vSAN NIC Teaming Behavior with Active/Passive Uplinks and Partial Network Connectivity Failures
Article ID: 444047
Updated On:
Products
Issue/Introduction
Questions may arise regarding NIC teaming and load balancing behavior for vSAN networks configured with Active/Passive uplinks.
Administrators may observe scenarios where:
- The physical NIC link remains up.
- Standard or small MTU traffic is successful.
- Jumbo frame traffic fails.
- vSAN communication experiences intermittent or degraded connectivity.
- Traffic does not fail over to the standby uplink despite communication issues.
There may also be questions regarding whether NIC teaming load balancing policies can detect and recover from such conditions.
Purpose:
This article explains the NIC teaming configuration (Active/Passive) for vSAN networks and clarifies failover behavior when partial network connectivity issues occur.
Environment
- VMware vSAN 8.x
- VMware vSAN 9.x
Resolution
For vSAN networks, VMware recommends using multiple physical NICs for redundancy and bandwidth utilization.
When using NIC teaming for vSAN traffic, the recommended load balancing policy is Route Based on Originating Virtual Port in an Active/Passive configuration.
With this policy, each VMkernel adapter is pinned to a specific uplink based on the virtual port assignment. Traffic remains on the selected uplink unless a failover event occurs. This provides a simple and predictable design and is the recommended configuration for vSAN traffic on both Standard and Distributed Switches.
In an Active/Passive configuration:
- Only the Active uplink carries vSAN traffic during normal operation.
- The Passive uplink remains on standby and is used only when the Active uplink is declared failed.
- This configuration provides redundancy but does not provide load balancing across both uplinks.
Regarding the reported network condition, failover and load balancing depend on the ESXi host detecting a network failure. If the physical NIC link remains up and the switch port remains operational, ESXi will continue to consider the uplink healthy.
In the described scenario:
- The NIC link remains up.
- Small MTU traffic succeeds.
- Jumbo frame traffic fails.
Since the uplink is not considered down by ESXi, NIC teaming will not automatically move traffic to the standby uplink. From the hypervisor perspective, the network path is still available because link state monitoring reports the uplink as healthy.
As a result:
- Automatic failover is not expected to occur.
- vSAN traffic using jumbo frames may continue to experience communication issues.
- Load balancing policies do not detect or correct MTU mismatches or partial network failures.
- The underlying network issue must be corrected at the physical network layer.
To detect failures beyond simple link state loss, VMware supports beacon probing. However, beacon probing is generally not recommended for vSAN networks. For vSAN environments, proper switch configuration, consistent MTU settings across the entire network path, and network monitoring are the recommended approaches.
For this environment, the following recommendations apply:
- Verify MTU consistency across:
- ESXi VMkernel adapters
- Physical switch ports
- Switch trunks
- Intermediate network devices
- If the requirement is only redundancy, but failover will occur only when the uplink is detected as failed.
- Resolve the underlying network issue causing jumbo frame traffic to fail, as NIC teaming and vSAN load balancing mechanisms are not designed to detect partial connectivity failures where the physical link remains operational.
Additional Information
Feedback
