In a perfect world, you configure a point-to-point link for HDLC or PPP and it just works. However, you may quite often find yourself in a situation when the link fails to come up while you strongly believe you configured everything right. We will briefly discuss in this section, how to isolate and fix problems on point-to-point WAN links.

A simple ping command is a good way to determine if a serial link configured with HDLC or PPP can or cannot forward IP packets. If you are able to successfully ping the IP address on the serial interface of the router at the other end of the link, it is enough proof that the link works.

If the ping does not work, you have a reason to worry. The problem may be related to functions at Layers 1, 2, or 3 of the OSI reference model. The best way to isolate the problem to one of the OSI layer is to use the show ip interface brief command and examine the line and protocol status. 

Table 12-3 Interface Status and Problematic Layer

Line Status Protocol Status Problematic Layer
Administratively down Down Interface is shutdown
Down Down Layer 1
Up Down Layer 2
Up Up Layer 3

Once you have identified the problematic layer, you know where to look. We introduce a few common problem you may face on point-to-point serial links.

 Keepalive Failure

The keepalive feature requires routers to to send keepalive messages to each other, every 10 seconds by default. Keepalive messages are treated as ordinary packets, and they exists for both HDLC and PPP. The HDLC keepalive message is Cisco proprietary, whereas PPP defines a keepalive message as part of Link Control Protocol (LCP).

The keepalive feature enables a router notice a dysfunctional link. A router expects to receive regular keepalives from its neighboring router over an HDLC or PPP link. If a router does not receive any keepalive messages from the other routers for 5 keepalive intervals by default, the router brings down the interface, believing the router on the other end of the link is no longer working. This allows routing protocol to converge and use other valid routes if they exist.

You can change the keepalive interval from the default of 10 seconds using the keepalive command in interface configuration mode. It is possible to speed up failed link detection by reducing the keepalive interval. But this strategy is not useful in all situations. For example, a typical failure of serial link involves losing the Carrier Detect (CD) signal. This sort of failure is detected very quickly, within a few milliseconds. Reducing the keepalive interval cannot speed things up in this case. In most cases, the default keepalive interval is used.

You can disable keepalives using the no keepalive command in interface configuration mode. However, either both routers should use keepalives, or both should disable them. If there is a mistake in which one end leaves keepalives enabled while the other end disables keepalives, the link is bound to fail. This mistake only breaks HDLC links; the PPP keealive feature can prevent the problem.