Advanced Failover cluster
Add and remove a cluster Node
- To add a node to a cluster, right-click a cluster and click Add Node to start the Add Node Wizard
– Enter the server name
– The selected server must have the Failover Clustering feature already installed
- The server is added and the quorum settings are updated
- To remove a node from a cluster, right-click the node, point to More Actions, and click Evict
Shutdown and restart a cluster
- Shutting down a cluster stops all clustered roles and the cluster service on all nodes in the cluster
–This approach is preferable to stopping the cluster service separately on each node
- To shut down a cluster, right-click the cluster name in the FCM
–Point to More Actions and click Shut Down Cluster
- To restart a cluster that has been shut down, right-click the cluster and click Start Cluster
Copying Cluster Roles
- Use the Copy Cluster Roles Wizard to migrate clustered services and apps from one cluster to another
- The following roles can’t be migrated:
–Microsoft SQL Server
–Microsoft Exchange Server
–Volume Shadow Copy Service tasks
–Task Scheduler tasks
–Cluster-Aware Updating settings
- To copy cluster roles, run the FCM from a node in the target cluster
– Right-click the cluster, click More Actions, and click Copy Cluster Roles
- In the Copy Cluster Roles Wizard, specify the source cluster and roles you want to copy
– The source cluster can’t be the cluster form which you’re running the wizard)
Moving cluster resources
Core cluster resources:
–Quorum resource (the witness disk or share)
–IP address resource that provides the cluster IP address
–Network name resource that provides the cluster name
Steps:
–Right-click the cluster
–Point to More Actions
–Point to Move Core Cluster Resources
–Click Best Possible Node or Select Node (you can specify which server to transfer resources to)
Configure Advanced Quorum settings
To configure quorum settings:
–Right-click the cluster in the Failover Cluster Manager
–Point to More Actions
–Click Configure Cluster Quorum Settings to start the Configure Cluster Quorum Wizard
In the Select Quorum Configuration Option window, you have three choices: Use default configuration; select the quorum witness; advanced quorum configuration.
The default quorum configuration
The cluster will automatically select the best quorum type for you to simplify the deployment. This choice is based on the number of nodes and available storage. The logic is as follows:
- Odd Number of Nodes – use Node Majority
- Even Number of Nodes
- Available Cluster Disks – use Node & Disk Majority
- No Available Cluster Disk – use Node Majority
Selecting Quorum Witness
The “Select the quorum witness” option allows you to specify how you want the quorum witness to be configured
–The cluster determines the other quorum options
Three options for configuring the quorum witness:
–Configure a disk witness
–Configure a file share witness
–Do not configure a quorum witness
Advanced Quorum Configuration Options
The advanced quorum configuration option allows you to choose which nodes have a quorum vote in the cluster
–By default, all nodes have a quorum vote
–You can remove nodes from the voting process
–If you remove voting from all nodes, you must have a disk witness
Dynamic quorum
A new Windows Server 2012/R2 feature that assigns a cluster node vote dynamically
–In earlier versions of Windows Server the number of votes was static. Another new feature is the capability to change node votes dynamically to create a tie breaker for 50% node split.
Highly available share
Two main options for configuring highly available shares by using the File Server role
–Default option is “File Server for general use”: Supports standard Windows SMB shares and NFS shares and role services
–The “Scale-Out File Server for application data” option is used for applications that leave files open for extended periods. Scale-out file server – provides the reliability of failover clusters and the load distribution of NLB. Client connections are distributed equally among all cluster nodes.
Upgrading a Failover Cluster
Options when upgrading to a Windows Server 2012/R2 cluster:
- Add two nodes running Windows Server 2012/R2 to the existing cluster and transfer the roles and resources to the new nodes. Evict (remove) the older nodes once resources and roles have been transferred
- Perform an in-place upgrade
- Build a new cluster with nodes running Windows Server 2012/R2 and run the Copy Cluster Roles Wizard from one of the nodes in the new cluster
Creating AD-detached clusters
Active-Directory-detached cluster – allows deploying a failover cluster without needing Active Directory for network name management.
–Still creates the DNS entries for name resolution but computer accounts are not created in AD
All cluster nodes must be running Windows Server 2012 R2 and be joined to the same domain
Created with the following PowerShell cmdlet:
New-Cluster ClusterName -Node Server1, Server2 -AdministrativeAccessPoint DNS
clustered storage space
your cluster and storage configuration must meet the following requirement:
–All servers must be running at least Windows Server 2012
–You must have at least three disks with no volumes allocated on them
–Disk controllers must be SAS and must not have RAID enabled
–Use a Windows Server 2012-certified JBOD storage enclosure that has a SAS connection for each node
–Disks used for clustered storage pool must be dedicated to that pool
Two main methods for deploying a clustered storage pool:
- Create a storage space before you create the failover cluster, then add the storage to the failover cluster when running the Create Cluster Wizard
- Create the failover cluster, then create a clustered storage space in the Failover Cluster Manager
Steps to configure:
step 1. Install the Multipath I/O feature on all cluster nodes
Step 2. Verify that the shared disks are available in Disk management or File and storage Services.
Step 3. Create the storage spaces by using File and storage services. ( skip this step for method 2)
Step 4. Create a failover cluster
Step 5. Don’t do this step for method 1. Create storage spaces in the failover cluster manager instead of using File and Storage services. To do so, expand the storage node in the Failover cluster manager, click pools, and click New Storage Pool in the Action pane.
Backup and resotre
We use Windows Server Backup to backup a failover cluster configuration.
Step 1. Install the Windows Server Backup feature on at least one node
Step 2. Before backing up the cluster configuration: Verify the cluster node where doing the backup is active in the cluster, the cluster is running, and it has a quorum.
Step 3. Full server backup is recommended. For custom backup, it should include system state. Because witness disk also includes cluster configuration data, include it in the backup. Also backup data on clustered disks
If you need to restore the cluster configuration on a single node, you have two options:
- Authoritative restore – configuration contained in the backup is replicated to all cluster nodes after backup is finished. Use when you need to roll back the configuration to an earlier point.
- Nonauthoritative restore – restore a node from the system state backup and it rejoins the cluster. Current cluster configuration stored on the other cluster nodes is replicated to it.
Next part we will focus on VM high availability