Ceci est pas directly a DNS problem, c'est a network routing problem entre certains parts of le internet et le DNS servers for serverfault.com. Since le nameservers ne peut pas be reached le domain stops resolving.
As far as I can tell le routing problem is on le (Global Crossing?) router avec IP address 204.245.39.50.
As shown by @radius, packets to ns52 (as used by stackoverflow.com) pass depuis here to 208.109.115.121 et depuis there work correctly. However packets to ns22 go à la place to 208.109.115.201.
Since those two addresses are les deux in le même /24 et le corresponding BGP announcement is aussi for a /24 this ne devrait pas happen.
J'ai done traceroutes via mon network qui ultimately uses MFN Above.net à la place of Global Crossing to get to GoDaddy et il y a no sign of tout routing trickery ci-dessous le /24 level - les deux name servers have identical traceroutes depuis here.
The seulement times J'ai ever seen something like this it was broken Cisco Express Forwarding (CEF). Ceci est a hardware level cache used to accelerate packet routing. Unfortunately juste occasionally it gets out of sync avec le real routing table, et tries to forward packets via le wrong interface. CEF entries can go down to le /32 level even si le underlying routing table entry is for a /24. C'est tricky to find these sorts of problems, mais once identified they're normally easy to fix.
J'ai e-mailed GC et aussi tried to speak to them, mais they ne va pas créez un ticket for non-customers. If tout of you are a customer of GC, please try et report this...
UPDATE at 10:38 UTC As Jeff has noted le problem has now cleared. Traceroutes to les deux servers mentioned ci-dessus now go via le 208.109.115.121 suivant hop.
