*//*
Matt Oquist <moquist@xxxxxxxxx> wrote:
I agree with you, but sadly this doesn't always happen. School IT
departments (a big one in my area might have 5 people) stay just the
way they are, and the former Winders(TM) admins suddenly become Linux
admins. A "good" Linux admin can manage 50-100 servers, but these
aren't "good" Linux admins (yet); many of them get upset because they
can't administer everything with a GUI anymore.
I'm an old Unix DBMS guy, so I've done my share of awk & ksh scripts,
etc., but anything that can alleviate some of the complexity of Linux
network admin is hugely welcome...
I've mentioned it before, when I took a close look at Karoshi
http://www.karoshi.org.uk I found a wonderful & underappreciated
effort... What Paul & Jo have done is make the adoption of a complete
multi-server SAMBA/NIS Linux network easy for GUI-driven users. They
did this using little more than BASH scripts w/ a GUI frontend.
Karoshi fully integrates up to 11 servers including two backup servers,
an on-line classroom & e-mail server, a Proxy filter (Dans Guardian) &
internet access server w/ room-by-room blocking, a primary domain
controller, at least three file servers (for both Windows & Linux
roaming and mandatory profiles), an on-line library system, a media
server, etc...
> But this transition is something that must be understood before it is
> undertaken, and I worry that many schools don't take the required
> changes in personnel (either the existing personnel must change, or
> new personnel must be hired) sufficiently into account.
The answer in this case is to pilot a Linux network - first on the
bench, then in a semi-live demo....
The most-complex part of setting up Karoshi, for instance, was
installing Mandrake (now Mandriva) - which is no harder than installing
Windows. I believe that Mandrake & Karoshi combined takes most of the
QOS risk out of Linux adoption (relative to Windows "ease-of-use") and
the payoff in terms of costs will be immediately obvious -- no more
server CALs plus a large bevvy of really wonderful free systems like
Dans Guardian, Moodle, etc. A full Karoshi Install represents easily
$40,000 of equivalent commercial software, and doesn't require stacks
and stacks of the hottest, latest "blades" to run it....
> Even so, as we all work to get the world to recognize the viability
> and benefits of Linux and FOSS in education, we can change these
> perceptions and thereby increase the number of "good" Linux admins in
> the long-term.
Hmmm.... well, "good admins" of any flavor are more likely to be good on
any platform, even if they are Windows-heads, they're also working with
Wifi security, WEP, DNS, NAT, bridging, subnets, etc., so they're
probably ready to take on Linux given the opportunity. School districts
have a pay problem, but the OTJ stress is lower & attracts older IT
workers who want school district retirement & insurance and 40-hour
weeks w/ calm summers. Altho I've seen some Windows admins (at a state
agency) shy away from Unix & Linux, most were up for the challenge given
that they had good over-the-shoulder support from a senior Unix admin....
What IT manglers seek is an available knowledge-base that can be
"transferred in" OR is easy enough to adopt. Windows networks, being of
critical mass fits the former criterion, and a GUI-driven system like
Karoshi fits the latter. Were I an IT manager and I saw the differential
cost savings by going with an integrated Linux server suite like
Karoshi, then the adoption gains far outweigh all the other adoption
issues....
One thing I suggested to the Karoshi developers was that they consider
adding VPN &/or LTSP functionality -- I believe they are considering it
as part of a later phase.... To me that would make Karoshi a true TKO
solution - and vice versa for K12LTSP -- K12LTSP integrated into Karoshi
would push K12LTSP from a narrow LTSP-only solution (not that is a small
thing, per se, but only one part of the overall K-12 IT picture) to
being part of a bigger, broader integrated solution that'd bring solve a
great many problems in one swell foop.