nanog mailing list archives

Re: Teaching/developing troubleshooting skills


From: Eric Brunner-Williams <brunner () nic-naa net>
Date: Mon, 28 Jun 2004 15:30:35 +0000


* The faulty assumption there is but one problem
* Incorrectly-formed causal relationships

Mythology.

Some may recall the adventures of the CTO who ran a sweep of an net 10.*
in a rather modest machine room somewhere in Maine, resulting in memory
exhaustion (arp table) in the router, which resulted in 1918 leakage into
public address space.

The operational mythology of the ever-so-security-minded-yolks was that
the initial and very poorly understood presenting problem was an external
act of malice, rather than self-inflicted DoS by the security-yolk itself.

I've seen many people struggle to fit what little they know into predefined
mythos of what could be happening, rather than starting like Sgt. Schultz,
who "knew nothing", at least until he really _knew_ it.

Eric


Current thread: