On 2011-05-24, Chris Wopat <m...@falz.net> wrote: > On Tue, May 24, 2011 at 3:11 PM, Claudio Jeker <cje...@diehard.n-r-g.com> > wrote: >> >> Are you running 4.9 or -current? Up until the code generating the LSA >> update packets (and sending them) did not change between 4.8 and 4.9. >> In -current this code got rewritten to fix a issue. IIRC the problem was >> that an LS Update got so big that it did not fit into a MTU sized packet. >> If my memory serves me right then the result was this kind of packet >> storm. >> >> You should try and compile a -current ospfd on your 4.9 system. I think it >> should run without any problems. > > I'm running 4.9, not -current. I didn't see anything about people > having this issue in the mailing lists.
If this is related to sending huge LS updates, I don't think many people currently running ospfd would hit it (you'd need to be announcing quite a lot of networks into ospf), so you probably wouldn't have read about it on the lists. > I see that there was a patch > applied only about 15 minutes ago to something else related to ospfd, > I'm not sure what new issues will occur if I go to -current here since > this box is in production. You have a confirmed issue now, so -current isn't likely to make things worse. (With daemons like this I'm usually happier running -current in production than older code). > * Can a reliability fix errata be created for this? If compiling -current ospfd on your system fixes the problem, it might be worth picking out the relevant commits and applying them to -stable but that is more risky than just running -current ospfd. > * Is the same bug in ospf6d? It looks like this probably is the case.
