-- Warren Kumari Sent with Sparrow (http://www.sparrowmailapp.com/?sig) On Friday, October 5, 2012 at 2:29 PM, ryanL wrote:
so, redistributed eBGP into IGP? how does one 210x their routing table? i appreciate the openness, but it doesn't sound like a "perfect storm" to me. it sounds like someone screwed up.
Could be someone screwed up, could have been caused by a bug. Your assumption that this was redistribution of eBGP -> IGP is not really backed up by the text. if I had to guess, I'd suppose that they have devices (like cisco 4900 series) that don't have enough memory / FIB / TCAM for full tables. Normally they only send partial routes (such as internal only / large aggregates / something) and accidentally or though a bug dropped filter and overflowed these devices… But, whatever the case, GoDaddy has released a somewhat reasonable post-mortem, sounds like they are / have addressed the issue and os it shouldn't happen again, etc. As many know, I'm not a real of GoDaddy, but credit where it is due… Running a large network is non-trivial, sometimes s%#@t happens -- at least GoDaddy was willing to publish something and not just hand wave this away… We keep kvetching that folk never publish postmortems when they have an outage -- lets not attack them when they do… W
On Fri, Oct 5, 2012 at 10:18 AM, Jason LeBlanc <jleblanc@godaddy.com (mailto:jleblanc@godaddy.com)> wrote:
I made a commitment to get everyone on here the RFO. Its presented in layman's terms:
http://inside.godaddy.com/inside-story-happened-godaddy-com-sept-10-2012/
Regards,
Jason LeBlanc Network Engineering (480)505-8800 x7202 (tel:%28480%29505-8800%20x7202)
Please contact my direct supervisor mdob@godaddy.com (mailto:mdob@godaddy.com) with any feedback.
Outages mailing list Outages@outages.org (mailto:Outages@outages.org) https://puck.nether.net/mailman/listinfo/outages