Hi all.

I have doubts about the operation of a machine haproxy
debian virtualized with 10 ethernet cards.

I have eth0 configured management server and the rest as frontend,
each with their backend. Everything works fine, but I'm going crazy
because I am getting graphic
bandwidth used by different frontend and traffic does not match the
configuration.
One example that is happening:

The eth1 is 192.168.1.5 and my haproxy I have the frontend
squid listening on that ip. Everything works fine, but the graph of
nagios not see activity on eth1. I watch with
wireshark and see the traffic coming out of the eth4.

I tried to configure haproxy to accept only a frontend
packets from one network interface, but I have not worked. Squid
refuses connections. example:

listen squid
     bind 192.168.1.5:3128 interface eth1


Why can it be?

Reply via email to