Deployment Guide

OS10(config)# interface ethernet1/1/3
OS10(config-if-eth1/1/3)# ip ospf 100 area 0.0.0.0
OS10(config-if-eth1/1/3)# exit
OS10(config)# interface loopback 1
OS10(config-if-lo-1)# ip ospf 100 area 0.0.0.0
Each VTEP switch in the underlay IP network learns the IP address of the VXLAN source interface. If a remote VTEP switch is not
reachable, its status displays as
DOWN in the show nve remote-vtep output.
2. Configure the MTU value on L3 underlay network-facing interfaces in Interface mode to be at least 50 bytes higher than the MTU on
the server-facing links to allow for VXLAN encapsulation. The range is from 1312 to 9216.
mtu value
3. Return to CONFIGURATION mode.
exit
Configure VLT
(Optional) To use VXLAN in a VLT domain, configure the VLT domain — including the VLT Interconnect (VLTi) interfaces, backup
heartbeat, and VLT MAC address — as described in the
OS10 Enterprise Edition User Guide in the Virtual link trunking section.
Required VLT VXLAN configuration:
• The IP address of the VTEP source Loopback interface must be same on the VLT peers.
• If you use a port-scoped VLAN to assign tagged access interfaces to a virtual network, to identify traffic belonging to each virtual
network, you must configure a unique VLAN ID for the VLT Interconnect (VLTi) link.
• Configure a VLAN to transmit VXLAN traffic over the VLTi link in VIRTUAL-NETWORK mode. All traffic sent and received from a
virtual network on the VLTi carries the VLTi VLAN ID tag.
Configure the same VLTi VLAN ID on both VLT peers. You cannot use the ID of an existing VLAN on a VLT peer or the reserved
untagged VLAN ID. You can use the VLTi VLAN ID to assign tagged or untagged access interfaces to a virtual network.
virtual-network vn-id
vlti-vlan vlan-id
• Although a VXLAN virtual network has no access port members that connect to downstream servers, you must configure a switch-
scoped VLAN or VLTi VLAN. The presence of this VLAN ensures that the VLTi link is added as a member of the virtual network so
that mis-hashed ARP packets received from the VXLAN tunnel reach the intended VLT node.
Best practices:
• If a VLT peer loses connectivity to the underlay L3 network, it continues to transmit routing traffic to the network through the VLTi
link on a dedicated L3 VLAN to the other VLT peer. Configure a L3 VLAN between VLT peers in the underlay network and enable
routing on the VLAN; for example:
OS10(config)# interface vlan4000
OS10(config-if-vl-4000)# no shutdown
OS10(config-if-vl-4000)# ip address 41.1.1.1/24
OS10(config-if-vl-4000)# ip ospf 1 area 0.0.0.0
• To reduce traffic loss when a VLT peer boots up and joins an existing VLT domain, or when the VLTi links fails and the VLT peer is still
up as detected by the VLT heartbeat, create an uplink state group. Configure all access VLT port channels on the peer as upstream
links. Configure all network-facing links as downstream link. For example:
OS10(config)# uplink-state-group 1
OS10(conf-uplink-state-group-1)# enable
OS10(conf-uplink-state-group-1)# downstream ethernet1/1/1-1/1/2
OS10(conf-uplink-state-group-1)# upstream port-channel 10
L3 VXLAN route scaling
The S4100-ON series, S5200-ON series, S4048T-ON, and S6010-ON switches support native VxLAN routing — routing in and out of
tunnels (RIOT). RIOT requires dedicated hardware resources reserved for overlay routing. You cannot use these dedicated resources for
underlay routing.
VXLAN
13