report, which explains whether each resource is eligible for migration. (The wizard also provides a report after it finishes, which describes any additional steps that might be needed before you bring the migrated resource groups online.)

    If storage is available on the new cluster, the Specify Storage for Migration page appears, giving you the option to migrate to new storage. If storage is not available on the new cluster, the wizard automatically retains existing storage settings and does not display the page.

  11. If you want to use new storage for a service or application:

    1. On the Specify Storage for Migration page, select the cluster disk that you want to migrate to new storage, and then click Select Storage.

    2. In the Select Storage for Resource Group dialog box, under Available Storage in New Cluster, select the cluster disk that you want the service or application to use in the new cluster, and then click OK.

    3. Repeat these steps for each cluster disk that you want to migrate to new storage. Then click Next.

    Important

    The Migrate a Cluster Wizard does not move existing folders and data to the new storage. You must copy the folders and data manually.

  12. Follow the instructions in the wizard to perform the migration. From the Summary page, we recommend that you read the Cluster Migration Report, which contains important information about any additional steps that you might need to complete before you bring the migrated services and applications online. For example, if you have not already installed needed applications on the new cluster node, you might need to install them.

When the wizard completes, most migrated resources will be offline. Leave them offline at this stage.

Steps for making existing data available to the new cluster and bringing it online

Use the following procedure to make existing data available to the new cluster and bring it online.

To make existing data available to the new cluster and bring it online

  1. Confirm that the settings for the migrated services and applications appear correct.

  2. Prepare for clients to experience downtime, probably briefly.

  3. On the old cluster, take each clustered service or application that you migrated offline.

  4. Complete the transition for the storage:

    • If the new cluster will use old storage, follow your plan for making LUNs or disks inaccessible to the old cluster and accessible to the new cluster.

    • If the new cluster will use new storage, copy the appropriate folders and data to the storage. As needed for disk access on the old cluster, bring individual disk resources online on that cluster. (Keep other resources offline to ensure that clients cannot change data on the disks in storage.) Then, on the new cluster node, use Disk Management to confirm that the appropriate LUNs or disks are visible to the new cluster and not visible to any other servers.

  5. If the new cluster uses mount points, adjust the mount points as needed, and make each disk resource that uses a mount point dependent on the resource of the disk that hosts the mount point. For more information about mount points, see Cluster Migrations Involving New Storage: Mount Points.

  6. Bring the migrated services or applications online on the new cluster.

Steps for adding the second node to the new cluster

Use the following instructions to prepare the second node and then add it to the new cluster. As part of this process, you will run validation tests that include both servers.

To add the second node to the new cluster

  1. On the new cluster, confirm that the migrated services or applications are functioning and that clients can connect to them.

  2. On the old cluster (the server that is running Windows Server 2008 R2 or Windows Server 2008), delete the migrated services and applications, and then destroy the old cluster:

    1. From the Start screen, open Failover Cluster Manager.

    2. Remove services and applications that were migrated. In Failover Cluster Manager, expand the cluster, and expand Services and applications. To delete a service or application, right-click the item, and click Delete.

    3. To destroy the cluster, right-click the cluster, click More Actions, and then click Destroy Cluster.

  3. On the same server, perform a clean installation of Windows Server 2012.

  4. Add the Failover Clustering feature in the same way that you added it to the other server, and install any needed services, applications, and server roles.

  5. Connect the newly installed server to the same networks and storage that the existing failover cluster node is connected to.

  6. Identify the disks or LUNs that are exposed to the new one-node failover cluster, and expose them to the newly installed server also.

    We recommend that you keep a small disk or LUN accessible to both nodes, and unused by clustered services and applications, throughout the life of the cluster. With this LUN, you can always run storage validation tests without taking your services and applications offline.

  7. On either server running Windows Server 2012, open Failover Cluster Manager from the Start screen.

  8. Confirm that Failover Cluster Manager is selected, and then, in the center pane, under Management, click Validate a Configuration.

