[vox-if] Installfest Hardware, etc.

[javilk at mall-net.com]

> >> javilk at mall-net.com wrote:
> >>>      I am considering coming to the installfest.  I've been running 
> >>> Linux for 12 years, etc. etc.  No Windows in this house!  But I'm an 
> >>> applications programmer, not a systems administrator.
> > 
> >     A little bit late now...
> 
> Look out for April 13th...

    A thought.

> >     Ubuntu... I run three root sessions, two webmaster sessions, and 
> > four or more login sessions on consoles (plus two to five gooey web 
> > browser sessions) when I'm working.  I hear Ubuntu does not like root 
> > sessions... Anyone who does not allow the user to fry his own machine is 
> > a crook!  It's how we learn...
> 
> Well, it sounds like you know what you are doing.  In as such, I don't 
> know whether you'd experience a substantial gain from the large support 
> base.

     Knowing what you are doing is a hinderance.  It is when you don't 
know what you are doing, that you find shortcuts...

     But then, just about everything I do is a research project.

> And no, I believe you are incorrect in your statement regarding Ubuntu 
> being inadequate.  Well, Ubuntu takes a little bit of encouragement, but 
> not a lot to achieve what you speak of.

      Perhaps if I knew which conflaguration files to ignite...

> >     Ah!  Didn't think of that. Like I say, I'm an applications guy. (She 
> > is my sys, and I admin her, but...) Actually, I think it can... And I do 

> >      How could I set that up?
> 
> There are a number of protocols for ethernet booting.  Mostly they 
> involve the card sending out a broadcast request, then a booting server 

     I have the option on the BIOS.

> >       I tried that. It produces a catatonic machine because all the 
> > custom I/O chips, etc. won't let me talk to the peripherals.  Done this 
> > on several notebooks, I just can't seem to get that to work.  The other 
> 
> It's because your install wasn't minimal enough.  You need to install it 

     Ach.  

> Then from there you do a custom compile of the kernel on the destination 
> machine.
  
    c... c.. Compile???!!!  I mean, simple things I compile.  10k, 50k 
source files.  But the whole dinking thing with non ECC, not even Parity 
checked RAM???  That's on the edge of insane.  

> > method... seems you can't stick (or I don't know how to stick) a boot 
> > sector on a USB mounted drive... I tried.  Won't boot when the drive is 
> > put inside the notebook.
> 
> This is certainly BIOS specific - like the ethernet booting.  Depending 
> on how ridiculous your BIOS decided to be, you may also be able to boot 

     I did some consulting for HP in 95...  could be quite ridiculous!

> >      I can mount the rest of the "cluster" (a group of five to twelve 
> > NFS coupled machines using a common queue to run tasks); but the rest of 
> 
> Don't cluster with NFS, just don't... use something modern like CODA or 
> OpenAFS, you'll see significant performance gains, I promise.

     My "cluster" simply parcels out long tasks to be run on machines 
with duplicate local resources. Simple, no real swapping, etc.  Use to 
be I'd have to build up a dedicated machine to run for days to weeks on 
some of those projects; now, I just chop them up and toss them in a 
queue.  (If I use SuSE, does that make it a Chop Suse cluster?)

> > the cluster can't mount any of it's drives.  (And right now, that 
> > machine is down.  A migration script moved /bin and /sbin ... I'll have 
> > to drive down to the office some night soon to put them back where they 
> > should be... The joys of doing things remotely as Root.  The first time 
> > this happened, I was able to deduce where those were and put them back 
> > before the session link was interrupted. This time, mv won't execute.  
> > Ah well, life in the fast lane.  That's what rescue CD's are for.)
> 
> So you just have NFS on a NAS?  Have you LVM'd over the NFS mounts?  I'm 
> sorry, I still don't get the exact problem here.

    Several machines share drives for low I/O (usually final 
destination) files, served to the web by a low front end machine. Took 
two hours to write.  And if I need to repurpose a machine, no problem!

      As for LVM, No!  I rig things so onle one process is writing to or 
reading from ONE file per drive.  Dramatic speed improvements that way, 
as the head movements are minimal.  Hardware level optimization, simply 
by the number of drives per machine; though you do have to write the 
application yourself...

> >       And a whole bunch of other issues with no TCPIP on the Novell 
> > demo, inadequate tag queue depth in the Fedora version... etc.  If I saw 
> > these guys had a GOOD WORKING version, I'd toss them the cash!
> 
> Are you talking about SUSE?  Really, most of the commercial linuxes are 
> inadequate for various reasons.

