Questions After a DNS Server Crash

Martin McCormick martin at
Tue Jun 28 14:09:38 UTC 2005

Brad Knowles writes:
>	This is one of the reasons I like putting things behind Layer 4 
>load balancing switches, and having those switch IP addresses be the 
>ones that get published.  That way, when a machine behind the switch 
>dies, you might get an e-mail about the problem when you come into 
>the office in the morning, but everything else should "just work", 
>and the traffic should automatically get routed to other servers, 

	Thanks to all who responded with thoughts on this and on
update configuration.  I will definitely look in to the use of
load-balancing switches in this context because there is fundamentally
something scary about single-point failure modes.  No matter how good
the software or hardware, there is that capacitor in the power supply
with a pin hole developing in its dialectric, that buffer that hasn't
filled to capacity yet, but today at 13:30, it will for some reason,
etc.  Redundancy is how nature protects itself when possible so it
makes since to me.

Martin McCormick WB5AGZ  Stillwater, OK 
OSU Information Technology Division Network Operations Group

More information about the bind-users mailing list