Unexpected static route behavior when next hop address is disabled

[kkudlacek]

I am having difficulty interpreting the Cisco Configuration guide relative to static route entries and valid next hop addresses. As I interpret the Cisco Documentation, when a static route to a remote network is configured to a valid next hop address and the next hop address goes away (i.e. fails) the route should be pulled from the active routing table... not true according to my lab tests.

Here is the quote from the configuration guide (URL:http://www.cisco.com/univercd/cc/td/doc/product/software/ios111/mods/4mod/4cbook/4ciprout.htm#xtocid13404175) "... Also, when the software can no longer find a valid next hop for the address specified as the forwarding router's address in a static route, the static route is removed from the IP routing table."

Anyone have any knowledge of this? Testing suggestions? My setup: on R1 I am setting a static to a remote net via a router R2. When R2 is unplugged from the LAN, the remote net still shows in R1's routing table, even after a clear ARP with the ARP table showing no R2.

Maybe I'm reading the documentation wrong...am I going nuts?

Thanks in advance!

Kale

[svermill]

Your routing table will not reference your arp cache in any way. I think your problem is rooted in the fact that you are referencing a next hop on an ethernet. The next hop address exists in the subnet assigned to the ethernet segment, which is directly attached, which is probably up/up regardless of whether or not R2 is up and running. This is different than, say, ppp, where if R2 were down, ppp would follow (up/down) and thus your next hop subnet would in its entirety be down. I think it is far more common that static routes be defined out serial interfaces rather than ethernet interfaces, so the generalization that the static route will flush out when the next hop goes down is probably true most, but not all, of the time. My guess anyway.

[Erick Bergquist]

Are you using a IP address as the next-hop or the interface name? If using the interface name, devices with proxy-arp enabled who know how to get to that network will respond. This may be what is happening in your test if you multiple routers.

[kkudlacek]

I'm using the next hop IP address in my tests... I think the first poster may have it right when he speculates that Cisco meant the interface has to go down in order for the static route to go away (not just the next hop router going down.

[milan.kulik]

Another situation when the route entry remains in the routing table even while the next hop router is down is using "permanent" option in the "IP route ..." command (which is a usefull way of routing loops preventing in the case you are absolutely sure there is the only one path to the destination network).

Regards,

Milan

[fullerms]

I feel that there is no issue here. If a next hop interface goes down, there might be another route to reach the ip address in question. Take the case of unnumbered interfaces connected to a cloud. All unnumbered interfaces connected to the cloud will have the same IP address. So, even if the directly connected interface goes down, it still might be possible to reeach the IP address in question via another router.

I feel that the local interface must go down for the route to be removed unless of course it is defined as "Permanent"

Am I getting this right ?