running the nsxcli get command through the CLI shows options

In this section some basic CLI commands for nsxcli will be provided. These can be used for troubleshooting.

Please take into account the NSX-T Data Center Administration Guide for NSX-T 3.2 here.

# Logical Switching

Common switching problems
N-VDS is incorrectly configured on a host
Overlay tunnel (GENEVE) is misconfigured
TEPs unable to reach each other

# Validate switch

esxcfg-vswitch -l # verify switch configuration

nsxdp-cli # Verify nsx local datapath services and statitics

# Verify network interfaces

ifconfig
net-stat -I
Verification Process

# ssh to NSX manager node

su admin # enter nsxcli command mode
get logical-switches # Verify all logical switches/segements configured in NSX manager
get logical-switch <segment-uuid> ports # verify the logical switch ports connected to the segment
get logical-switch <segment-VNI> transport-node-table # list the transport node table of the segment logical switch
get logical-switch <segment-VNI> arp-table
get logical-switch <segment-VNI> map-table
get logical-switch <segment-VNI> vtep
get nodes # list all the transport nodes
get cluster status # check all nodes and its service statuses
get cluster vip # on which node does the vip run
st eng # from admin to root (exit to return to admin mode)

# ssh to Edge Transport Nodes

get controllers # which controller is node connected to
get managers # check if all connected

# ssh to ESXi host

nsxcli # enter nsxcli command mode
get logical-switches # It will list the switches VNI, UUID, DVS name, VIF numbers
get logical-switch <segment-VNI>
get logical-switch <segment-VNI> mac-table
get logical-switch <segment-VNI> arp-table
get logical-switch <segment-VNI> vtep

DNS and NTP configuration NSX-T nodes

To query and configure DNS/NTP configuration for NSX-T Manager nodes, you can login using SSH to the IP address of the NSX-T (3.x) Manager node(s) using the admin, and run the following commands here:
  • nsxt01-1> get name-servers
    • Output
      [date and time]
      172.0.0.11
      172.0.0.12
  • nsxt01-1> get ntp-servers
    • Output
      [date and time]
      10.0.0.7
      10.0.0.23
      10.0.0.39

For more details, please check this post.

# ssh to KVM host

sudo -i # enter root mode
virsh dumpxml <vm-name> | grep interfaceid # obtain the interfaceid of the required vm
nsxcli # enter nsxcli command mode
get logical-switches # It will list the switches VNI, UUID, DVS name, VIF numbers
get logical-switch <segment-VNI>
get logical-switch <segment-VNI> ports
get logical-switch <segment-VNI> map-table
get logical-switch <segment-VNI> arp-table
get logical-switch <segment-VNI> vtep

# Check GENEVE VMKernel

esxcli network ip interface ipv4 | get vmk10
vmk10 is the TEP for NSX
esxcli network ip interface ipv4 | get vmk50
vmk50 is for intra-tier networking/routing and containers.
Verifying overlay tunnel reachability
Ping destination TEP interface from the source host

vmkping ++netstack=vxlan -s Vxlan is used by host rather than GENEVE. It’s the same stack for ESXi.
Try 1572 if 1575 fails This is the minimum size needed to support GENEVE. GENEVE adds 72 bytes to a 1500 byte data packet.
If 1572 fails try 1472 if that works, the overhead for the overlay hasn’t been configured.

Example

vmkping ++netstack=vxlan -s 1572 -d <TEP-IP> # using 1572 data bytes, and ping destination TEP

By DM