Interesting. I wonder if we're discussing two separate events. The one I knew of supposedly involved a regex with a misplaced .*, which seems like a fairly fundamentally different lesson. Both valuble, though...
I was downstream from 2551 at the time of their, um, big event. A botched regex that led to external BGP routes being redistributed into OSPF is what sources that I considered reliable explained to me as the cause. Reports on the length of the outage varied, but I think official reports were from the time of the initial failure until their core was stable, because my network was down well over 24 hours, and I seem to recall the "official" times as shorter. As for the previous question in this thread, 12.1 that I'm currently running treats "no red bgp subn route-m <name>" the same as "no red bgp" (for any value of <name>, including values that don't match what's already there). I have no reason to believe that the 11.x that Netcom was presumably running at the time was any different. -- Brett