OSPF stuck in 2-way state

[Kenth Eriksson]

On Thu, 2019-10-03 at 16:46 +0200, Ondrej Zajicek wrote:
> CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.
> 
> 
> On Thu, Oct 03, 2019 at 01:49:23PM +0000, Kenth Eriksson wrote:
> > Hi!
> > 
> > We are having a problem were sometimes bird ospf neighbor state machine
> > becomes stuck in a 2-way state. This happens when two broadcast
> > interfaces are connected, one of them is running quagga and the other
> > end is running bird. The quagga router has prio 0 and bird prio 5, i.e.
> > quagga is not eligible to become DR.
> > 
> > Why does not bird transition from 2-way to ExStart?
> > 
> > I believe the following lines causes me.dr to become 0.0.0.0. Bird adds
> > itself as an eligible router as described in section 10.4, but with the
> > address of zero?
> > 
> >   me.dr  = ospf_is_v2(p) ? ipa_to_u32(ifa->drip) : ifa->drid;
> >   me.bdr = ospf_is_v2(p) ? ipa_to_u32(ifa->bdrip) : ifa->bdrid;
> 
> Hi
> 
> I do not get why do you think that it is added with the address of zero.
> There is a line above specifying local address is used:
> 
>   me.ip = ifa->addr->ip;
> 
me.dr is 0 because ifa->drip is 0, but should it have declared itself
as an eligible DR? 
> 
> It seems to me that quagga elected bird, but bird for some reason does
> not elect itself.
> 
Correct. 
> 
> How often does it happen in that configuration? The election should be
> relatively deterministic, it should happen just on order of wait timers
> expiration and hello packet reception.
> 
This does not happen every time. Not sure at this point how frequent it
is. But when it happens it is stuck there forever.
> --
> Elen sila lumenn' omentielvo
> 
> Ondrej 'Santiago' Zajicek (email: santiago at 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."