On Tue, 2006-06-20 at 07:37 -0400, G. Matthew Rice wrote:
> I had listed AFS when we were talking but it got "marketed" out.  Same with
> Kerberos :(  It seems that some enterprise clients just don't use it.  I'm a
> little surprised but...

The problem with AFS is that it's standalone.  Unlike NFS, Samba and
many other RPC file services that work on local files, AFS uses a
virtual filesystem.  So anything and everything is AFS-only, including
its own access controls, permissions, authorization, etc...  NFS, Samba
and other services leverage the local filesystem and all its existing
meta-data (plus whatever, additional emulation is added for Samba).  So
what you cover for one of them is very applicable to another.

> Actually, let me write down the main points from the research:
>    - linux in a mission critical and mixed environment

Exactomundo.  ;->

>    - core network services
>    - Samba, NFS, LDAP

Yeah, I see two exams out of those first 3.  Most people are
implementing basic auth/dir/name just for basic services on their
network.  Most enterprises are not using LDAP for more complex
schema/store configurations than that, which means their use of LDAP is
typically intertwined with other core network services.  Again, basic
auth/dir/name for the entire network.

>    - capacity planning

That would very much go into the "Availability and Redundancy" domain.

> these were the areas of of interest for the first exam.  I agree that the
> auth vs. file/print makes a sensible breakdown.


> You got it.  All of that should be in the 'LDAP' exam.
> Perhaps, we should rename the pages to the more generic titles of 'auth' and
> 'file/print', ... to help avoid the confusion?

Or just use subtitles, like ...

- LDAP, authentication and naming
- Samba and network file services

> It just seems that a few of you are starting at opposite ends.  You are
> approaching it from the the top-down and others want to start getting at the
> details.  We need a balance between the two.

I've got details, trust me.  I'm not going to make this broad/generic.
There are only so many ways to do something.

I just don't want it to be only 1-2 ways.  ;->

> You keep working on the generic and others can do the details.  There's room
> for both.

Oh, again, I can give specifics on how things work -- the exact
services.  Those can directly lead into tasks.

I think I've been covering how specific Samba services tie into the
greater role of the enterprise auth/dir/name.  I have done the same for
NFS.

That's why the ELResource wiki design is concept, practice then task.

> Well, they should be put in.  This exam is part 'the ideal Linux-centric'
> setup but it's also part 'what is common practice today'.
> And I think that I mentioned the JTA helping with the weighting of importance
> on these tasks.

> Sounds okay to me.  The first two are our focus this year.  I think that
> security makes a smart focus next year.  Although, I'm hearing rumblings of
> some embedded stuff being talked about.

Oh, well, I can help there.

In fact I'm doing a phone screen today with a client in New England to
do some embedded work for the financial industry (yeah, I get around).
My most recent, permanent job was embedded Linux work too -- targeting
ARM/X-Scale and other, non-x86 platforms.  And most of my clients (and
then employers before that) have either been defense or finance.

[ I'm really an old VxWorks software engineer from the missile defense
industry of the mid-'90s.  Linux/Solaris were the popular server/host
platforms for GNU/VxWorks. ]

Maybe it's because I only spent 3 months working the ISP/ASP arena (and
learned to get the hell out of that penny-pinching industry), but I'm
used to LAN/WAN/Intranet environments.

> I should also point out that a major focus on the LPIC-3 development is to
> build up the tools to make exam development easier.  That means we could
> actually start oodles of exams in various areas.  The main problem will be
> cost of publishing with the test networks (let alone that whole hands-on
> issue ;).

Hey, I'm all for hands-on testing.  When someone shows me its feasible
for LPI to do.  ;->


-- 
Bryan J. Smith           Professional, technical annoyance
mailto:[EMAIL PROTECTED]     http://thebs413.blogspot.com
----------------------------------------------------------
The existence of Linux has far more to do with the breakup
of AT&T's monopoly than anything Microsoft has ever done.


_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/mailman/listinfo/lpi-examdev

Reply via email to