Changing the network teaming failback delay fails in ESXi

Products

• VMware vSphere ESXi

Issue/Introduction

• Changing the failback delay fails
• Setting the failback delay to any other value does not work.

Environment

VMware vSphere ESXi  6.5, 6.7, 7.0 and 8.0

Cause

• When teaming policy is in place, connections are switched to another uplink when an uplink goes down.
• Connection is restored once the uplink is up again. When the primary link is flapping, this may result in packet loss due to frequent failback operation.
• The TeamPolicyUpDelay setting is set to 100ms by default under Configuration > Advanced Settings > Net > Net.teampolicyupdelay
• Net/TeamPolicyUpDelay setting is added to change the threshold in milliseconds (ms). This setting delays the up notification for the primary interface therefore prolong the time the active uplink stays with the standby interface. 
• When an up event is received, a timer is triggered to notify this. If a down event occurs when the trigger is pending, then the up notification is cancelled.

Resolution

This is resolved in ESXi 6.7 P02

While this issue is resolved in ESXi 6.7 P02, it is important to note that the configuration option causing the issue remains valid in ESXi 7.0 and ESXi 8.0.

The resolution addresses the behavior under certain conditions rather than the Configuration setting itself. Therefore ESXi 7.0 and ESXi 8.0 are included in impacted environment list to ensure clarity. 

 

 

Kindly refer to NIC Teaming Policy Documents below for further details

https://techdocs.broadcom.com/us/en/vmware-cis/vsphere/vsphere/7-0/vsphere-networking-7-0/networking-policies/teaming-and-failover-policy.html

NSX-T Perspective:

Please note that the information provided in this section is NSX-T perspective.
If NSX is running on this host and using NVDS, You can also apply these changes using below command from ESXi root login. 

        nsxcli -c set vswitch runtime TeamPolicyUpDelay "value"

Example:
        nsxcli -c set vswitch runtime TeamPolicyUpDelay 600000. 

However, this will take effect at runtime and will not persist after the host rebooted the value will reset to its default. 

To persist the change, please follow the steps below:

Option 1

1] Make changes from ESXi via vSphere console -pick host > Configure > Advanced system Settings > Net.TeamPolicyUpDelay > edit (default is 100) And modify/set the value you want to and save by clicking ok. 

2] Reboot host for changes to take effect 

3] Once node is back online, ssh to host with root and verify the setting using below command :

        nsxcli -c get vswitch runtime

Option 2

1] Make changes on ESXi via command line with below command 

        esxcli system settings advanced set -o /Net/TeamPolicyUpDelay --int-value xxx

Example: esxcli system settings advanced set -o /Net/TeamPolicyUpDelay --int-value 300000

2] Reboot host for changes to take effect 

3] ssh to host with root and verify the setting using below command :

      nsxcli -c get vswitch runtime

Note: N-VDS is not supported on NSX 4.X version. With the release of VMware NSX 4.x N-VDS is not supported for ESXi hosts anymore.
However, if you are still using NVDS or have already migrated to VDS, this change is being made directly to ESXI and will apply to both.

Additional Information

Customer may also consider disabling the failback feature completely to allow manual switching to the primary interface per redundancy design of the environment.