I am trying to get a lab setup to work, but NAT connections from the VMs fails after a few minutes.
I have three hosts: HOST1, HOST2, and GWHOST1. GWHOST1 runs the gateway VM, GATEWAY1. HOST1 runs VM1 and VM3 and HOST2 runs VM2.
VMQ is not enabled anywhere.
The gateway VM has one default route (to the external gateway 172.20.0.1).
VM1 and VM2 are connected to the virtualized VM network Test1 (192.168.1.0/24).
VM3 is connected to the virtualized network Test2 (192.168.2.0/24).
The provider address space is 172.21.0.0/24.
The external (front end) address space is 172.20.0.0/24.
Scenario 1:
After initial setup I start ping -t 172.20.0.1 (an external router) on VM3 and it starts to ping for a while. After a few minutes ping fails with 'Request time out'. As long as ping in the VM works I can ping -p 172.21.0.3, the provider address for GW-External, from HOST2 and HOST1. When ping stops responding in the VM ping -p to 172.21.0.3 also stops responding to HOST1 and HOST2 (Detsination host unreachable). Why does 172.21.0.3 disappear?
Disconnecting and reconnecting the Test1 from the gateway reestablishes 172.21.0.3 and ping starts to work again from the VM.
Scenario 2:
I start ping -t 172.20.0.1 on all VM machines and it works for a while. After a few minutes ping fails with Request timed out. At the same time 172.21.0.3 stops responding to ping -p.
Disconnecting and reconnecting Test1 from the gateway reestablishes NAT for Test1 (ping works). Disconnecting Test2 from the gateway makes NAT for Test1 start to fail and 172.21.0.3 disappears again.
I am at a loss to get this to work. There seems to be some kind of software glitch but I can't believe MS releases something with such a fundamental bug, on the other hand DPM 2016 was a real horror story when it was released.
I have tested the above on different physical hosts, different switches, and changed cables. No luck.
Any ideas?
Regards,
Hans Carlsson