Long e-mail to follow on the topic of certification pros and cons.
I agree with all points of Giannis. It's a Network Administrators job
to learn commonly used commands and associated switches and arguments.
Every once in a while, I have to interview candidates for IT positions
at the company I work for. The first thing I do, is hand the candidate
a test with questions that deal closely with the job they are applying
for. Usually about 10 questions or so. Some of the questions require
written answers, and others require short commands. Here are some examples.
What is the difference between a routing protocol and a routed
protocol? Give examples of each.
Please give the absolute paths to the folders where the following types
of files would be found on a Linux Computer: 1) Kernel 2) Logs 3)
Global Configuration Files
Please describe, in whatever manor you feel comfortable, how DNS works.
These kinds of questions are designed to ensure the applicant knows the
basics, and ensures that he/she will be able to learn more advanced
topics. Using their answers as a baseline, certain assumptions can be
made. If the applicant missed the routing question above, then I could
assume they have very little, if any, routing knowledge, and I need not
ask any advanced questions about it. However, if they answered the DNS
question correctly, then I would ask much more advanced/detailed
questions. Such as "What command would you use to find the MX record
associated with a given domain?"
This kind of questioning is really designed to traverse the knowledge
tree of the candidate. It give the interviewer a reasonably good
understanding of what the candidate knows intimately, and what they know
little about. Standardized tests, like the LPI, can't really use this
kind of testing, because it is designed without limits, to test
everything that the user knows within given categories of information.
Albeit, they're similar in that the kinds of questions in both forms of
testing are similar.
OK, so let's say after the interviews, I like 2 candidates for the job.
My next step would be to ask both of them to come in at different times,
and perform hands-on tasks. Examples might be: Setup this router to do
xxx, Fix this server, it's not working. And from there, the best
candidate would get the job. It's interesting that this phase of the
interviewing process is like the RHCE exam.
What can I draw from this? It's interesting that my real-world
experiences in hiring Network Administrators have consisted of
candidates passing both LPI-like and RHCE-like tests. So which is
better? IMHO, neither. You can't get to the second interview without
passing the first, and vice versa.
I can't image hiring someone strictly on the hands-on test above, but
for the sake of argument, what if I did. Well, the candidate may have
mastered the router setup and fixed the server, but what about theory,
or base knowledge? In fact, the user might not know what DNS is, or how
it works, or even worse, does he care? The applicant might be an "I
just play with it until it works" kind of person, which IMHO, is a
dangerous personality trait for a Network Administrator to have.
Which brings me to this conclusion: Having just an LPIC-2 or RHCE, is
IMHO, not enough to judge a Network Administrator competent in the work
place. However, an individual with an LPIC-2 and 3 years of real-world,
Network Administrator Experience, might be just the employee you are
looking for.
On a side note, I passed the LPI 201 a few weeks back, and will be
taking the LPI 202 test tomorrow, wish me luck. :-)
Chad
On 5/17/07, Lennart Sorensen <[EMAIL PROTECTED]> wrote:
I agree entirely. They are dumb questions. They are just as bad as
computer courses that expect you to memorize the order of function calls
to certain system calls so you can write perfect source code on a
written test. And will penalize you for missing a ';' somewhere. It
doesn't show if you know what you are doing or not. Being able to fix
it when the compiler tells you it isn't right is at least as important.
I was interview by Google some months ago, for an admin position. They
kept asking me some anal questions for kill parameters and signals and
what does traceroute show when that X specific network problem exists
and stuff like that. I knew most of them but some where quite obscure
and although I consider myself an experienced system administrator, I
didn't know them by heart, since I didn't ofter use them.
After the interview, I finally got it. Yes, we admins can find any
solution to probably any problem by Googling it. However, think about
it. How does a google admin solve an infrastructure problem preventing
google from working in the first place? Yes, experience counts, but
sometimes you really need to know exactly how stuff works inside, even
if it is just boring theory, even if you never use it. Most of the
times our job is infrastructure maintenance, so our dependance to
resources should be as low as possible.
Yes, man pages are always available. However, what do you do when the
damn knoppix cd has bad sectors EXACTLY on the man page you are
looking for?
When I fell in love with linux, one of the first thing I did was to
read every single man page. Although I certainly can't remember the
details, I have at least a vague knowledge of what lies inside the
/usr/bin treasure chest. And that gives me confidence.
That said, although there is room for improvement, I like LPI's
current question approach.
Regards
- Giannis Stoilis
_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev
_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev
