[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [school-discuss] Re: Karoshi, anyone?

On Mon, 2006-07-03 at 19:57 -0700, lee wrote:
> Daniel Howard <dhhoward@xxxxxxxxxxx> wrote: 
> > Good stuff Lee, I'll look into it and see if it could help our
> pilot 
> > project to roll LTSP out to 6 other schools here in Atlanta.

I am looking at the Karoshi stuff now. It looks very useful. I'll need
to "Peek under the hood" for details on things like server naming,
> Neat... well I don't know much enuf about how LTSP is structured but
> I'm assuming that integrating r-mounted /home volumes along w/ NIS/yp
> into Karoshi's server schema wouldn't be too hard.
> My only complaint about Karoshi at the time I installed it was that
> the server names were fixed (nothing that enhanced shell scripts
> couldn't change if there were a customer need...) and that there was a
> minor (but easily allayed) dependency upon a local NTP server getting
> universal time from a local IPcop firewall (I rigged my network to
> lookup on the world time service since I wanted to use a LinkSys SOHO
> firewall/router instead of IPCop... no big deal).
> Karoshi's installed Windows "user profiles" use a UK keyboard & UK
> settings, so one would have to build US-oriented workstation profiles
> & place them in the correct shared directories on the Linux servers.
> > And I 
> > agree completely with the comment about coupling Linux experts ('a 
> > senior Unix admin') with schools moving or considering Linux; this
> is 
> > exactly what we did here in Atlanta, and our local expert, Jim
> Kinney, 
> > is really helping them not only with transitioning to Linux, but
> also 
> > with getting out of the box of buying brand names, etc. 
> Well there you go... also it'll be important for doing a pilot the
> most-critical aspect is proving disaster recovery, incl.
> proof-of-concept backup & restore, etc. This is what bugs IT managers
> most is interruption of service & QOS. Once a school becomes dependant
> upon an online classroom system, for instance, then the expectation of
> 24x7 follows. Giving them the warm fuzzies is critical....

For many places (this pilot included) just having working machines is
such a leap forward they haven't even considered backups yet (we have,
> > But it's to the 
> > point for me now, where if any other school asked me about moving
> to 
> > Linux, I would: 1) give them a tour of our school to see what a 
> > difference it makes, 2) if they don't already have a *nix
> admin/expert, 
> > then identify a local *nix admin/expert in their city that can help
> them 
> > plan and convert in stages. 
> One idea I've been cooking in my head has been a demo cart that I can
> roll into a school... basically buying a bunch o' laptops from the
> refurbished outlet or gov't surplus sites ($500 for 20 laptops w/out
> harddrives...), building a demo Karoshi network with them & rolling
> the whole kit in already-running condition. Maybe lease the cart as a
> lease-loan kit for them to kick the tires, do feasibility studies,
> integrate some of their own desktops into the kit to see how well
> it'll work with their existing hardware, etc., taking all the
> time/risk worries out of the equation, plus it's there right in front
> of their eyes...  Maybe do a point & click disaster recovery demo ...
> wreck a server with rm -rf from root, show them the turn-around
> procedure to get the shop back online...

The demo cart is a great idea. The recover-ability will be a discussion
that is best reserved for techs of the system. The thing that gets the
teachers attention is how readily available it is. With a K12LTSP setup
and a server on battery (so it can be moved from room to room and
replugged) the teachers eyes pop out when they see how fast things go
from power on to login screen! 
> Just a thought... if a demo cart sounds like overkill, there's always
> bringing in just one or two laptops running multiple Linux servers on
> UML Linux "VM" instances (which alone has a very major "cool" factor).
> One great wonder of UML Linux, BTW, is the ability to mirror Linux
> instance images & then (if I recall correctly) provide a
> failover-capable set of servers (not true HACMP clustering, just HA
> fail-over clustering). I don't know if anyone's doing this yet, but it
> oughtta be feasible & probably a value-added $ feature one could
> provide as a wrap-around for a Karoshi setup.... heh... all these
> groovy things could be boiler-plated given the ability to write
> effective shell scripts in Linux...
> > That's why I'm excited about having a 
> > really big list of consultants, I really think that is a key part
> of 
> > helping existing school IT departments make the transition.
> > And we learned in our case that they really preferred someone 
> > local that they could call up on a moments notice, meet with
> frequently, 
> > see their face, etc. without the associated travel costs.
> I ran into this w/ my former school when my family decided to
> relocate... My replacement was not nearly so technical, so
> understandably mgm't sought expertise. Unfortunately they fell back on
> their fav Windoze vendor who was expensive and marginally
> competent ... So a good, qualified Linux support admin is going to be
> important. 

One thing that needs to be understood by the management types is unlike
windows, the Linux platform is designed for remote maintenance. I have
pssh on my palm treo and can get to any of my supported systems and
"kick things" if I need to. There are many things that just can't be
done in the windows world remotely without significant expense.
> What I learned when I started off on getting Linux servers in the
> school was that I wasn't going to get rid of Windows on the users'
> desktops. There's a good reason for the end-users to want Windows: In
> K-6 there's Reading Blaster, Math Blaster, the teacher's fav gradebook
> program, etc. But if the schools are well-managed, they've retained
> all the licenses for their machines, so the existing Windows machines'
> licenses are already sunk cost. 

There are also so many schools that have win95, win98, winME systems
that can't ever be used for Internet again due to security issues and
can't be upgraded to XP for hardware issues.  It boils down to a
cost/benefit analysis that most schools don't have the resources to do.

I have managed to get several windows98 edutainment games to run under
> Some of the loss-lead development costs may be producing manuals, case
> study brochures,  maybe a test lab (or cart) that the consultants
> could quickly boilerplate for a potential client school (ala a "one
> day seminar" with snacks, food, glossy propaganda...), etc. 

Actually, things like manuals are viewed better as high-value items. As
the community develops better documentation that production cost gets
spread over a wider base. The United Nations put out an excellent manual
on how to use Linux several years ago. They released it in both PDF and
OpenOffice formats.
> There are some Microsoft licensing problems that some local gov'ts
> have faced that entails MS marketing strong-arm tactics ... some
> municipality trying to stray away from the MS fold have found MS
> marketeers threaten to bring in their license cops in to examine CAL &
> COA discrepancies. All this hinges on whether the local gov'ts have
> kept good records, kept their COAs, etc. But for schools, the outlook
> is better, b/c if they took in donor machines w/out licenses, the the
> MS FreshStart program can rectify the discrepancies (for smaller
> school districts or indep. schools?) for PIII-class machines or lower.

I view this as the main argument for moving away from the Microsoft
systems as fast as possible. The strong arm/Gestapo tactics used by the
SBA have encouraged many, many schools to by needless Microsoft licenses
in extra capacity as "insurance" against legal action. The legal teams
at the school systems need to take a much closer look at the use of FOSS
systems as an alternative to outrageous court expenses brought on by M$
and others working to milk the resources away from teaching.
> /lee
> ======
> /lee
> +-----------------------------------+
> | This concludes our broadcast day |
> +-----------------------------------+
James P. Kinney III          \Changing the mobile computing world/
CEO & Director of Engineering \          one Linux user         /
Local Net Solutions,LLC        \           at a time.          /
770-493-8244                    \.___________________________./

GPG ID: 829C6CA7 James P. Kinney III (M.S. Physics)
Fingerprint = 3C9E 6366 54FC A3FE BA4D 0659 6190 ADC3 829C 6CA7

Attachment: signature.asc
Description: This is a digitally signed message part