BGP received-routes no such neighbour

I’m seeing 2 things on 1.3.2 (the LTS image) which makes me wonder what is going wrong or what I’m missing.

  1. I can see on my bgp session that supposedly I’m sending a full route to my peer (which should not happen since there is a route-map that only sent a single route which isn’t existing currently.), however according to this I’m sending the full table. Since I’m also in control of the other side I could see that as it should I’m not sending anything.

Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd PfxSnt
185.xxx.xxx.xxx 4 123456 468243 425969 0 0 0 12:24:48 899880 899880

The asn is obviously changed, i’m not peering with as 123456 :wink:
Why does it say I’m sending 899880 prefixes here while in fact I’m sending none? Even during a advertised-routes it says I’m sending a full table.

This is what I see on the other side:
185.xxx.xxx.xxx 123456 ESTAB 12h41m25s 0 0 904554 0

The first 0 is what I’m announcing to it.

I have a prefixlist which includes a single /24, which is tied to a route-map, the routemap is listed as out for the session.

Clear ip bgp neighbor x.x.x.x soft out
show ip bgp neighbor x.x.x.x adv

show conf com | match “policy|bgp”

I think you mean reset instead of clear.

The second command just show the full table, not going to paste it here since that would just be pollution.

As for the last part:

set policy prefix-list announce rule 5 action ‘permit’
set policy prefix-list announce rule 5 le ‘32’
set policy prefix-list announce rule 5 prefix ‘185.10.xxx.0/24’
set policy route-map transit-out rule 5 action ‘permit’
set policy route-map transit-out rule 5 match ip address prefix-list ‘announce’
set policy route-map transit-out rule 10 action ‘permit’
set policy route-map transit-out rule 10 set tag ‘666’

set protocols bgp 123456 address-family ipv4-unicast
set protocols bgp 123456 neighbor 185.xxx.xxx.xx address-family ipv4-unicast nexthop-self
set protocols bgp 123456 neighbor 185.xxx.xxx.xx address-family ipv4-unicast route-map export ‘transit-out’
set protocols bgp 123456 neighbor 185.xxx.xxx.xx address-family ipv4-unicast soft-reconfiguration inbound
set protocols bgp 123456 neighbor 185.xxx.xxx.xx remote-as ‘23456’
set protocols bgp 123456 parameters always-compare-med
set protocols bgp 123456 parameters graceful-restart
set protocols bgp 123456 parameters log-neighbor-changes
set protocols bgp 123456 parameters network-import-check
set protocols bgp 123456 parameters router-id ‘185.xxx.xxx.xx’
set protocols bgp 123456 timers holdtime ‘30’
set protocols bgp 123456 timers keepalive ‘10’

Which version?
I guess “reset ip bgp x.x.x.x soft out” for some reason could not work in your case
Did you try reset hard peer, I mean with re-establish session?

show ip bgp neighbor x.x.x.x

By the way you didn’t send full “bgp|policy”config

This part configuration should work as expected

The route-map “transit-out” contains only “permit” actions, which will I think result in the full routing table being advertised to your peer. You need to “deny” routes you do not wish to advertise.

Version 1.3.2 which the latest LTS downloaded from the support area.

I tried your complete command but it errors out:
show conf com | match “policy|bgp”
Invalid command: [bgp”]

So I did them separately.

From what I can gather from other posts, only putting in that prefix in the list should only allow that one out.

Edit:
I think I fixed my own issue, but not sure why it would be wrong, on my route-map at the end I had:

set policy route-map transit-out rule 10 action ‘permit’
set policy route-map transit-out rule 10 set tag ‘666’

even though rule 5 matches on a specific prefix list, with the tag included it re-announces everything, without it announces nothing (which is correct since the subnet it needs to advertises isn’t on it yet.)

I guess I need to set my community’s differently.

Your rule 10 has action permit, without any filter.
Try same filter on rule 10,
or
combine permit and setting tag into rule 5 and get rid of #10