Linux-Misc Digest #707, Volume #21                Mon, 6 Sep 99 19:13:07 EDT

Contents:
  Re: Linux+NT dsaster (help wanted) ([EMAIL PROTECTED])
  ZIP parallel port mounting kernel 2.2 ([EMAIL PROTECTED])
  Re: General Rant from a Linux Newbie (newsseeker)
  Re: Kernel-Panic ???? (Leonard Evens)
  Re: General Rant from a Linux Newbie (David Steinberg)
  Re: Help: RH6 Installation Problem (Leonard Evens)
  procps funny under smp (Avery Earle)

----------------------------------------------------------------------------

From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.setup
Subject: Re: Linux+NT dsaster (help wanted)
Date: Mon, 06 Sep 1999 21:52:36 GMT

Hi Emilio,

I did the same thing a while back to my NT PC. . Restored the C: from
a backup image, no go. I reinstalled NT, still did not work.
Fortunately I had saved a note about fdisk /mbr.  Worked great.  NT
was back!

I then reinstalled Linux and specified that I wanted to boot from a
floppy disk.  NT was safe.

Hope this helps,

Ken


"D. Emilio Grimaldo Tunon" <[EMAIL PROTECTED]> wrote:

>
>A colleague has an NT Dell Latitude CPt and according to NT's 
>disk manager this is how the partitions look like:
>
>    +-------------+---------------+---------------+------------+
>    |   FAT       |  NTFS   C:    |  NTFS   D:    | Free       |
>    +-------------+---------------+---------------+------------+
>
>Now since the install was done when you try to boot the laptop
>you get a black screen with "Invalid partition table" and nothing
>else. The only way to get the machine to boot is with the NT
>boot disk which then boots NT. We copied the boot.ini etc
>from the boot disk into C: but still the same results (it says
>to boot from partition 2).
>
>  I think that during installation the MBR got overwritten by
>LILO and somehow it cannot find itself. We have been thinking
>about doing "fdisk /mbr" from NT but are worried that perhaps
>that will screw up the whole thing. Will that restore the 
>ability to boot NT from HD? or what? We can install Linux again
>but NT install/reconfigure would be a nightmare, so is there
>a way to first make this thing boot NT from HD? any plans on
>how to proceed? any help would be greatly appreciated.
>
>                       Emilio
>
>-- 
>D. Emilio Grimaldo Tunon       Compuware Europe B.V. (Uniface Lab)
>Software Engineer             Amsterdam, The Netherlands
>[EMAIL PROTECTED]  Tel. +31 (0)20 3126 516
>*** The opinions expressed hereby are mine and not my employer's ***


------------------------------

From: [EMAIL PROTECTED]
Subject: ZIP parallel port mounting kernel 2.2
Date: Mon, 06 Sep 1999 21:50:52 GMT

Hi,

        I have noticed that a few people are having
problems mounting their parallel port iomega zip drive
when moving from kernel 2.0.x to 2.2.x. I initially had
similar problems with either modules or builtin kernel
support for the zip drive, so here is what worked for me:

kernel 2.2.x:

General Setup:
        Parallel Port Support  y
        PC-style hardware      y

Character devices:
        Parallel printer support M

SCSI  support:
        SCSI support       M
        SCSI disk support  M
        IOMEGA parallel port (ppa - older drives) M


I suppose you could also choose to build in the scsi support,
but I wanted to use it this way. To get things
going at boot time , and as I use debian,
I added the following lines to my /etc/modules file:

        # Zip parallel type
        scsi_mod
        sd_mod
        ppa


Then it is ready to mount:

         mount /dev/sda4 /mnt



or if you use autofs auto.amnt:

zip  -fstype=auto,nosuid,nodev :/dev/sda4



Cheers




Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

------------------------------

From: newsseeker <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy
Subject: Re: General Rant from a Linux Newbie
Date: Mon, 06 Sep 1999 17:33:03 -0500

Ross Smith wrote:

> <snip> [EMAIL PROTECTED] is obviously only familiar with Win9x. You're right, NT's
> Task Manager is far more powerful than Win9x's, and hardly ever gives
> any problems with killing anything.

