Routed setup

[ellenverheyen]

Hello,

i just reinstalled proxmox at our server in the DC where we need to use a routed setup. (all trafic must be routed by the main IP and macaddress)

Can someone guide me through the changes we need to make, given we start with a clean install from the proxmox cd?
I read the wiki page about routing... but it was not clear to me.

What we have:
main IP 78.46.33.xx1
mask 255.255.255.255
gateway IP 78.46.33.xx6

extra IP's
78.46.33.xx2
78.46.33.xx3
78.46.33.xx4

3 dns server IP's yyy.yyy.yyy.yy1 (.yy2 yy3)

Kindly regards

Ellen

 

[anthonysomerset]

your anwser is here: http://pve.proxmox.com/wiki/Network_Model

use the routed setup and plug your IP's into it

 

[ellenverheyen]

I've read this, but it's still very unclear to me.
Could you help me with setting this up?

Ellen

 

[tom]

I've read this, but it's still very unclear to me.
Could you help me with setting this up?

Ellen

what about going for a commercial support ticket? looks like you need direct support as the howtos are not enough for you.

 

[ellenverheyen]

I did put this into the /etc/network/interfaces

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet static
        address  78.46.33.xx1
        netmask  255.255.255.224
        gateway  78.46.33.xx6
        post-up echo 1 > /proc/sys/net/ipv4/conf/eth0/proxy_arp


auto vmbr0
iface vmbr0 inet static
        address  78.46.33.xx2
        netmask  255.255.255.0
        bridge_ports none
        bridge_stp off
        bridge_fd 0

but when rebooting I cant ping the host anymore, and I had to startup the rescuesystem to change it to the defaults again:

auto lo
iface lo inet loopback

auto vmbr0
iface vmbr0 inet static
        address  78.46.33.xx1
        netmask  255.255.255.224
        gateway  78.46.33.xx6
        bridge_ports eth0
        bridge_stp off
        bridge_fd 0

Any idea what's going wrong?

 

[anthonysomerset]

i am stumped on this one too my thoughts would be to setup proxmox in kvm mode to take over eth0 with your main ip and provide DNAT routing for the rest of the vm's on a second virtual interface (using internal ip's and mapping them with 1:1 nat)

thinking about it i am not sure how possible this is unless you have full console access to your box (not just ssh)

 

[ellenverheyen]

OK, server is up again.

Had to reboot in the rescue system and change this:

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet static
        address  78.46.33.xx1
        netmask  255.255.255.224
        gateway  78.46.33.xx6
        pointopoint 78.46.33.xx6

auto vmbr0
iface vmbr0 inet static
        address  78.46.33.xx2
        netmask  255.255.255.224
        gateway  78.46.33.xx1
        bridge_ports none
        bridge_stp off
        bridge_fd 0

I can now ping the main IP, but from outsite the vmbr0 IP is not reachable.
I can ping the xx2 address directly on the server itself.
So I suppose this is a routing problem?

Any idea's?

Ellen

 

[anthonysomerset]

with this setup it would work if the host node knew what to do with the traffic sent from vm's to it

you have set the gateway of the vm's to the server itself so they will be able to ping each other but without setting up some actual routing between the 2 interfaces the host node wont do anything, hence needing this line

post-up echo 1 > /proc/sys/net/ipv4/conf/eth0/proxy_arp