     Yes!  But I'd rather plunk down my dollars to let someone else give 
me a STANDARD, tested, Known Good system. Then, if something goes 
strange, and does so ONLY on my machine, I know it must be my software, 
my hardware, not some screw up in the operating system.

      I've fought that war before on lots of systems... Nothing like a 
Known Good Standard! (I've seen enough OS level bugs, etc... standards 
make bug replication easier, and more conclusive as to cause if you 
can't replicate it on another machine.)

> >       I know I really ought to compile my own... but then I'd really get
> 
> Compile your own what? Sorry, I don't understand. If you are doing 
> clustering, there are quite a few distros slated for this exact purpose.

     I am ignorant.  With my skills, it's often easier to write my own 
software than do the research.  Not always...

> > lost!  The only thing I'd really like to do that would be worth while, 
> > is fixing the absurd variable processing in BASH.  
> 
> You don't need to use BASH. There is no lack of languages out there. 
> Especially ones that can do reflection with variable names.  PHP, Ruby, 
> Python, etc...

     None of them offer the kind of facilities that BASH offers.  I can 
whip up a custom app in Bash in a matter of hours to day that youd take 
weeks in Perl/Icon/Python, etc. and months in C.

     BASH delivers the most computation (using standard Linux utilities) 
per line of code and per hour of human programming; far more than any 
other language when used by a skilled programmer.  (That's not to say 
it's efficient in terms of compute time... But in most cases, the goal 
is to get a prototype up, let the user play with it, and realize what he 
REALLY wanted...  which is not usually what you thought he described.)

     Double substitutions, recursive substitutions, done right, give you 
sparse matrix abilities, abilities to define rules, other things. In 
Bash, you can't even do a variable controlled substring, because the 
substitution is not recursive. Once you've done ${a}, you can't do 
${a:$n:3}; that blows 90% of the utility of substringing!

     What I want, is to be able to say:

a=de
b=fg
defg=testing

and get echo {{a}{b}} to say "testing".

     That way, you can do all kinds of things to select the values of a 
and b, and use that to pick up rules, what have you.  Messy, but 
powerful and FAST to program!  

     You could do that with Word Star Mail Merge.  I used Mail Merge to 
generate 180,000 statements of code on one project, did that in four 
months, did it over and over and over again as the other group tinkered 
with the specifications they needed to get their part of the project 
running.  My ability to deliver 120,000 - 180,000 lines of working C 
code to them in 24 hours made their part of the project immensely 
easier, resulting in a success rather than the expected failure.

     It took me one month to understand the project, one day to deliver 
the prototype, a few more for the first production version, and then it 
was just tweaking the rules and code over and over again as the other 
team tried to figure how to get their half of the system working.  In 
the end, we had a working rule base, so I translated the output section 
to fold that into bit strings, and delivered their functionality in 
under 4,000 lines of code.  But I could ONLY get there by generating the 
180,000 lines of C code till they simplified their part.  And they could 
only do that, if they got a new batch of interface code for every change 
of their protocols.  It was fun.  It's always fun watching people's eyes 
light up. Bonus points if the jaw drops; but only if the corners of the 
mouth smile.

     That's the kind of crazy stuff I do. I've programmed in over 35 
different computer languages, been a Software Linguist on an project for 
IBM, etc. Don't claim to be good at any language... but I get things 
done.  And do it faster and simpler than most anyone else.

     If you can do that kind of crazy recursive substitution, you can do 
fantastic rule based AI. I've done things like that. That is power that 
blows all compiled and many interpreted languages away when used by 
someone like myself.  Even blows prolog away.  And as I said, it can be 
used to generate other code.

     The first cut of anything I'll do in BASH, if I can.  Next, we try 
making portions more efficient.  And finally, I'll use something like C, 
or even assembler, for the parts that really need speed.

     What else I am looking for, is a good, simple editor that has macro 
capabilities, and the ability for the macros to kick off into BASH to 
run utilities.  I had a version of CRISP way back that could do it under 
Interactive Unix (pre Slowaris) and DOS...  but I lost the source.  (No, 
I don't want to learn Emacs.  My fingers do word star.)

     Anyway, it sounds to me that if I could set up an ethernet boot, I 
might be able to get something installed.  (Gentoo would be 
interesting... but again, I don't trust this hardware for compiles.)

 -JVV-