V7.20 BGP automatic Blackhole adding

RouterOS Forwarding Protocols
1

Hello everyone.

Since v7.20.x a blackhole route is automatically added for every advertised prefix even if this route is already in the main routing table. For example:

I’m advertising prefix 10.0.0.0/24 to my remote BGP peers. This is a local IP range, which is already in the main routing table:

/ip route print
Columns: DST-ADDRESS, GATEWAY, ROUTING-TABLE, DISTANCE

DAc 10.0.0.0/24 local-bridge main 0

But now with v7.20.x on top of that I have an additional dynamic Blackhole route with the same prefix and distance=255.

How can I stop blackhole routes to be added to the routing table automatically for prefixes that are already in the routing table?

Thank you in advance.

2

I confirm this behaviour. I personally tried every kind of rule matching and found no solution to this.

Finally I thought that this is not a bad thing anyway since the route have a distance set to 255 and would not be preferred compared to a dynamic learned or connected route and it’s a nice thing to have in case internal routing would be down and no route to advertised subnets it’s available.

3

Hi mmmigoro,

Thank you for your answer. I agree it may be a very handy feature in some cases.

However, I still would like to know how to stop this. :slightly_smiling_face:

4

You can wait for the release of 7.21 or install the current beta, from the changelog:

  • bgp - added output.network-blackhole setting;
  • bgp - do not auto-generate blackhole routes by default (introduced in v7.20);
5

CGGXANNX,

Thank you very much for you answer. Before I was going to roll back to release 7.19.x because what happens with these blackholes is they are generated per connection per prefix. I have roughly 80 connections that advertise 3 prefixes. So I have 80 the same blackholes for each prefix, 240 all together. :slightly_smiling_face:

But since a fix coming in the new version I think I'll wait for release 7.21.

Thank you again.