I work on a NT Workstation every day.  I still don't find WinNT's Task Manager any
more powerful other than to show me CPU Usage and processes vrs just the Shutdown
Dialog Box in Win98.  I'll give you a perfect example I've seen on my NT machine.  I
communicate with a Single Board Computer that we are developing that uses no
Handshaking.  Someone on my machine once turned on Hyperterm to talk to the board and
inadvertently turned on Hardware Handshake.....Predictably Hyperterm froze.
Unpredictably my machine just about froze, and bringing up the task manager showed
100% CPU usage.  No matter how hard we tried, we couldn't get Task Manager to shut
down Hyperterm...period (I quit counting the number of time Task Manager kept saying
"Wait to see if the process responds").

Another example....MKS source-version over the network.  Should the Machine holding
the license file go down, MKS will hang forever, and the Task Manager will just stay
on the desktop telling me to wait on MKS to see if it responds....over and over
again.  Task Manager, However, will not shut down MKS regardless of the number of
times we hit the 'End Task' button.

If these apps were in Linux, I would be able to say killall -9 MKS and killall -9
Hyperterm and they would absolutely be gone.

So, are you going to be saying I'm obviously not familiar with Windows NT Patched
Task Manger version X just so you can discredit again?

And, no, I don't care about how to fix this NT problem.  I'm just very thankful that
I don't have to put up with this on my own machines.

--nwskr




------------------------------

From: Leonard Evens <[EMAIL PROTECTED]>
Subject: Re: Kernel-Panic ????
Date: Mon, 06 Sep 1999 16:20:56 -0500

Sebastian Koball wrote:
> 
> please help !
> what does it mean : Kernel panic ?
> thank you
> sebastian.koball(at)stud.uni-rostock.de

It means the kernel has met a situation it can't deal with.
There are usually other error messages following which give
some information about what might have gone wrong.  Also,
is is important to know when this happened.  One common
situation is when booting, one gets a kernel panic message
followed by an error message about VFS being unable to mount
root on some number (of a device).   This is usually the
result of an error in lilo configuration or partitioning.   
But while this
is a common istuation, there are many other situations which
produce the kernle panic messages.
-- 

Leonard Evens      [EMAIL PROTECTED]      847-491-5537
Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208

------------------------------

From: [EMAIL PROTECTED] (David Steinberg)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: General Rant from a Linux Newbie
Date: 6 Sep 1999 22:49:41 GMT

K. Bjarnason ([EMAIL PROTECTED]) wrote:
: > You know, if you had a clue about any of this, you'd realize that it is 
: > usually source that is packaged in tarballs.  

: Golly gee, like I didn't know that.  I also happen to know that out of 
: the times I've downloaded apps for Linux that weren't part of the 
: distribution package they usually came as... source.

Funny, I've found most applications that I've downloaded as RPM's.  
A typical app that I haven't been able to find in RPM format is a 
dockable application for Win Maker that basically produces one 
executible.  Hardly a tragedy.  Also, applications that are really early 
in their lives often aren't in RPM's, to discourage the end-user from 
using them before they're ready.

Typically binaries distributions of large, important apps can be found.

Was your search just a mission to find things to whine about, or would you 
like to be useful?  If the latter, why not make an RPM?  A clever 
developer such as yourself can probably handle "man rpm," where it 
explains how to build an RPM.  Then, you can say you've helped the world 
and made life easier for the end users that you care so much about.

: Indeed.  Now, of the package systems available, which is the standard 
: one?  Or even three?

Standard?  Well, the standard package management for a distribution 
generally comes with it.  If you got RedHat, Mandrake, or SuSE, you'll 
probably use RPM.  Debian or Corel (soon), you'll use DEB.

: WTF?  What's ivf?  Something wrong with shipping the app ("package" if 

Well, we already talked about i.  v is verbose. h is hash -- it draws a 
little progress meter with #'s.  As I said before, if it's over your head, 
just go with with i...for install, remember?

