Okay, after breaking down my entire NSX setup, and creating it newly, this time the only difference being that VXLAN transport VLAN was set to 0. Previously, I had defined this to VLAN100. Why would this cause a VM and ESG connected to the same VNI, unable to communicate?! Since their respective NICs are on the same port group, and on the port group, it is still VLAN 100 that is defined, why would the communication not take place?
Any advise? Thank you!
If the VM and the Edge are on different ESX hosts, then the Edge internal interface and the VM nic communicates through VTEP ports, and the VTEP vlan should be configured on the Physical switch port connected to the Uplink interface as 802.1q trunk port with that vlan allowed. If it works with Vlan 0 which means untagged, this may be the reason,
1 person found this helpful
Is your lab setup a "Nested" setup? If so, VLAN tagging is problematic with nested vSphere hosts. Use of a native VLAN will operate just fine. I am uncertain if there is a documented way around this issue since it has been around for some time.
Yeah, it is a "nested" setup. On the top level vDS, I wasn't specifying the tagging for the transport VLAN. But, on the internal lab vDS, I was tagging with the transport VLAN.
I have below setup of windows vms
192.168.1.1 NSX Edge gateway
All the above machines are in same host. Able to ping NSX gateway from all the windows machines but windows machines cannot ping each other. what could be the issue?