Description

1. The user needs to modify the physical network port corresponding to the VXLAN network or the IP address of the physical network port corresponding to the VXLAN network due to network planning changes.
2. If a conflict with the VXLAN default interface due to the deployment of applications such as containers, the interface used by the VXLAN network needs to be modified.

Precautions

1. Modifying the physical interface, IP address, or interface corresponding to the VXLAN network will cause the VXLAN network to be temporarily interrupted. Please modify the configuration during the downtime window.
2. The overlay network interface requires a Gigabit (or 10 Gigabit) network interface and a Gigabit (or 10 Gigabit) switch for interconnection. (When only two nodes exist, the overlay network interface can be directly connected). To improve the interface bandwidth and redundancy of the Overlay Network Interface, it is recommended to configure the aggregation interface as the Overlay Network Interface. In this case, the peer switch must be configured with the corresponding aggregation mode.
3. To obtain high network forwarding performance, it is recommended to set the management communication port and Overlay Network Interface (VXLAN) to different interfaces.
4. Direct aggregation of two interfaces with different roles is not supported. You can cancel the role of one interface before aggregation. Example: Cannot directly aggregate eth0 (management interface) and eth1 (VXLAN interface). The VXLAN interface can be adjusted to eth2. At this time, eth0 (management interface) and eth1 (no role) can be aggregated.

Steps

1. Using an aggregated interface is recommended: Navigate to Nodes > Physical Interfaces, click Add multiple aggregated interfaces, and select two network interfaces—for example, eth2 and eth3. The aggregation mode is recommended based on the MAC or IP address load.

2. Find the Overlay Network Interface and click Edit Configuration to enter the configuration page to modify the Overlay Network Interface.

3. By checking the Enable high performance mode checkbox, the MTU of the interface of the HCI node will be set to 1600 bytes so that the data encapsulated by VXLAN will not be fragmented when sent to the physical network, which can significantly improve the virtual network forwarding performance. At this time, the physical switch at the opposite end must enable Jumbo Frame.