Hello to all...
I installed an infrastructure (in lab) with 3 hosts ESXi 7.2 with VCSA 7.1 and NSX-T 3.1.1
I've configured 2 Edge , 1 Gw-T0 in Act-Act and 2 Gw-T1 with SR and 1 Segment for each T1.
I've powered on 1 VM on each segment in this way the tunnel on 2 Hosts is up.
All network routing works fine from and to inside VMs in NSX-T environment.
Now I wanted to verify the Gateways present in the hosts that have tunneling up, and I would have expected both the DR of the GW-T0 and of the GW-T1 created on the hosts.
When I ran the get logical-routers command on the hosts, as output I only received the DR of the GW-T1.
Why is the DR of the GW-T0 not present?
Thanks to all for any suggestions.
oki....sorry for post....I had misread the article ... I created the same type of connections and I am with what is written in the article.
Why didn't I see the DR of the Gw-T0?
Because my Gw-T1s were connected to the Edge Cluster (so they have an SR)
(it's the second scheme in the article)
By creating a Gw-T1 without SR, on each host where the DR of the Gw-T1 is present, the DR of the Gw-T0 is also active
Thanks and sorry for the post ...
That's correct, because you have SR / stateful services configured, only the T1DR will exist on the hosts.
Then the data passes to the T1SR on the edge, after that it hits the T0DR on the same edge. After this it is then forwarded in whichever destination it needs, whether it be the T0SR or the next T1.
If you didn't have stateful services configured or an edge cluster attached to the T1, then you would have a T0DR on the host.
ok...
I asked this, because reading this post:
I would have expected the DR of T0 and T1 on the hosts and the SR of the GW on the Edge.
So this post is incorrect or is it old, as on NSX-T 3.x this behavior has changed?
I add some info ...
On the Gw-T1-A a Segment-A is connected to which a VM-A is connected
On the Gw-T1-B a Segment-B is connected to which a VM-B is connected
Gw-T1-A and Gw-T1-B are connected on Gw-T0
Now what happens ....
If the VM-A and VM-B are on two different hosts, the DR of the Gw-T1-A is active on the Host-A and the DR of the Gw-T1-B is active on the Host-B.
If the VM-A and VM-B are on the same host, the DR of the Gw-T1-A, Gw-T1-B and also of the Gw-T0 are active on the host (it appeared when I happened to vmtion)
This is what verified by doing tests.
Does this mean that the Gw-T0 is activated on the host if necessary?
And when it is not active, does the traffic between VMs of different Segments, is routed through the Edge, instead of taking place in the host kernel?
Thank you.
oki....sorry for post....I had misread the article ... I created the same type of connections and I am with what is written in the article.
Why didn't I see the DR of the Gw-T0?
Because my Gw-T1s were connected to the Edge Cluster (so they have an SR)
(it's the second scheme in the article)
By creating a Gw-T1 without SR, on each host where the DR of the Gw-T1 is present, the DR of the Gw-T0 is also active
Thanks and sorry for the post ...