Looks broken to me and I don't remember seeing it reported.  File an issue on it and see if someone picks it up.  If anyone else knows better, feel free to respond.

On Thu, Feb 28, 2019 at 8:48 AM Chris Malton <chris@deltav-tech.co.uk> wrote:
Hi all,

Weird one here:

In my config I have an eBGP Multihop peer, and for some reason, despite
the config looking like this:

router bgp 65001
  neighbor w.x.y.z remote-as nnnnn
  neighbor w.x.y.z local-as 205804
  neighbor w.x.y.z description Upstream - Peering Router - b
  neighbor w.x.y.z ebgp-multihop 2
  neighbor w.x.y.z disable-connected-check
  neighbor w.x.y.z update-source 81.187.56.158
  address-family ipv4 unicast
   neighbor w.x.y.z remove-private-AS all replace-AS

nnnnn is the correct peer ASN.

The resulting PCAP in wireshark shows:

Path Attribute - AS_PATH: 205804 65001
     Flags: 0x50, Transitive, Extended-Length, Well-known, Complete
     Type Code: AS_PATH (2)
     Length: 10
     AS Path segment: 205804 65001
         Segment type: AS_SEQUENCE (2)
         Segment length (number of ASN): 2
         AS4: 205804
         AS4: 65001

I am convinced I have done something wrong, as I expected FRR to strip
65001 from the sequence.

I'm running frr 5.0.2 on a pfSense (so alas, upgrading to 6 isn't an
option at the moment).

Any hints what I've done wrong - or is this a known bug?

Regards,

Chris Malton

--
Delta V Technologies Limited
03333 402 402    www.deltav-tech.co.uk
Office: 17 Elm Close, Southampton, SO16 7DT
Company No. 11006104 Registered in England and Wales


_______________________________________________
frog mailing list
frog@lists.frrouting.org
https://lists.frrouting.org/listinfo/frog


--
Don Slice
Cumulus Networks