Loopback interface as OSPF stub interface (IPv6)
Tried setting a loopback interface as an IPv6 OSPF stub interface, but this is not picked up as an OSPF interface. Doing the same on IPv4 loopback interface picks it up as an OSPF interface and announcing it as a stub in the area. Why not for IPv6? Thanks, Kenth
On Wed, May 05, 2021 at 02:50:18PM +0000, Kenth Eriksson wrote:
Tried setting a loopback interface as an IPv6 OSPF stub interface, but this is not picked up as an OSPF interface. Doing the same on IPv4 loopback interface picks it up as an OSPF interface and announcing it as a stub in the area. Why not for IPv6?
OSPFv3 uses link-local addresses, while OSPFv2 uses regular addresses. Loopback has regular address (both IPv4 and IPv6) but does not have link-local address, so it is skipped by OSPFv3 iface scanning code. Perhaps it will work if you attach link-local address to it. -- Elen sila lumenn' omentielvo Ondrej 'Santiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
OSPFv3 uses link-local addresses, while OSPFv2 uses regular addresses. Loopback has regular address (both IPv4 and IPv6) but does not have link-local address, so it is skipped by OSPFv3 iface scanning code. Perhaps it will work if you attach link-local address to it.
Correct, when I configured a LL address on the loopback interface it is picked up by bird. But does that make sense? A stub interface does not send packets, so why does it need a LL address?
On Fri, May 07, 2021 at 11:37:03AM +0000, Kenth Eriksson wrote:
OSPFv3 uses link-local addresses, while OSPFv2 uses regular addresses. Loopback has regular address (both IPv4 and IPv6) but does not have link-local address, so it is skipped by OSPFv3 iface scanning code. Perhaps it will work if you attach link-local address to it.
Correct, when I configured a LL address on the loopback interface it is picked up by bird. But does that make sense? A stub interface does not send packets, so why does it need a LL address?
Not really. It is just an unintended result of chosen implementation decisions, same as in unnumbered PtP thread: https://bird.network.cz/pipermail/bird-users/2021-May/015465.html -- Elen sila lumenn' omentielvo Ondrej 'Santiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
Correct, when I configured a LL address on the loopback interface it is picked up by bird. But does that make sense? A stub interface does not send packets, so why does it need a LL address?
Not really. It is just an unintended result of chosen implementation decisions, same as in unnumbered PtP thread:
So should we change the behavior to allow stub interfaces without LL address? Patch? đ
Anno domini 2021 Kenth Eriksson scripsit:
Correct, when I configured a LL address on the loopback interface it is picked up by bird. But does that make sense? A stub interface does not send packets, so why does it need a LL address?
Not really. It is just an unintended result of chosen implementation decisions, same as in unnumbered PtP thread:
So should we change the behavior to allow stub interfaces without LL address? Patch? đ
That would be really helpful and would help to repair the current "violation" of the principle of least surprise :) Are you offering a patch or are you looking for one? :) Best Max
participants (3)
-
Kenth Eriksson -
Maximilian Wilhelm -
Ondrej Zajicek