I don't recall this being a problem in v9.1 or v9.5
When I deploy a new Server 2012, 2016 or 2019 VM from template from the Catalog and deploy "powered off" then add the network with IP Pool it works.
Changing NICs thereafter is hit and miss.
If I deploy from template powered on which is before adding the nic, I get a disconnected nic. If I then add a nic while powered on, nothing happens, if I restart the VMs sometimes, it will connect the nic but it does not apply the "IP Pool" ip, instead it applies a dhcp pool IP. Sometimes it remains disconnected.
If I add a second NIC, then either the 2nd NIC works or both suddenly start working.
This is very concerning as it means if a client deploys a VM from template with Power On ticked, getting the NIC to work properly and set the correct IP becomes a problem.
I was on VCD 9.7 so upgraded to VCD 9.7.0.3 with no luck.
I then thought maybe we hardened our templates too much so I used one from a year ago with the same results. These are the same templates we use in IaaS (vCenter 6.7 same as my VCD, same hosts, same storage, same network infrastructure) with no issues.
We use vlan port group backed network so it should be simple layer 2 stuff.
We also found that if you add NICs after VM deployment or make a few nic changes then the VM can no longer ping the gateway or other VMs on the same network and subnet.
It is almost like NSX is blocking access unless the nic settings are perfect. NSX is on version 6.4.5 Build 13282012. I will apply any newer version if available.
I just launched VCD v9.7 but had to move clients off into IaaS due to these issues.
Has anyone else seen the same problem?
