Well, my previous post wasn't solved, after all...

I am trying to connect to a VirtualBox VM from a remote, through an OpenVPN tunnel that is hosted by the same computer.

If I am logged-on to that computer, e.g. by VNC or SSH, then I am able to ping the VM (which has a "bridged" adapter #2 ...) at 192.168.1.183. When running WireShark on that machine, I see both the ICMP ping request (from host 192.168.1.199), and the ICMP ping reply.

If I now attempt to ping that same address from my own computer, which is connected through the tunnel, I see an ICMP ping request, this time from 10.44.55.66 [i](the VPN-assigned address of my computer), but no reply.

I would conclude from this that the packets aren't making it.

Just for kicks, I tried issuing route add -host 19.168.1.183 192.168.1.199, but this is clearly incorrect: now the packets just come flying back to me and don't show up in Wireshark at all. (If I use -net, I simply get "time-to-live exceeded" and once again no packets show up.)

I'm beginning to wonder if I must do something with static routes on the office router ... as I had to do in order to get VPN to work, for instance. Do I have to, say, route packets that are addressed to 192.168.1.183 to 192.168.1.199, on that router?

The problem remains un-resolved but here's a few more things about it:

(1) The VM does, indeed, appear on the local network and it can be accessed from there. Also, I can ssh (or telnet) to the device, through the tunnel, and from that (local) session ping any virtual network-adapter of the VM. But I still cannot ping it through the tunnel. I see how the VM has a promiscuous-mode adapter setting (on the host ...) to allow it to receive traffic not intended for it and to thereby mimic a second physical box on the local network.

(2) I set up a "Host-Only Network" (vboxnet0, if I recall ...), and see that listed as an interface in the host's ifconfig output. I see that this is handled by "virtualbox." Once again, I can ping that adapter when I am logged-in to the box. But, I cannot ping even that adapter from the other side of the tunnel.

I'm still scratching my head about why this isn't working, and what to do about it.

Well, I also "solved" this one ... and I'm still not quite sure what was wrong.

The VM has two interfaces: one NAT, and one Host Only (which I defined as an experiment in connecting to it ...). Both of these IPs can be pinged from the host.

I finally succeeded in being able to ping the public (NAT ...) address of the VM, while connected through the tunnel. I'm honestly not sure why it did not work before.

I have not been able to ping the Host Only address, even though I am tunneling to the same machine that hosts both OpenVPN and VirtualBox. But I did not entirely expect to be able to do so.