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

Re: [K12OSN] k12ltsp release future?



On 6 Oct 2003, Steve Wright wrote:

>On Mon, 2003-10-06 at 00:51, Thomas E. Haynes wrote:
>> The fedora core releases are going to make it tougher to use Redhat in an
>> educational environment. We have the same sort of security issues that
>> anyone else does. I understand that the frequent upgrades are a cost of
>> doing business for users in education that don't have a budget, and I
>> realize that these are economic realities, etc. Sometimes when you don't
>> have a budget, you're not properly staffed either and you get smacked twice.
>> There are lots of cold realities here.
>
>
>I do not understand this.  I regularly apt-get upgrade, and I have the
>latest of everything.  This does not cause me problems.  "Frequent
>upgrades" are not a cost for me at all.  They are free, totally
>automated, and 100% successful (except for broken rpm - not package
>related.)
>
>Why is this a bad thing for K12LTSP ?

As I just posted a few minutes ago, I agree with Steve's assessment here
that tools like apt-get/up2date/yum make it pretty easy to keep up with
the latest upgrades and security patches - even with a large number of
servers.

The Red Hat Network makes it much, much easier to maintain a large
number of servers, but if you don't have the budget apt-get/yum are
easy enough.


>> I think probably the list/developers should consider a move to ES for
>> schools that can afford it and let everyone else install the "taroon"
>> version for the demo or small lab environment on a shoe string. The cost of
>> ES will put some people out of the game, but that was one of the things that
>> RH knew would happen when they went to the fedora/enterprise model.
>
>We basically *have* ES.  By the time you have SAMBA + LDAP on top of
>K12LTSP - it is pretty well "enterprise ready".
>
>Why would R/H ES be better than what we have now ?

Here I agree with Tom. K12LTSP is not a "enterprise ready" solution, at
least not by most enterprise's definition.

I run K12LTSP 2.1.2 and 3.1.2 on about 50 servers (not terminal servers,
but regular file/mail/proxy/etc servers). We are evaluating using Linux
for some of our mission critical databases - there we looking at Red 
Hat Enterprise Linux, not K12LTSP. $2,500/server/year is a HUGE BARGAIN
in some circumstances ;-)

There is also an educational deal of some sort that offers the low-end
RHEL for schools at $25/box/year. But in lots of 100. One of our peer
agencies have approached me about lumping our servers together to make
the lots-of-100 easy to swallow. We may do this for our more critical
file and mail servers.

>> Another option that needs to be seriously considered is a move to Debian
>> perhaps for the k12ltsp base. This is a somewhat radical suggestion given
>> the host for this list, but practical realities may require it. If the initi
>> al cost, complexity, and security issues scare folks off, then we don't have
>> the user base (or the ltsp labs) to really make much of an impact.
>
>
>But Why ?  What practical realities ?  What initial cost?  Which
>security issues ?
>
>
>sorry, but I am not following you, Tom.

The community support of the Fedora project is yet to be demonstrated.
If it works out reasonably well, moving to Debian would not buy us
much - Debian is completely community supported as well.

The Fedora project will have to tank pretty hard before I would make
the jump to Debian.  I'm a big fan of Debian, don't get me wrong, but
it is simply not geared to be an "appliance" system that we're trying
to accomplish with K12LTSP. If I handed a high school CS teacher, who
has never even seen Linux before, a Debian boot disk and told them
to GO FOR IT without any further instructions - I would assume that
they would fail to get it installed. With K12LTSP, I would assume that
they would end up with a working server. Debian is great, but it just
doesn't fit well into this niche as-is.

-Eric




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