Hello, 

 

Context :
I have two interfaces on my unraid server (6.12.5), I would like to have :

- one interface (eth0) for dockers, web api access, that can access from my private network (1Gbits/s) .  

- and one (eth1) that will be use by a container to access internet with a VPN (with a 10Gbits/s). 

My private network has already Vlans but have switch and router limited to .
The first test I want to do, is to force the container with theVPN to pass in a specific Vlan on eth1.

In a close future, eth1 would be on a completely separate network. 

Concrete case:
On my private network I have a Vlan Server: 10.10.10.1/24 and a Vlan Fast : 10.10.20.1/24
If both eth (0 and 1) are one 10.10.10.1 (ex : eth0 = 10.10.10.2 and eth1 = 10.10.10.3) I can connect to Unraid but when I change eth1 to  10.10.20.2, I can't acess the web ui anymore. 

Actually (both interface on the same subnet) they are like :
PROTOCOL    ROUTE            GATEWAY                    METRIC
IPv4        default            10.10.10.1 via eth1        0
IPv4        10.10.10.0/24    br0                        1    
IPv4        10.10.10.0/24    eth1                    1

I think it is weird that the default is "via eth1" 
Looking for an answers, I seen a post indicating that metric on route need to be modified.
Maybe it should be like :
IPv4        default            10.10.10.1 via br0        0
IPv4        default            10.10.10.1 via eth1        1
IPv4        10.10.10.0/24    br0                        1    
IPv4        10.10.10.0/24    eth1                    1

And in if I change the subnet to eth1 :
IPv4        default            10.10.10.1 via br0        0
IPv4        default            10.10.20.1 via eth1        1
IPv4        10.10.10.0/24    br0                        1    
IPv4        10.10.20.0/24    eth1                    1

Am I wrong on that supposition ? 

Hi,

 

I have my Unraid with 2 interfaces however I just use eth0 for Unraid mgmt and VM traffic and eth1 for dockers, in my scenario I just have an IP on eth0, I don't see why add another IP to Unraid on another interface, unless your goal is Unraid mgmt redundancy.

Quick question, if I understood eth0 is 1Gb interface and eth1 is a 10Gb interface correct?

 

If both are 1Gb, create a bond and add the VLANs you need, you can let the Unraid mgmt on a native VLAN and create VLANs for the VMs and Dockers.

Hi, thanks for the answer !

Yes "eth0 is 1Gb interface and eth1 is a 10Gb interface". 
I succeed to manage the Unraid part. Not the docker part. 
For unraid I needed to set bond to false for both interfaces, like that br0 use eth0 and br1 is create to use eth1.
I also set the Metric like :
IPv4        default            10.10.10.1 via br0        1
IPv4        default            10.10.20.1 via br1        2
IPv4        10.10.10.0/24    br0                        1    
IPv4        10.10.20.0/24    br1                    2

Now the contenair I use is from Hotio, from some week now, the images has integrate a client vpn in it (wiregard in my case).
I tried to set the network to br1 and bind the port 8080 to the eth0 interface. But that didn't works. 
I bind it using -p 10.10.10.2:8080:8080 (10.2 is the ip of unraid, so eth0 ip). 

I think there is something I am missing but don't know what