MAE-east weirdness

Karl Denninger, MCSNet karl at mcs.com
Mon May 8 20:56:23 UTC 1995


 
> Did this happen on Friday?  If so I heard from a directly knowledgeable 
> source that there was a power failure in Dallas that affected the MFS pop 
> there and that MFS had to bring in a truck with generators to get the 
> node operational again.  If this is correct wold the looping be really 
> blamable to either UUNET or Sprint?
> 
> ********************************************************************
> Gordon Cook, Editor & Publisher       Subscript.: Individ-ascii  $85

The trace shows bouncing between sites which are both local to Washington
DC.  I don't see how Dallas comes into this, unless at least one site 
involved is pointing default somewhere (which could cause this).  My
understanding is that both Sprint and Alternet run defaultless (as do 
many multihomed providers who have a reasonable route table such as MCS.)

Discontinuities in a defaultless network do not show up as route loops -- 
they show up as !Hs or black-holes (ie: no returns at all) somewhere along 
the line, depending on ICMP configuration issues.  You *can* get them if one 
or more of the routers in the game are pointing default (or have a 
less-specific CIDR route) in a way which can be looped back.  This is an 
architectual problem which needs to be found and fixed -- looping packets 
will destroy the efficiency of your circuits (consider the TTL of most 
packets and the number of bounces before they TTL out).

Hard-wired and "pulled-up" routes can also cause this kind of silliness.

--
--
Karl Denninger (karl at MCS.Net)| MCSNet - The Finest Internet Connectivity
Modem: [+1 312 248-0900]     | (shell, PPP, SLIP, leased) in Chicagoland
Voice: [+1 312 248-8649]     | 7 POPs online through Chicago, all 28.8
Fax: [+1 312 248-9865]       | Email to "info at mcs.net" for more information
ISDN: Surf at Smokin' Speed  | WWW: http://www.mcs.net, gopher: gopher.mcs.net



More information about the NANOG mailing list