2811 router cuts servers off for a few minutes

[zephyran]

We have a 2811 router at one of our sites that seems to cut off connectivity to a random server or computer in that site from time to time, and then re-allow connections a few minutes later.

When a server is cut off, it can ping through the router to the remote site (the router connects 2 sites via a PtP T1), but none of the application-layer protocols (like RPC or Exchange X.400) can get through, nor can the server telnet to the router.

The router has a single T1 DSU WIC connected to the T1, and 2 VLANs set up on the FastEthernet0/0 interface (as 0/0.1 and 0/0.2), with dot1q encapsulation configured. It's connected to a port on a 2950 switch that is also configured for dot1q trunking. As far as I can tell, there are no address conflicts on either VLAN, and the problem only happens to servers on VLAN1.

We have an identical router at the other end of the PtP with the same configs (different IP addresses, obviously), but no such problems seem to be manifesting there. Has anyone run into this kind of issue before?

 

[zephyran]

In case anyone else has this issue, I fixed it.

Well over a year before this router went into place, we removed a server from that building which used the IP address that was later held by the router. I had checked the network from top to bottom and couldn't find where that IP was living (Pings didn't find it, and the old server was gone long enough for the ARP caches to all clear).

I ended up changing the router's IP address and re-pointing the devices, servers, and clients to that address. Once I did that, it took right off and hasn't missed a beat since.