2007/10/1, Pascal Hambourg <[EMAIL PROTECTED]>: > However it is surprising that the router answers ARP requests for > addresses that are on the same side. This would not break only DHCP but > all ARP operation on the subnet. > > > What would you advice to get rid of this sittuation ? > > I suppose that turning arp_proxy only on eth0 should work > > I cannot give any advice without knowledge about your network layout, IP > subnets, routing tables, and what you need proxy ARP for.
It's simple . I've got 2 routers. One that is a BGP peer with worldwide routing table that holds a C IP class and has IP x.x.x.1 My router/firewall/TC machine is connected directly to the router with x.x.x.2 IP . It has 3 interfaces : -eth0 for internet default via x.x.x.1 -eth1 for 10.0.x.x clients (DHCP assigned ) -eth2 for 10.0.x.x clients (pppoe-server assigned) Actually 80% of the clients are masquaraded on x.x.x.2 but those who have public IP addr are masqueraded also (DNAT/SNAT) so things like active ftp don't work well . I want to assign IP adressess directly to the interface of my clients - i've been testing it for a while and it works - except for one thing. After about 10 (?maybe less, maybe more?) minutes of inactivity - there's no way to ping or connect to an IP addr from the internet. I think it has something to do with the way I assign IP adresses statically through DHCP - my iptables and arp table are denying customers to connect with another mac address . So when I enable arp proxy on eth0 - there's everything OK - but AFAIK i have to enable it on eth2 and eth1 and then clients get the "DHCPDECLINE" message. I can provide as much info as You want (and I'm able to ;) ) Regards. Wojtek -- Wojciech Ziniewicz Unix SEX :{look;gawk;find;sed;talk;grep;touch;finger;find;fl ex;unzip;head;tail; mount;workbone;fsck;yes;gasp;fsck;more;yes;yes;eje ct;umount;makeclean; zip;split;done;exit:xargs!!;)} -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]