Re: [Sheflug] Good linux learning resource

["Chris J" <cej [at] xxxxxxxxxxxxxxxx>]

And Lo! The Great Prophet " Chris Johnson" uttered these words of wisdom:
> All these questions on the list make me realise how little I know.  Is
> there
> a good online resource or a book where I can start learning more about
> Linux
> administration?  The Suse manuals are OK but not as much in depth as I'd
> like.

First thing to remember - each Linux distribution is different, so no book
is going to be perfect. Second thing to remember - Linux is essentially
UNIX, so any decent UNIX admin book is good for the job.


> Any suggestions appreciated.
>

As well as all the HOWTO's et al, books that I class as essential for me:

- Newnes UNIX pocket book ... an invaluble reference when you're starting
off in UNIX. Newnes pocket books though seem to be hard to find these
days, but all of them are excellent if you can find them. I don't use it
much these days but when I was starting off with UNIX, it went everywhere
with me :)

- Essential System Administration (aka the "Armadillo book") ... from the
O'Rielly guys. And then there's all the other O'Rielly books - DNS,
Sendmail et al - which cover specifics to a deeper level. There's very few
people I know that don't have a copy somewhere :)

Then there's the need to learn how to use standard tools such as awk, sed,
vi etc - this way no system can catch you out all the time (just
occasionally as you come across a bizarre or old implementation of awk, or
it's a system that doesn't have GNU tools, so there aren't any of those
nice extensions you've always relied on). O'Reilly's Sed & Awk book should
do the job adequately here (I've never used it myself, but O'Rielly books
are good so shouldn't cause any problems).

>
> I suppose I need a project really so that I have a target to aim for and a
> list of tasks that I really need to know well.
>

Writing scripts to process log files is always a good one. When in Uni, I
wrote a script to generate statistics for my webpage (and that's not
either trivial to parse when the script was written purely with grep, sed
and cut). In the event, a few people ended up using it though. The
downside was that it was not light on processor load. At one point it was
running on the main university web server and had an the webmaster on my
back telling me not to run it every night as it was putting a fair strain
on maple :)

These days I'd right such a beast in awk, perl or C.

But those are the sort of things to look at if you want a project; scripts
that generate statistics can be useful in both teaching and having
something that's also going to be useful at the end of the day. Similarly
monitoring scripts that run once a minute/hour/day that do things like
send you an email if a disk is filling up, or CPU load seems to be
extraordinarly high for an amount of time (read up on CPU load averages),
scripts that roll log files once they're a week old or they reach a
certain size (or both!) etc.

Essentially it's a case of getting under the hood and wondering what
happens if you remove /that/ bolt, and actually doing so, then learning
how to put it back together again as you find out it was securing the
engine :-) I'm still learning new stuff - mostly tweaks to things I
already do to get more performance out of less lines, but occasionally
something life changing does appear that I end up adopting :)

A list of tasks could be endless ... you've got everything from adding a
user and adding a new disk to wondering what is causing bob's mail to
bounce, when all the config looks okay, or why the DNS server is reporting
duff data.

Chris...

-- 
\ Chris Johnson                 \
 \ cej [at] nightwolf.org.uk          \
  \ http://cej.nightwolf.org.uk/  ~-----------------------------------+
   \ Redclaw chat - http://redclaw.org.uk - telnet redclaw.org.uk 2000 \____

___________________________________________________________________

Sheffield Linux User's Group -
http://www.sheflug.co.uk/mailfaq.html

  GNU the choice of a complete generation.