I have WAN on ether1-gw with IP adress 176.76.242.175 and DMZ on bridge-dmz with different IP range 176.76.240.16/28.
I add IP adress 176.76.240.17/28 to address list for DMZ GW.
The route is added automatically to the route list (176.76.240.16/28, bridge-dmz, Pref.source 176.76.240.17).
I have two servers with static ip addresses in the IP range 176.76.240.16/28: 176.76.240.22 (server1) and 176.76.240.23 (server2) with GW 176.76.240.17.
When I try to ping from server1 to server2, I occasionally get an error:
From 176.76.240.17: icmp_seq=2103 Redirect Host (New nexthop: 176.76.240.22)
Anyone can explain to me where I am making a mistake.
Is the DMZ network and route set up correctly
Since server1 and server2 are in the same IP subnet, they should communicate directly without any gateway. So first step would be to check IP settings (address, subnet mask) on both servers to verify they match intended use (error message shows that router is involved in at least some of ping exchange but it shouldn’t be).
Another thing would be to verify arp settings on router …
If you don’t get it working, post full text export of router configuration.
First off: are the two servers supposed to communicate with each other a) through firewall or b) are they allowed to communicate directly?
If it’s b), then they should be able to communicate even if they are connected to a dumb switch. Hence you should check if they have proper IP settings, specially the network mask (it should either be /28 or 255.255.255.240). Again: getting that next hop message indicates that server1 communicates with server2 via router. And that doesn’t have anything to do with router configuration.
If it’s a), then you’ll have to to redesign DMZ config a bit because in this case router should be handling all the traffic without issuing next hop ICMP packets … which again is result of using subnet mask, but in this case on the router.
So you have ether11 and ether12 bridged for the DMZ in question (and ether12 is actually disabled). I don’t see error which would force servers to communicate via gateway. Since router isn’t running DHCP server for that subnet I assume servers have IP settings configured manually. So I’m asking you (again) to tripple check settings on servers.
Since both servers are (obviously) connected to same router port (ether11), what is the topology of network between router and servers? Are servers virtual machines running off same hypervisor?
BTW, you’re using various 172.100.x.y subnets. I’m sure you’re aware these are not RFC 1918 addresses for private use. They are registered to Charter Communications Inc.
Ether12 is temporarily disabled. The router does not use DHCP for this subnet. The servers have manually configured IP settings.
The servers are virtualized and run on the same hypervisor.
The Hyperv server, where virtual systems run, is connected via a switch to the router’s ether11 interface.
DMZ subnet 176.76.240.16/28 is with an external provider (Telekom) - external IP addresses of Telekom.