jibun no kage
New Member
Wondering if the following is the best way to force clients to route through a specific gateway? Here is the scenario, we have two routers, that are not bridged and can't be. Each router is connected to a given ISP. Our internal DHCP servers are configured to be local authoritative, so each client gets two router IP addresses, for discussion 192.168.1.1 and 192.168.1.2. Our DNS servers forward non local resolves to the ISP DNS servers. This works until the ISP DNS servers for a given ISP stop responding, stop resolving. The given ISP DNS servers are not down, just stop resolving once in a while. This effectively blinds our internal DNS servers. Because the router is pingable, and the ISP DNS servers are pingable, we have an issue since no DNS resolution is working. Now if we could some how test the DNS 'path' and then block the broken path, this would be the best of both worlds, so to speak. In any other environment, we would just bridge the routers, setup cross over to each ISP, and call it done. But in this case, we can't, so we need to do something creative.
For example... In a full working scenario...
Route 1, workstation 1, local DNS 1 forward, using gateway 1, transports to router 1, then ISP 1 DNS responds
Route 2, workstation 1 local DNS forward, using gateway 2, transports to router 2, then ISP 2 DNS responds
But if say ISP1 DNS does not respond or resolve but is still pingable, the DNS query fails. How can I force workstation 1 to use then use IPS 2 DNS? The first thought was to have a watch dog service, test a DNS query and if it fails, just delete the route 1 entry from the routing table on local DNS 1, so that it can't use or see the non-responding ISP 1 DNS source. But is that the only way? Again the routers are limited, we can't use any of the usual fancy features or methods, enterprise routers have, since they are very simple routers.
For example... In a full working scenario...
Route 1, workstation 1, local DNS 1 forward, using gateway 1, transports to router 1, then ISP 1 DNS responds
Route 2, workstation 1 local DNS forward, using gateway 2, transports to router 2, then ISP 2 DNS responds
But if say ISP1 DNS does not respond or resolve but is still pingable, the DNS query fails. How can I force workstation 1 to use then use IPS 2 DNS? The first thought was to have a watch dog service, test a DNS query and if it fails, just delete the route 1 entry from the routing table on local DNS 1, so that it can't use or see the non-responding ISP 1 DNS source. But is that the only way? Again the routers are limited, we can't use any of the usual fancy features or methods, enterprise routers have, since they are very simple routers.