I was going through the below document regarding the CDO mode.
In the document figure 12 & 13 is similar but it says that ping is successful in figure 12 & in figure 13 the ping has failed.
The diagram is the same for both figure 12 & 13 but why ping fails in figure 13.
In figure 12, there are 2 Hosts (Host1 and Host2) and 3 VMs (VM1, VM2 and VM3)
Initially, two VMs VM1 and VM2 were communicating in Host1 and there is another VM (VM3) on Host2.
Because of VM3, Host2 already part of the logical switch VNI 900002 prior NSX Controller Cluster failure (control plane loss).
Therefore, when VM2 vMotions to Host2 the communication continues to work (sucessful ping)
Now in the other picture figure 13, there is no VM3 on Host2 so initially Host2 was not part of the logical switch VNI 900002
When VM2 vMotions to Host2, NSX Controller cannot be notified that Host2 are now the new member of logical switch VNI 900002 and this is why Ping Fails.
This scenario can be handled by enabling CDO mode
In figure 12, there are 2 Hosts (Host1 and Host2) and 3 VMs (VM1, VM2 and VM3)
Initially, two VMs VM1 and VM2 were communicating in Host1 and there is another VM (VM3) on Host2.
Because of VM3, Host2 already part of the logical switch VNI 900002 prior NSX Controller Cluster failure (control plane loss).
Therefore, when VM2 vMotions to Host2 the communication continues to work (sucessful ping)
Now in the other picture figure 13, there is no VM3 on Host2 so initially Host2 was not part of the logical switch VNI 900002
When VM2 vMotions to Host2, NSX Controller cannot be notified that Host2 are now the new member of logical switch VNI 900002 and this is why Ping Fails.
This scenario can be handled by enabling CDO mode