BGP Route advertise issue

[SajeshB]

Hi Everyone,
Need help in this issue, bgp is still advertising the route even though the next hop for the static route is down.
Scenerio.(I've attached a rough NW diagram.)
Their is a client managed Nexus switch connected to Router R2, they did some activity and gig0/0/0 of R2 was down.
However, the path did not failover to R3 from R1. Also, the WAN failover has been tested mutiple times before; if I make R2 peer down to R1, R1 will learn the route from R3.
Because the static route pointing to the Nexus switch through interface gig0/0/0 was down, BGP should stop advertising the network, as predicted. When I checked, the route had been up for more than 20 hours, and the path was via R2 only. As a result, the LAN failover to the Nexus switch did not result in a path failover.
Outside NAT is configured in R2 for Server IP(10.105.0.38) to 10.210.231.214, and this NAT ip(10.210.231.214) is configured as the tunnel IP in R2. So maybe due to this path didn't failover or else there's bgp peering towards client FW, but we have route map applied in R2 (IN direction), so we only learn 100.x.x.x prefix from CLient FW but advertise 10.105.0.38 prefix to client.
I'm trying to figure out if this NAT or BGP peering towards CLient FW is a problem or we need to shutdown both the interface connected to client fw and nx switch to make failover or if it's something else.
I made a connectivity between client FW and Nexus switch(Im not sure about how client device connectivty is)

[MHM Cisco World]

static route is remove when the outlet interface down or next-hop is unreachable. 
when you config the static route with outlet interface and this interface connect to SW "and via SW the next-hop is attach" then the static route never remove because the outlet interface is not down "it up because it connect to SW not direct connect to Next-hop".
solution here use IP SLA to check the reachability of next-hop.

[SajeshB]

The static route is pointed to the next hop IP and to reach Next hop Ip it is via Gig0/0/0 which was Down so as per your point and my understanding the next hop is unreachable and the interface is also down in my case. But the bgp was still advertising the route and I didn't collect the live logs Bcuz when the issue was reported they restored the link Gig0/0/0. So this is the actual issue I still don't know if the static route is removed from the routing table or not. But even the Gig0/0/0 was down which I know from my logging in router, the bgp should have stop advertised that route.

[MHM Cisco World]

OK, if the static route is remove when the outlet interface is down then:-
BGP with network without mask<- this make BGP advertise the network with CLASSFUL mask. 
and since you use 10...... "major subnet" then the mask is 255.0.0.0.
so try config the network with mask.

[SajeshB]

In bgp config, it is advertised as a host network.

Network 10.105.0.38 255.255.255.255

Network is config with mask only.

[paul driver]

Hello

The static route is not conditional and it should be and that condition should be set to state its only valid when its necthop is reachable.

to do that you need to append some ipsla tracking on the static route.

Example:
ip sla 1

icmp-echo 10.252.38 22 source-interface gig0/0
timeout 1000

ip sla schedule 1 start now time forever
track 10 rtr 1 reachability

ip route 10.105.0.38 255.255.255.255 gig0/0 10.252.38.22 track 10

[SajeshB]

hi Paul,

I have similar architecture where when the next hop get unreachable the bgp stop advertising the route so ipsla is not needed.

However I figure out the issue but I wanted to know if that is causing an issue.

there is one more static route with longest 10.x.x.x prefix

10.0.0.0 255.0.0.0 tunnel 620

But in bgp we advertised

Network 10.105.0.38 255.255.255.255

Please help me to figure out if this static route is an issue here  even after the host route is unreachable and interface is down.

[SajeshB]

hi Paul,

I have similar architecture where when the next hop get unreachable the bgp stop advertising the route so ip sla is not needed.

However I figure out the issue but I wanted to know if that is causing an issue.

there is one more static route with longest 10.x.x.x prefix

10.0.0.0 255.0.0.0 tunnel 620

But in bgp we advertised

Network 10.105.0.38 255.255.255.255

Please help me to figure out if this static route is an issue here  even after the host route is unreachable and interface is down for route 10.105.0.38 255.255.255.255 10.235.16.1.