Have a GRE tunnel between 2 locations. All works fine has been for years..

Have multiple connections into this location so wanted to setup bgp like I have in the rest of my network.

Tunnel is stable... passes static routes fine.. and I have OSPF running over it also..

When I went to setup BGP .. it times out after 180seconds(default hold time)
In error log I get "Hold Timer Expired , Subcode=0"
I assume this is because its not getting keepalives no Idea why keepalives are not making it there.
Can ping at higher than 1540MTU ,

anyone doing this? or have seen this oddity ?

Thanx

Look at routing table after BGP is established. Most likely because of how networks are advertised you are trying to reroute already encapsulated packets inside tunnel interface. Which obviously is not going to work.

Just to confirm. what should I be looking for?
Ip scheme is as follows
Router A
Interface IP 192.168.1.1/24
Tunnel 10.10.1.1/30
OSPF Neighbor 10.10.1.2 (UP, and passing routes)
Bgp Peer 10.10.1.2

Router B
Interface IP 192.168.1.50/24
Tunnel IP 10.10.1.2/30
OSPF Neighbor 10.10.1.1 (UP, and passing routes)
BGP Peer 10.10.1.1


I see in the routing table before and after they peer up that the BGP peer ip's have only one route to its peer, and that is through the tunnel via OSPF Route.
If traffic was already encapsulated I would expect the OSPF sessions to do a similar thing(reseting)?

they are peering on the same IP's as the BGP Sessions.

More--INFO
I did some debugging trying to see where the keepalives are being lost.
Here is a snippet of the log data. Every 3 minutes the process just repeats
It looks like router B is sending and receiving. But router A is sending only?
----------------------------------------------------
ROUTER A
12:37:44 route,bgp,debug,timer KeepaliveTimer expired
12:37:44 route,bgp,debug,timer RemoteAddress=10.10.1.2
12:37:44 route,bgp,debug,packet Sent KEEPALIVE message
12:37:44 route,bgp,debug,packet RemoteAddress=10.10.1.2
12:37:44 route,bgp,debug,packet Length=19


12:48:15 route,bgp,error HoldTimer expired
12:48:15 route,bgp,error RemoteAddress=10.10.1.2

---------------------------------------------

Router B

12:33:35 route,bgp,debug,packet,raw Sent KEEPALIVE message
12:33:35 route,bgp,debug,packet,raw RemoteAddress=10.10.1.1
12:33:35 route,bgp,debug,packet,raw Length=19
12:33:35 route,bgp,debug,packet,raw FF FF FF FF FF FF FF FF FF FF FF FF FF FF
FF FF
12:33:35 route,bgp,debug,packet,raw 00 13 04
12:33:35 route,bgp,debug,packet Sent KEEPALIVE message
12:33:35 route,bgp,debug,packet RemoteAddress=10.10.1.1
12:33:35 route,bgp,debug,packet Length=19
12:33:38 route,bgp,debug,packet KEEPALIVE Message
12:33:38 route,bgp,debug,packet RemoteAddress=10.10.1.1
12:33:38 route,bgp,debug,packet MessageLength=19
12:33:38 route,bgp,debug,packet,raw Received KEEPALIVE packet
12:33:38 route,bgp,debug,packet,raw RemoteAddress=10.10.1.1
12:33:38 route,bgp,debug,packet,raw Length=19
12:33:38 route,bgp,debug,packet,raw FF FF FF FF FF FF FF FF FF FF FF FF FF FF
FF FF
12:33:38 route,bgp,debug,packet,raw 00 13 04

Can you export your firewall settings just for clarification on both devices? You are right, router A is not receiving keepalives from router b starting the clock on the dead timer

In the situation that I have there are no firewall rules..
This device is just routing.
So three should be no reason anything should be blocked.

Show routing table output of both routers when BGP is established and specify what is the address tunnel is connected to.

Could it be related to having 2 Loopbacks on this router? One for GRE Tunnels and another for peering sessions?

I clipped out all the unnecessary routing lines.. would make this a unnecessarily long post

Ip scheme is as follows
Router A
Interface IP 192.168.1.1/24
Loopback1 192.168.2.1/32 (used for tunnel peering only)
Loopback0 10.10.11.1 (loopback0 Used for OSPF and BGP PEERING)(
Tunnel ip 10.10.1.1/30 (on tunnel Interface)
OSPF Neighbor 10.10.11.2 (UP, and passing routes)
Bgp Peer 10.10.11.2


When bgp session is up.

Connected route in routing table
ADC 10.10.1.0/30 10.10.1.1 CUST-4-TUNNEL 0

Bgp peer router in routing table
ADo 10.10.11.2/32 10.10.1.2 110


Loopback0—in routing table
ADC 10.10.11.1/32 10.10.10.1 LOOPBACK0 0


Default route
0 ADo 0.0.0.0/0 192.168.2.1 110



Router B
Interface IP 192.168.1.50/24
Loopback0 10.10.11.2/32
Tunnel IP 10.10.1.2/30
OSPF Neighbor 10.10.11.1 (UP, and passing routes)
BGP Peer 10.10.11.1

When bgp session is up.

Default Route
0 A S 0.0.0.0/0 10.10.1.1 1
1 Do 0.0.0.0/0 10.10.1.1 110

Bgp peer router in routing table
266 ADo 10.10.11.1/32 10.10.1.1 110


Loopback0—in routing table
273 ADC 10.10.11.2/32 10.10.11.2 LOOPBACK0 0

Connected route in routing table
460 ADC 10.10.1.0/30 10.10.1.2 CUST-4-TUNNEL 0


Route pointing tunnel endpoint outside of tunnel traffic.
421 A S 192.168.2.1/32 192.168.5.1 1

That is incomplete information.

1. On router A you have loopback for tunnel peering, on router B you don't
2. You have only 4 routes in routing table? what about BGP installed routes?

As asked before provide IP addresses used for tunnel peering for both routers.
And post the output of commands:
/ip route print where <remote_tunnel_peering_ip> in dst-address
/ip route print where <remote_bgp_peering_ip> in dst-address