Host is down on Windows Failover Cluster

[ejdgr8]

Hello, I have 2 nodes of Hyper-V and windows failover cluster enabled. when I check on the nodes, the other node is down.

When I checked on the network I noticed that the vSwitch have a Unidentified status.

and this is the status on the Windows Failover Cluster on the network configuration

Steps Taken:
1. disable and enable the vSwitch/NIC
2. Manually input the IP Address for the vSwitch
3. Restart the host

Thank you!

 

[ChatGPT]

Hi ejdgr8,
From your screenshots, this does not look like a simple “GUI glitch.” It looks like hyperv02 lost cluster network communication, so the cluster marked that node Down. Your screenshots show:

• hyperv02 = Down
• Cluster Network 2 = Down
• both vEthernet adapters on hyperv02 = Unavailable / Unidentified
• the affected network is marked Cluster Only

That points first to a network path issue on hyperv02: physical NIC, team/SET, vSwitch binding, VLAN, cable/switch port, or host vNIC config.
Also, an “Unidentified network” label by itself is not always the actual problem on Hyper‑V host vNICs. The real problem is that Failover Clustering sees those interfaces as Unavailable/Down and removes the node when heartbeat traffic is lost. Microsoft’s cluster guidance says node removal commonly happens when cluster communication is lost, and recommends validating network configuration, checking latency, drivers/firmware, firewall/ports, and switch/NIC issues. Microsoft also recommends that cluster traffic networks be isolated properly, and that each dedicated physical or virtual adapter use a unique subnet per traffic type.

Most likely causes​

Check these first on hyperv02:

1. Physical NIC or uplink failure
   • cable/SFP
   • switch port down/err-disabled
   • FC is unrelated here; this is Ethernet cluster traffic
2. NIC team / SET / vSwitch issue
   • one team member failed
   • uplink adapter removed from vSwitch
   • wrong load balancing / broken team state
3. Wrong VLAN or IP settings on host vNICs
   • replication-vsw / uplink-vsw may be on the wrong VLAN
   • duplicate IP
   • subnet mask mismatch
   • gateway on non-management NIC
4. Cluster heartbeat network broken
   • cluster uses UDP/TCP 3343
   • if the network is reachable at L2/L3 but blocked by firewall/filtering, the node can still be evicted. Microsoft specifically calls out port 3343 plus related cluster ports and recommends cluster validation/network checks.
5. Outdated NIC drivers / firmware / vendor utilities
   • Microsoft explicitly notes old NIC drivers/firmware or adapter issues as common causes of lost heartbeats / Event ID 1135.

---

What to do now​

1) Check whether hyperv02 is actually alive​

On hyperv02 console / iLO / iDRAC:

Get-Service ClusSvc
Get-NetAdapter
Get-VMNetworkAdapter -ManagementOS
Get-VMSwitch
ipconfig /all

What you want to see:

• physical NICs = Up
• team/SET = Up
• vEthernet adapters = Up
• correct IPs on the cluster/replication/uplink networks

If the physical NIC is down, fix that first.

---

2) Verify the cluster network from hyperv01​

From hyperv01:

Get-ClusterNode
Get-ClusterNetwork | ft Name, State, Role, Address, Metric, AutoMetric
Get-ClusterNetworkInterface | ft Node, Network, IPAddress, State

If hyperv02 interfaces on Cluster Network 2 show Unavailable, that confirms the cluster cannot talk to that node on that network.

---

3) Test basic connectivity between nodes on each cluster-related subnet​

From hyperv01, ping the IPs of hyperv02 on:

• management
• cluster
• replication/live migration if used

Example:

ping <hyperv02-management-ip>
ping <hyperv02-cluster-ip>
ping <hyperv02-replication-ip>
Test-NetConnection <hyperv02-cluster-ip> -Port 3343

If management works but cluster-only network fails, the problem is isolated to that specific vNIC/VLAN/switch path.

---

4) Check the Hyper-V vSwitch and management OS vNIC config on hyperv02​

If you are using a converged design, verify the virtual NICs attached to the vSwitch:

Get-VMSwitch
Get-VMNetworkAdapter -ManagementOS | ft Name, SwitchName, Status, MacAddress
Get-VMNetworkAdapterVlan -ManagementOS

If a VLAN tag is wrong or missing on replication-vsw or uplink-vsw, Failover Cluster will see the network as down.
Microsoft’s Hyper‑V cluster network guidance says cluster, live migration, management, and replica traffic should be isolated correctly, and dedicated adapters should be on unique subnets. It also notes that if VMM is managing the cluster, networking changes should be done consistently through VMM.

---

5) Check NIC Teaming / SET​

If using LBFO:

Get-NetLbfoTeam
Get-NetLbfoTeamMember
Get-NetAdapter

If using SET:

Get-VMSwitchTeam
Get-NetAdapter

Microsoft recommends for clustered Hyper‑V:

• Switch Independent
• Hyper-V Port load balancing
  for NIC Teaming/LBFO scenarios.

If one uplink is failed or detached, recreate/fix the team before touching cluster settings.

---

6) Check Event Viewer on both nodes​

Look for:

• FailoverClustering
• Hyper-V-VmSwitch
• NIC vendor driver logs
• System

Especially:

• Event ID 1135
• link up/down
• miniport reset
• team member failure
• VLAN mismatch / adapter disconnect

Microsoft’s Event 1135 troubleshooting guidance is directly relevant here: when a node is removed from cluster membership, review cluster networks, run validation, check firewall/ports, latency, packet loss, NIC drivers, and switch issues.

---

7) Run Cluster Validation — especially Network tests​

From Failover Cluster Manager:

• Validate Configuration
• run at least Network tests

Or:
Test-Cluster -Include "Network","System Configuration"
This is one of Microsoft’s first recommended steps for node communication issues.

---

Important design note​

Based on your screenshots, you have:

• Cluster Network 1 = Cluster and Client
• Cluster Network 2 = Cluster Only

That is valid. Microsoft’s recommended roles are:

• Management = cluster + client
• Cluster = cluster only
• Live migration = cluster only
• Storage = no cluster communication
• Replica = usually cluster + client, depending on design.

So the role itself is probably not the issue. The issue is more likely that Cluster Network 2 on hyperv02 is no longer reachable.

---

I would troubleshoot in this order​

1. hyperv02 physical NIC/link state
2. Switch port / VLAN config
3. Team/SET health
4. vSwitch bindings
5. vNIC VLAN/IP config
6. Event 1135 / NIC reset logs
7. Cluster validation

---

Likely fix​

In real-world cases like this, the fix is often one of these:

• reconnect/fix failed NIC or switch port
• correct VLAN tagging on host vNIC
• repair/recreate NIC team or SET
• update NIC drivers/firmware
• remove bad stale vNIC and recreate it
• re-run cluster validation and bring node back

---

Quick commands to paste on hyperv02​

Get-NetAdapter
Get-VMSwitch
Get-VMNetworkAdapter -ManagementOS
Get-VMNetworkAdapterVlan -ManagementOS
ipconfig /all
Get-Service ClusSvc

Quick commands on hyperv01​

Get-ClusterNode
Get-ClusterNetwork | ft Name, State, Role, Address
Get-ClusterNetworkInterface | ft Node, Network, IPAddress, State
Test-NetConnection <hyperv02-cluster-ip> -Port 3343

If you want, send me the output of those commands and I’ll tell you exactly which adapter/network is broken and whether it’s VLAN, team, vSwitch, or cluster heartbeat.