EFW Support

Hi,

We are working on a migration from ESXi to Hyper-V. We've already compiled a kernel (3.12.43 -- longterm kernel branch) with the Hyper-V modules, etc. and it runs flawlessly, as far as we can tell.

There's a VM though, that's giving us some grief. It's running Endian 2.5.1, still on VMware ESXi. We've tried creating a new VM with Endian 3.0.0dev on Hyper-V with our custom kernel and restoring a backup (we've also followed the recommendation of deleting /etc/businfotab, etc. to get back eth0, eth1, due to the new MAC addresses). When doing this, the tap0 interface goes away (i.e. no longer present in the ifconfig output) although in the GUI everything seems to be fine. Even loading the new VM with the original Endian kernel wouldn't bring up the tap0 interface.

We've tried installing our custom kernel on the existing machine and converting it to Hyper-V. After doing this, the tap0 shows up in ifconfig and we can connect to the tunnel. However, the firewall DHCP server is now handing 2 IP addresses to the client VPN, instead of just one. We can ping from the remote client the firewall interfaces, but nothing else within our LAN segment. I also noticed some Connection Refused (error code=11) in the OpenVPN logs.

Any help will be much appreciated.

Thanks

ERRATA: The OpenVPN log message reads:

openvpn[4231]: read UDPv4 [CMSG=8|ECONNREFUSED]: Connection refused (code=111)

ERRATA2: Sorry guys, just noticed that the VPN connection to the original VM (the one on VMware) also hands 2 IP addresses to the VPN client (at least on my Mac using the TunnelBlick VPN client), and yet the VPN works fine.

But there's definitively something going on with the tap0 interface.