Community discussions

MikroTik App
 
carlnrtn1
just joined
Topic Author
Posts: 1
Joined: Sat Jul 17, 2021 1:59 am

Strange OSPF default route behaviour

Sat Jul 17, 2021 2:08 am

Hi all,

I'll start with a basic topology to describe my problem:

R1 ETH1 Area 0 >>> R2 ETH1 Area 0 > R2 ETH2 Area 1 >>> R3 Area 1 >>> R4 Area 1 >>> R5 ETH1 Area 1 > R5 ETH2 Area 0 >>> R1 Area 0

Basically, R1 is our core is in area 0. R1 to R2 is a layer 2 circuit provided by a third party provider. It is our main uplink for this leg of the network. However, we have an alternative route out via R5 which is through an FTTP connection tunneled back to our core router.

Tonight a strange problem occured. There has been a planned outage scheduled for the third party L2 connection. The line went down as expected, however nothing else on the rest of the network recovered. R5 was accessible so a bit of mac-telnet over mac-telnet got us logged in up the network and we found that R2 was learning a default route from R2.... even though R2 had no default route in it's routing table! None of the routers are configured to distribute default.

The problem seems to arise from this R2 L2 connection. Even though the L2 circuit was down, the interface was still up (the outage was within the third parties core, so the direct connection was still up) - and for some reason this meant R2 was sharing this default route even though it didn't have one. As soon as we disabled the uplink interface, it stopped sharing the non-existent default route and all was well.

This makes zero sense to me - even though the interface was up, OSPF was down, and there was no default route, so all the routers on the network should have learnt the default route from R5.

We are running 6.47.7. Is this some kind of bug we have run into, because it's certainly not usual OSPF behaviour.

Thanks.

Who is online

Users browsing this forum: loloski and 11 guests