    Follow the instructions in the wizard, but this time, be sure to specify both servers (not just the existing cluster name) and specify that you want to run all tests. Then, run the tests. Because two nodes are now being tested, a more complete set of tests runs, which takes longer than testing one node.

    Important

    If any clustered service or application is using a disk when you start the wizard, the wizard asks whether to take that clustered service or application offline for testing. If you choose to take a clustered service or application offline, it remains offline until the tests finish.

  9. The Summary page appears after the tests run. To view Help topics to help you interpret the results, click More about cluster validation tests.

  10. While still on the Summary page, click View Report and read the test results.

    To view the results of the tests after you close the wizard, see

    <SystemRoot>\Cluster\Reports\Validation Report <date and time>.mht

    where <SystemRoot> is the folder in which the operating system is installed (for example, C:\Windows\).

  11. As necessary, make changes in the configuration and rerun the tests.

    For more information about failover cluster validation tests, see Validate Hardware for a Failover Cluster.

  12. If the new cluster is not displayed, in the console tree, right-click Failover Cluster Manager, click Connect to a Cluster, and then select the new cluster.

  13. In the console tree, select the one-node cluster, and then in the Actions pane, click Add Node.

  14. Follow the instructions in the wizard to specify the server that you want to add to the cluster. On the Summary page, click View Report to review the tasks that the wizard performed.

  15. On the Summary page, click View Report if you want to review the tasks that the wizard performed. Or view the report after the wizard closes in the <SystemRoot>\Cluster\Reports\ folder.

    Note

    After you close the wizard, in the center pane, you might see a warning about “Node Majority.” You will correct this issue in the next few steps.

  16. In the console tree, expand Storage. Check to see if all the disks that you want to make available to the new cluster are shown, either in one of the clustered services or applications or in Available Storage.

    In most cases, you need at least one disk in Available Storage for the next step (specifying a witness disk). If you need to add a disk, in the Actions pane, click Add Disk and follow the steps in the wizard.

    Before you can add a disk to storage, it must be accessible from both nodes in the cluster. To be used for a witness disk, a disk can be a relatively small, but must be at least 512 MB.

  17. In the console tree, right-click the new cluster, click More Actions, and then click Configure Cluster Quorum Settings.

  18. Follow the instructions in the wizard to select the most appropriate quorum setting for your needs. In most cases, this is the Node Majority quorum configuration, which requires that you specify an appropriate disk (from Available Storage) for the witness disk. For more information about quorum settings in Windows Server 2012, see Configure and Manage the Quorum in a Windows Server 2012 Failover Cluster.

  19. Expand Services and Applications, and then click a migrated service or application that you want to test.

  20. To perform a basic test of failover for the migrated service or application, under Actions (on the right), click Move this service or application to another node, and then click an available choice of node. When prompted, confirm your choice.

    You can observe the status changes in the center pane of Failover Cluster Manager as the clustered service or application is moved. If there are any issues with failover, review the following:

    • View events in Failover Cluster Manager. To do this, in the console tree, right-click Cluster Events, and then click Query. In the Cluster Events Filter dialog box, select the criteria for the events that you want to display, or, to return to the default criteria, click the Reset button. Click OK. To sort events, click a heading, for example, Level or Date and Time.

    • Confirm that necessary services, applications, or server roles are installed on all nodes. Confirm that services or applications are compatible with Windows Server 2012 and run as expected.

    • Review migrated resource settings and dependencies. If you are using new storage that includes disks that use mount points, see Cluster Migrations Involving New Storage: Mount Points for more information.

    • If you migrated one or more Network Name resources with the Kerberos protocol enabled, confirm that the following permissions change was made in Active Directory Users and Computers on a domain controller. In the computer accounts (computer objects) of your Kerberos protocol-enabled Network Name resources, Full Control must be assigned to the computer account for the failover cluster.

Migration Between Two Multi-Node Clusters

Migration of Highly Available Virtual Machines Using the Migrate a Cluster Wizard

Cluster Migrations Involving New Storage: Mount Points

Migration Paths for Migrating to a Failover Cluster Running Windows Server 2012