12

I have a dedicated Linux (Debian 7.5) root server, with a number of guests set up. The guests are KVM instances, and get network access via bridge-utils (NAT, internal IPs, use the host as a gateway).

E.g. one KVM is my WebServer guest, and it gets accessible via the host IP this way:

    iptables -t nat -I PREROUTING -p tcp -d 148.251.Y.Z 
--dport 80 -j DNAT --to-destination  192.168.100.X:80 

I do the same with other services, keeping them self-contained, NATed and isolated.

But one guest is supposed to be a network monitor, and shall perform network traffic inspection (like an IDS). Usually, in a non virtual setup I would use VACLs or SPAN ports to mirror the traffic. Of course, inside this one host, I cannot do this (easily, because I don't want to use complex virtual switching approaches).

  1. Can I get a port mirror using iptables, and redirect all ingress and egress traffic to one KVM guest? All guests have a dedicated interface, like vnet1.
  2. Is it possible to selectively forward traffic, based on the protocol (like a VACL forward rule, which only grabs HTTP)?
  3. do the guests need a specific interface setup, when I need to keep vnet1 as a management interface (with an IP)?

I would be happy for a point into the right direction:

iptables         1.4.14-3.1
linux            3.2.55
bridge-utils     1.5-6

Thanks a lot :)

1 Answer 1

15
+200

what about prepending the root server pre-Routing module Mangle table rules by something like:

iptables -I PREROUTING -t mangle -j ROUTE --gw 192.168.200.1 --tee

and then prepending the post-Routing module Mangle table rules by something like

iptables -I POSTROUTING -t mangle -j ROUTE --gw 192.168.200.1 --tee 

where 192.168.200.1 is the network monitor.

These rules will mirror all the incoming and outcoming traffic forwarding it to 192.168.200.1

edit:

mangle table specific
  -j ROUTE            (explicitly route packets, valid at PREROUTING)
      options:
      --iface <iface_name>
      --ifindex <iface_idx> 

but you could also use use something like

iptables -I PREROUTING –t mangle –i eth0 –j TEE –gateway 192.168.200.1

and

iptables -I POSTROUTING –t mangle –j TEE –gateway 192.168.200.1

where TEE now is a target which at PREROUTING takes more options like i.e. -i, -p, etc

1

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .