EMC

My setup is the following: I have 2 NIC s in my system, a pci-e one and a MOBO integrated one. I run pfsense in a virtual machine witch acts as my router, the pci-e nic is passed through to it, it gets the WAN connection from there. I have a virtual network interface set up between the unraid host`s bridge network (br0) adapter(the onboard NIC) and the pfsense guest, I would like to use the onboard NIC for my LAN port. My problem: When I have the virtual bridge configuration like this: <interface type='bridge'> <mac address='52:54:00:80:59:c1'/> <source bridge='br0'/> <target dev='vnet0'/> <model type='e1000'/> <alias name='net0'/> <address type='pci' domain='0x0000' bus='0x01' slot='0x00' function='0x0'/> </interface> Everything works fine, the unraid machine and other hosts on the network can communicate with each other and the internet, the problem is that the e1000 interface is slow, it theoretically supports 1Gbits/s, but I`m only geting about 500Mbits/s between the host and the pfsense machine (tested with iperf), and I`m planing to upgrade to 2.5G networking. If I change the adapter to use <model type='virtio'/>, the connection speed jumps up to 5Gbits/s, and other hosts can still communicate flawlessly, but my unraid host can`t. I`m able to ping 8.8.8.8 and the ip of the gateway from the terminal, but none of my docker containers have network access, and my apps tab is broken as well. I belive this is a routing issue between the virtio interface and unraid. My unraid routing table in both cases is the following (it works for e1000, doens`t work for virtio😞 Does anyone have any idea why unraid refuses to use the virtio bridge?

Hello everyone, I have a not too common unraid (6.6.7) server setup. I have a pf sense firewall running as a virtual machine with a dedicated pci e nic passed through to it. That serves as my router. I have been using this setup for months now without any problems, but there is something wrong in my setup, and that is that my host unraid server is connected to the virtual firewall via an external switch, so basically I have a cable coming out of my nic (witch is assigned to PS sense), going into an external switch, and then going back to my motherboards ethernet port. This config was fine for me, but I am planing to upgrade to a 10 GB network between my pc and my server and I don't want 2 cables coming out of the server running to my pc so I want to pass the 10 GB nic to PfS as well. The problem is that if I want to have a connection to my server via the 10 GB link the same way I would have a buy 2 nic s (1 with 2 ports) for my server, assign the 2 port one to PfS, keep the 1 port one to unraid, and basically having them pluged into each other. That config would be more than stupid, so I came up with a solution. I set up a virtual bridge between my PfS vm and the host. I disconnected the ethernet cable from the server and set up a route to my PfS firewall using the virbr0 interface, so my routing table looks like this: Most of the stuff works fine, I can access the web ui from any machine on the network, can transfer files, and my unraid server can access the internet with no problem as well.(I ran an iperf test between my PfS vm and the server and I was getting speed around 6-7 GB/s, so it would be enough for 10 GBe) There is problem however with this config and that is that docker does not recognize virbr0 (my virual bridge) as an interface and this causes 2 main problems: - As docker does not now what the machines IP address is, when I try to access a containers web ui it gives me a blank page, my port mappings looks like this: (I have a reverse proxy set up with letsencrypt, so most of my containers are in a custom docker network, their custom ip shows up, but the host ip is blank, only the port is shown) I can get around this issue by just typing my servers ip address in the browser manually and define the containers port, I can access the containers that way, and the containers can also access each other that way. - My other problem is a bit more serious, as docker does not recognize the virtual network bridge, containers can't directly see them as well: The pic above is my ovpn client servers network config, and it is seen that I can only select the containers own loopback interface, and the custom docker network for the server to listen on. This issue also effects delugevpn as well (and any other container that has to know the machines ip, mostly VPNs). So my question is: If there a way to make docker detect the virtual bridge as a network interface, or to force it to use a pre set ip for the containers?