First off the versions…
vRA 7.3.0.536 Build 5610496
NSX 6.4.0 Build 7564187
Issue: When the blueprint is deployed, the VM receives all the configured IP info from DHCP, except DNS.
Setup: I created a blueprint with a vSphere Windows 10 VM and added an On-Demand NAT Network. The VM IP settings are configured to receive the information via DHCP. The On-Demand NAT Network has a parent network profile, external network profile, NAT type (1 to many), subnet mask and gateway assigned in the General tab. In the DNS/WINS tab I have primary and secondary DNS server along with a DNS suffix, no WINS. The DHCP tab has the start and end of the range and the IP Ranges tab is empty. The parent network profile has the same General, DHCP and DNS info.
Troubleshooting: After I’ve deployed the blueprint and I look into the properties of the Edge VM that was created for NAT (vSphere Web Client, Networking and Security, NSX Edges, name of Edge, Manage) I can see that in the DHCP section the entry with IP range has “Auto configure DNS” turned on and of course the DNS info from the parent network profile (and On-Demand NAT Network) is blank. If I manually make the change and publish then all works fine. So it seems that during the deployment vRA does not pass on the DNS settings but instead turns auto configure on.
I’ve tested with other VMs (different OS) and the behavior is consistent. The Edge VM deployed by vRA does not receive the DNS settings but auto configure is on.
If someone can point me in the right direction it would appreciated.