Re: asrock, problem with nic after windows-boot - Exact Opposite issue the OP is having

On 17 Jun 2006, in the Usenet newsgroup comp.os.linux.networking, in article
<1150569738.676166.206750@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>, iforone wrote:

My small history with PCs started ~1990-ish, when I sold them, when I
worked briefly in sales (a few months) for a major appliance/tv/stereo
chain, ...yet I knew hardly anything about them (their inner workings).

And this is different from the norm exactly how? ;-)

You know? I never use/used a power strip in that manner (for *ATX*
systems), and I never condoned it to others (and there are quite a few
I've seen that use this method, with*out* shutting down properly
first), for the simple fact that; in windblows, the 'dirty' bit would
get set in that way, causing Scan/ChkDsk to have to always run on
startup (though that could be circumvented, it isn't such an
intelligent thing to do).

That's true of many O/S that use a write cache. Unix has had this for
decades, as did Novell Netware and MacOS predating windoze. SGI had a
neat switch on their Challenger boxes in the early 1990s - you punched
the "power switch" on the front panel, and the O/S would cleanly terminate
any running apps, then flush the buffers, and finally unmount the disks
before turning off the power. There was also a hidden power switch at the
top right of the back panel, but this one killed power without doing
anything safe.

(this is Scott Mueller's site (an Electrical Engineer, who publishes
very useful info IMHO) - I own the 13th Edition of his major
publication, "Upgrading And Repairing PCs")

0-88022-395-2 88-62746 First edition
1-56529-467-X 93-86245 Third Edition
0-7897-1636-4 98-84382 Tenth Edition
0-7897-2974-1 (2004) 15th Edition

Some of his stuff I take with a whole bag of salt, but I admit to using his
"swap the 79 cent crystal" improvement to an original PC-AT. It was working
at 8.5 MHz (rather than the stock 6 MHz) which pissed off my cube-mate who
had a real 8 MHz version with the anti-overclocking BIOS.

I hadn't realized that the earlier IBM design required the
'write-to-every-address' requirement in order to set parity, but allow
me to ask;
Is this a Registered(Buffered) thing, or a ECC(Unbuffered) related
thing - or perhaps both?

It is for the ECC for the same reason.

Yes - I _must_ do this (in earnest) - especially also to delve into
BASH scripting

For shell scripting, start with the Bash-Prog-Intro-HOWTO

-rw-rw-r-- 1 gferg ldp 31540 Jul 27 2000 Bash-Prog-Intro-HOWTO

then grab a copy of The Grendel's fabulous book "Advanced Bash Scripting
Guide" from the LDP (http://tldp.org/guides.html). I've lost track of how
many revisions it's gone through, but it's a _very_ polished book with lots
of examples. I also recommend setting down with a bucket of coffee and the
man pages, and trying to read through the boot scripts on your system. These
are generally written by someone who is _really_good_ at scripting, and is
_flaunting_ those skills.

But from a purely hardware level (no OS involved), using the BIOS ROM
code interaction with the PC components (x86), there is definitely
differences between a 'warm' and a 'cold' boot - and while I can't
pinpoint the specifics, and may be confusing the issue (on topic), I
certainly suspect this diff (perhaps in combo with the WOL/WOR BIOS
(DMI?) feature I mentioned earlier) is what's causing this ZeroConf
networking issues, whether one uses Linux -or- Microshaft, or any other
OS for that matter.

Well, I know there is that difference in the missing hardware reset action,
but I think your 'second reboot and it works' might be that the first is
only partially fixing the problem, and the second reboot sees something it
recognizes and is able to correct/handle.

I know 'grep', but would love to be able to use 'awk', 'grep' in small
scripts and for 'plucking-out' only the relevant info - I hear ya
though...in due time.

It's a fairly powerful tool - but you do realize that O'Reilly was able to
find enough materials to get two full size books and a reference guide for
it (my wife cringes when I get _near_ a bookstore).

Effective awk Programming, 3rd Edition May 2001 $39.95
sed & awk, 2nd Edition Mar 1997 429 pgs $34.95
sed & awk Pocket Reference, 2nd Edition Jun 2002 $12.95

I see, 'Zgrep" is for Gzip files I presume (perhaps also for most
compressed types)

except bzip2, but yeah it also handles regular files as well, which comes
in handy when searching through a directory that includes both types.

but those DIRs you list above are not occupied with anything ATM

Fumble finger - try /usr/share/man/man1/

ahhh.....I have a 32MB Pentium 166 in a BabyAT here - it has a
minimalist and tweaked win98 on it (runs great),

What is my P-133... let's see, it's running a log server, print servers,
and backup network access box. It was running X - it still has FVWM2 on it
which is a fairly lightweight desktop. I'm not into eye-candy, and mostly
use X to provide multiple xterms. X really doesn't belong on a server,
because X is a _user_ interface, and _users_ shouldn't be mucking about
the servers. It's astonishing how much memory and disk space X (and the
"popular" desktops) take.

yet I need to get it a NIC...there is an open PCI slot (as well as ISA
slots), but I think I need an earlier PCI type that isn't so easy to find
(New) nowadays...I'm sure I can find one, if only I can recall the
requirements

Possibly supply voltage. Look in Mueller's book.

I eventually plan on using it as either something headless (some sort
of Server) on the LAN

287028 Mar 31 2003 Remote-Serial-Console-HOWTO

A wise person has two ways into a box - over the net, and through a console.
As this is to be headless, string a null modem cable to another system so it
can be used if networking goes weird. I've got six headless boxes here, and
use cascaded A/B/C/D switches to one system to allow direct access. Big clue
is to put Dymo (embossed) labels on the switches, and computers, and set the
prompts so that they include hostnames - it's embarrassing to halt a computer
then yank the power cord out of the wrong one.

Old guy
.

Relevant Pages

  • Re: Ubuntu Server 10.04 LTS: initiate shutdown by pressing POWER button ?
    ... In the morning I just switch the Power ... Before the Ubuntu Server 64 install, ...
    (Ubuntu)
  • Re: server connectivity??
    ... I do not have a different switch to try. ... I was thinking switch as well. ... Everything seems to be not using power save. ... Check everything in the server to make sure there is no power saving going ...
    (microsoft.public.windows.server.sbs)
  • Re: Poweroff after X
    ... similar procedures can't switch the power off ... If I do not start X server and type poweroff - the computer will ...
    (comp.os.linux.x)
  • Re: Loopback as primary DNS
    ... figure the switch would boot up and be ready way before the server, assuming the switch and server are either on the same UPS or different UPS that all power back up after a power outage. ...
    (microsoft.public.windows.server.dns)
  • Re: Windows/Macro Language Info?
    ... Java scriptlets and servlets can also carry malware. ... other part of the HTTP server. ... PHP can deliver malware and coordinate attacks to precisely the ... server-side scripting languages are not. ...
    (comp.lang.cobol)