Martin Buck

I've got a preliminary bugfix and an updated topotest checking for both multiple ABR and single ABR with multiple paths, both for FRR 8.4.3: https://github.com/gromit1811/frr/tree/ospf6_ecmp_inter_area_bugfix_15777_test_merged_frr843 Problem: The updated topotest fails...

Rebased to current master: https://github.com/gromit1811/frr/tree/ospf6_ecmp_inter_area_bugfix_15777_test_merged Here, the topotest runs reliably, so the missing nexthops mentioned before must be caused by another bug in 8.4.3 which has been fixed in the...

I wanted to give the original reporter a chance to test the fix before opening a PR. But I've not heard anything from him, so I'll do the PR now.

> I can confirm the bug in version 10.0 too... @DS819src Can you give my fix linked above a try? I'd assume it should fix the problem you're seeing as...

Weird, then I don't know what could be causing this. I guess this is 100% reproducible and doesn't depend on the order that links/routers come up? Can you provide a...

@DS819src I had a closer look and I don't really understand what's going on :confused: First I added interface names to your diagram, hope they're correct: ``` Area 0.0.0.0 +----------+gi1...

It's "lo" by default: ``` >>> conf.iface >>> ``` If I set it to "veth0a", the warnings disappear and I get proper source addresses for the MLD packet.

Thanks for your explanation! And thanks a lot for such a useful tool, BTW! I wasn't aware of the different purposes of `conf.iface` and `sendp(iface=XXX)`. It's clear that the `sendp()`...

So, looking at the route towards `2001:db8:10:0:203:0:113:10/128`: ``` Destination: 2001:db8:10:0:203:0:113:10/128 [...] LS Origin: Inter-Prefix Id: 0.0.0.3 Adv: 203.0.113.3 [...] Metric: 30 (0) Paths count: 2 Nexthop count: 1 Nexthop: fe80::402b:76ff:fe56:4ed3...

Thanks for the script! With FRR 10.0 and my bugfix applied, I got the expected ECMP nexthops, but it was lacking some routes completely which were there without the fix....