Teaching/developing troubleshooting skills

I'm working on trying to teach others in my group (usually
less-experienced, but not always) how to improve their
large-network troubleshooting skills (the techniques of
isolating a problem, etc).

It's been so long since I learned network troubleshooting
techniques I can't remember how I learned them or even how I
used to do it (so poorly).

Does anyone have experience with developing a
skills-improvement program on this topic? If you've tried
such a thing, what worked/didn't work for you? Outside
training? Books? Mentoring? Motivational posters?

I'm particularly sensitive to the "I got my CCNA, therefore
I know everything there is to know about troubleshooting"
perspective, and how to encourage improving troubleshooting
skills without making it insultingly basic.

Thanks for your help.
Pete.

Pete Kruckenberg wrote:

I'm working on trying to teach others in my group (usually
less-experienced, but not always) how to improve their
large-network troubleshooting skills (the techniques of
isolating a problem, etc).

There are several vendors that offer these types of courses, and I am sure that if you search for courseware, you can find some good materials you could use to teach your own sessions in house.

Jon

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Pete Kruckenberg wrote:

I'm working on trying to teach others in my group (usually
less-experienced, but not always) how to improve their
large-network troubleshooting skills (the techniques of
isolating a problem, etc).

It's been so long since I learned network troubleshooting
techniques I can't remember how I learned them or even how I
used to do it (so poorly).

Does anyone have experience with developing a
skills-improvement program on this topic? If you've tried
such a thing, what worked/didn't work for you? Outside
training? Books? Mentoring? Motivational posters?

I'm particularly sensitive to the "I got my CCNA, therefore
I know everything there is to know about troubleshooting"
perspective, and how to encourage improving troubleshooting
skills without making it insultingly basic.

If you are looking for some courses on just analytical troubleshooting
and/or problem solving techniques, you might want to look at the Kepner
Tregoe stuff (www.kepner-tregoe.com). It is not network specific but
rather teaches techniques. Some of their courses include:

  Problem Solving and Decision Making
  Analytic Trouble Shooting
  Implementing Corrective and Preventive Actions

- --

On 04/6/24 at 5:09 PM -0600, Pete Kruckenberg wrote the following :

I'm working on trying to teach others in my group (usually
less-experienced, but not always) how to improve their
large-network troubleshooting skills (the techniques of
isolating a problem, etc)

I took a 5 day course in another era, fortunately for me at the
beginning of my career, on analytic trouble shooting (Kepner Tregoe?).

The 5 day course can be boiled down really to one concept that can be
taught in 5 minutes... 'half-splitting'. (The other 4.95 days were
spent making sure we understood the concept and learning to implement
it, the length of time was overkill but the course vendor had to make
money somehow :slight_smile:

(In another discussion* it was pointed out to that that wasn't the
correct name in the writers view... it was "binary search". Google
may have proved the writer correct, but I still refer to it as half
splitting as I spent a week learning to call it that :slight_smile:

The point of this note is troubleshooting boils down finding the
problem in the fewest steps.

Half-splitting ensures the number of steps are at a minimum.

The troubleshooters knowledge of the system and equipment provides
the ability to devise tests at the half-splitting point.
Half-splitting is a 5 minute concept. System/equipment knowledge is
of course a lifetime endeavor.

The reason I am writing this note is as I went through a career of
troubleshooting I was surprised at the number of colleagues who had
no concept of "half-splitting" and used "linear" or "random"
techniques to determine test points/tests with a corresponding
dramatic reduction in effectiveness.

Cheers,

Darrell

*p.s., I just remembered where my previous discussion was;

http://db.tidbits.com/tbtalk/tlkmsg.lasso?MsgID=15775

http://db.tidbits.com/tbtalk/tlkmsg.lasso?MsgID=15787

http://db.tidbits.com/tbtalk/tlkmsg.lasso?MsgID=15788

Searching with Google for "half-splitting" will produce some useful hits.

Date: Fri, 25 Jun 2004 20:04:38 -0700
From: Darrell Greenwood

[ editted for brevity ]

The 5 day course can be boiled down really to one concept
that can be taught in 5 minutes... "binary search".

Every half-decent programmer knows O(log(N)) is one's friend
unless the scalar coefficient is large. A good way to
demonstrate its efficiency is:

* Have someone pick an integer between 1 and n, inclusive
* Make guesses, going "higher" or "lower" according to the
  number-holder's feedback.

The uninformed are surprised that one can always guess the number
from 1 to 1000 in ten iterations or less.

The reason I am writing this note is as I went through a
career of troubleshooting I was surprised at the number of
colleagues who had no concept of "half-splitting" and used
"linear" or "random" techniques to determine test
points/tests with a corresponding dramatic reduction in
effectiveness.

Good point.

[ below text in response to nobody in particular ]

It's also important that one avoid:

* The faulty assumption there is but one problem
* Incorrectly-formed causal relationships (NANOG-L has some
  examples of these)
* Making too many changes in one iteration
* Attempting to tackle a system with more unknowns than are
  absolutely necessary.

A certain amount of troubleshooting can be taught, but IMHO it
requires a self-driven person with intuitive reasoning.

Finally: Apprenticeship. Have the novices follow along when
experts work actual cases. A certain amount of troubleshooting
is developing the intuition to make informed guesses -- e.g.,
"some idiot broke pmtud" -- and develop good leads without having
to search methodically through the entire problem space.

Eddy