bunny-lab/docs
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
b9aeaabbfbd508f125603ba64a088b01c37b5749
docs/Windows Server/Deploying Failover Cluster Host.md
Nicole Rappe b9aeaabbfb Initial Commit 
Bringing Documentation into Gitea
2023-12-21 01:15:09 -07:00

9.8 KiB
Raw Blame History

Deploy Additional Failover Cluster Node

Deploying a Windows Server Node into the Hyper-V Failover Cluster is an essential part of rebuilding and expanding the backbone of my homelab. The documentation below goes over the process of setting up a bare-metal host from scratch and integrating it into the Hyper-V Failover Cluster.

!!! note "Prerequisites" This document assumes you are have installed and are running a bare-metal Hewlett-Packard Enterprise server running iLO (Integrated Lights Out) with the latest build of Windows Server 2022 Datacenter (Desktop Experience). Windows will prompt you that your build is expired if it is too old.

Download the newest build ISO of Windows Server 2022 at the [Microsoft Evaluation Center](https://go.microsoft.com/fwlink/p/?linkid=2195686&clcid=0x409&culture=en-us&country=us)

!!! info "Assumption that Cluster Already Exists" This document also assumes that you are adding an additional server node to an existing Hyper-V Failover Cluster. This document does not outline the exact process of setting up a Hyper-V Failover Cluster from-scratch, setting up a domain, DNS server, etc. Those are assumed to already exist in the environment.

Preparation

Enable Remote Desktop

The first thing you will want to do is get remote access via Remote Desktop. This will enable higher resolution, faster response times with the GUI, and the ability to transfer files to and from the server more easily.

  • Connect to the server via the iLO Remote Console

  • Login using your Administrator credentials you created during the operating system installation

  • Open Server Manager

    • Navigate to "Local Server"
    • Under "Remote Management"
    • Click on "Disabled"
    • Un-check: "Allow Remote Connections to this Computer" !!! warning "Disable NLA (Network Level Authentication)" Ensure that "Allow Connections only from computers running Remote Desktop with Network Level Authentication" is un-checked. This is important because if you are running a Hyper-V Failover Cluster, if the domain controller(s) are not running, you may be effectively locked out from using Remote Desktop to access the failover cluster's nodes, forcing you to use iLO or a physical console into the server to log in and bootstrap the cluster's Guest VMs online.

    This step can be disregarded if the domain controller(s) exist outside of the Hyper-V Failover Cluster.

  • Locate the (current) DHCP-issued IP address of the server for Remote Desktop

    • You will want to use Remote Desktop for the next stage of deployment to transfer an ISO file to the server
    • Log into the server with Remote Desktop using the Administrator credentials you created when initially installing the operating system
      • You can use ipconfig /all to locate the current DHCP-issued IP address

Provision Server Role & Domain Join

You will want to rename the computer so it has the correct naming scheme before installing any server roles or domain joining it. The general naming convention is MOON-NODE-<0#>. Use a domain administrator credential for the join command when prompted. Restart the computer to finalize the changes. Increment the hostname number based on the existing servers in the cluster / homelab.

# Rename the server
Rename-Computer MOON-NODE-01

# Domain-join the server
Add-Computer MOONGATE.local

# Install Hyper-V server role
Install-WindowsFeature -Name Hyper-V -IncludeManagementTools

# Install the Failover Clustering feature
Install-WindowsFeature -Name Failover-Clustering -IncludeManagementTools

# Restart the server to apply all pending configurations
Restart-Computer

Failover Cluster Configuration

Configure Cluster SET Networking

You will need to start off by configuring a Switch Embedded Teaming (SET) team. This is the backbone that the server will use for all Guest VM traffic as well as remote-desktop access to the server node itself. You will need to rename the network adapters to make management easier.

  • Navigate to "Network Connections" then "Change Adapter Options"
    • Rename the network adapters with simpler names. e.g. (Embedded LOM 1 Port 1 becomes Port_1)
    • For the sake of demonstration, assume there are 4 NICs (Port_1, Port_2, Port_3, and Port_4)

!!! warning "10GbE Network Adapters" Be sure to leave the dual 10GbE network adapters out of the renaming work. They will be used later with the iSCSI Initiator.

# Switch Embedded Teaming (SET) team
New-VMSwitch -Name Cluster_SET -NetAdapterName Port_1, Port_2, Port_3, Port_4 -EnableEmbeddedTeaming $true

Configure Static IP Address

You may be booted out of the Remote Desktop session at this time due to how the network team changed the configuration. Leverage iLO to remotely access the server again to configure a static IP address on the new vEthernet (Cluster_SET) NIC using the following configuration. While in the NIC Properties, disable IPv6.

IP ADDRESS SUBNET MASK GATEWAY PRIMARY DNS SECONDARY DNS
192.168.3.5 255.255.255.0 192.168.3.1 192.168.3.10 192.168.3.11

Configure Static IP Addresses for 10GbE Networking

You will now want to set up the network adapters for the 10GbE iSCSI back-end. Configure both of the Intel(R) Ethernet Controller x540-AT2 10GbE NICs and change their IP addresses to match the table below. Rename the NICs to match a NIC1 and NIC2 naming scheme. Also disable IPv6. !!! warning Make sure that you test that each interface can ping their respective iSCSI target by performing a ping using the IP address in the "ISCSI PING IP" column of the table. If it fails to successfully ping, swap the IP addresses of the 10GbE NICs until it succeeds.

IP Address Subnet Mask Gateway Primary DNS Secondary DNS iSCSI Ping IP
192.168.102.200 255.255.255.0 192.168.3.10 192.168.3.11 192.168.102.100
192.168.104.200 255.255.255.0 192.168.3.10 192.168.3.11 192.168.104.100

Configure iSCSI Initiator to Connect to TrueNAS Core Server

At this point, now that we have verified that the 10GbE NICs can ping their respective iSCSI target server IP addresses, we can add them to the iSCSI Initiator in Server Manager which will allow us to mount the cluster storage for the Hyper-V Failover Cluster.

  • Open Server Manager
    • Click on the "Tools" dropdown menu
    • Click on "iSCSI Initiator"
      • You will be prompted to start the Microsoft iSCSI service. Click on "Yes" to proceed.
    • Click on the "Discovery" tab
    • Click the "Discover Portal" button
      • Enter the IP addresses of the "iSCSI Ping IP(s)" from the previous section. Leave the port as "3260".
    • Navigate to the TrueNAS Core server and add the "Initiator Name" seen on the "Configuration" tab to the Sharing > iSCSI > Initiator Groups > "Hyper-V Failover Cluster Hosts"
      • Example Initiator Name: iqn.1991-05.com.microsoft:moon-node-01.moongate.local
      • This is not explicitly documented at this time and is different from lab to lab in regards to the iSCSI implementation
    • Click the "Targets" tab to go back to the main page
      • Click the "Refresh" button to display available iSCSI Targets
      • Click on the first iSCSI Target failover-cluster-storage then click the "Connect" button
        • Check the "Enable Multi-Path" checkbox
        • Click the "Advanced" button
        • Click the "OK" button
      • Repeat the connection process seen above for all remaining iSCSI Targets
      • Close out of the iSCSI Initiator window
    • Navigate to "Disk Management" to bring the iSCSI drives "Online"

Initialize and Join to Existing Failover-Cluster

Validate Server is Ready to Join Cluster

Now it is time to set up the Failover Cluster itself so we can join the server to the existing cluster.

  • Open Server Manager
    • Click on the "Tools" dropdown menu
    • Click on "Failover Cluster Manager"
    • Click the "Validate Configuration" button in the middle of the window that appears
      • Click "Next"
      • Enter Server Name: MOON-NODE-01.moongate.local
      • Click the "Add" button, then "Next"
      • Ensure "Run All Tests (Recommended)" is selected, then click "Next", then click "Next" to start.

Join Server to Failover Cluster

  • On the left-hand side, right-click on the "Failover Cluster Manager" in the tree
    • Click on "Connect to Cluster"
    • Enter MOON-CLUSTER.moongate.local
    • Click "OK"
  • Expand "MOON-CLUSTER.moongate.local" on the left-hand tree
    • Right-click on "Nodes"
    • Click "Add Node..."
      • Click "Next"
      • Enter Server Name: MOON-NODE-01.moongate.local
      • Click the "Add" button, then "Next"
      • Ensure that "Run Configuration Validation Tests" radio box is checked, then click "Next"
      • Validate that the node was successfully added to the Hyper-V Failover Cluster

Cleanup & Final Touches

Activate Windows Server

You will need to change the edition from "Windows Server 2022 Datacenter Evaluation" to "Windows Server 2022 Datacenter". This will ensure that the server does not randomly reboot itself. If you have a license, you can install it now. Otherwise, you can force-activate using the Changing Windows Edition documentation.

Run Windows Updates

Ensure that you run all available Windows Updates before delegating guest VM roles to the new server in the failover cluster. This ensures you are up-to-date before you become reliant on the server for production operations.