: you prefer) so that you can run _it_?  You know, like I just downloaded 
: "LinuxApp 1.09", so I have a LinuxApp1.09 file which is, itself, 
: executable and simply clicking on it will run the installer?

What is the brilliant logic behind having each package an executable with 
the package management program built in?  Should every letter you write 
have a copy of Word embedded in it, too?  Why can't a package be input to a 
single program that deals with it?  Wouldn't it be easier if installing 
or uninstalling a package was done the same way each time, using the same 
program?

Oh, right.  That's not how Windows does it.  So it must be worse.

: Yes, yes, *you* like command-lines and switches.  So do *I*.  About 98% 
: of the desktop users out there do NOT; they want it point-and-click 
: easy; compare your method above to "click on setup.exe or the .EXE file 
: you just downloaded."

As I said, there are GUI front-ends.  Why did you ignore that part of my 
post and instead ramble on about how command-lines and switches are too 
hard for most desktop users?  I specifically said GUI's exist.  I just 
don't have much more information than their names, since I don't use 
them myself.

If you want to inform yourself before you continue rambling, you might 
go take a look at GnoRPM or KPackage.

: Seen MSI?  Guess not. Let me give you an overview.

What's MSI?  Is that the existing crap that leads to OS re-installation 
every six months on Window 9X/NT?  Or is this a brilliant new 
"innovation" in the phantom Windows 2000?

[snippage]

: Want to see what's there?  Simple; there's a standard location for 
: accessing application uninstallation and configuration settings: control 
: panel/add remove programs.  Let's take a peek, shall we?...

: So, anyhow, let's look.  Repair Offce.  Add/Remove Features.  Remove 
: office.  Repair is there in case someone else's install went blooey and 
: trashed something - it knows what was previously installed, how it was 
: installed, whether install-on-demand has fired for any components, and 
: refreshes the actual installation on disk the way one would expect.

Control Panel->Add/Remove Programs?  That sounds familiar...

Oh yes, that's where I went last week when Word 97 started crashing every 
five minutes for no reason.  Add and remove didn't do me much good.  I 
removed it, and then I re-installed it.  It didn't help.  I then tried 
removing all of Office, and re-installed it.  That didn't help either.  
Why do I try next?  What magical file is broken?  Hmmm...I've got nothing 
much to lose.  The only useful thing I do with Windows is use Office, 
anyways.  Try freshening Windows and re-applying service pack.  Oops, 
I've broken IE.  Dang, I forgot, that's part of the system.  Re-install 
IE.  Word still crashes for no reason every five minutes.  But, I've lost 
an evening that I should have been using to write my report.

I still have to write that report, so what's left to do?  Backup data, 
nuke NT and all installed apps.  Re-install.  Well, installing is sure 
easy.  Just run the EXE, and...oh wait.  Except for my Network Card 
drivers.  I had to move files from a subdirectory called WIN95NT into its 
parent for the installer to find them.  And my Sound Blaster drivers.  
This one is brilliant: the installer wanted files to be Synthgm._sbk, 
Synthgs._sbk, Synthmt._sbk, etc; however, they were actually called 
Synthgm.sbk, Synthgs.sbk, Synthmt.sbk, etc.  So, I had to hope that those 
were the files it wanted, and RENAME them.  Brilliant.  If only installing 
software was this easy in Linux...

Geez.  Anyways, the whole process ended up solving my problem.  Word now 
runs without the mysterious crashing.  But it took me three evenings to 
re-install and re-configure everything.  So, I had to do my whole report 
this weekend (the last of the summer).  I could have thought of better 
ways to spend it.

How could this have been solved without re-installing the whole system?  
As I see it, two things are missing.  First, a decent error message 
would have told me what program/library contained the code that was being 
executed at the time of the crash.  Then, a decent package management 
system would have told me where that file came from, so that I could 
replace it.  As it was, I had no idea what was broken, so I had to fix 
everything.

BTW, this is not an isolated incident.  Something similar happened to my 
brother a few weeks ago in Windows 98.  Internet Explorer started crashing 
for no reason.  It gave a decent error message, informing us that the 
problem was in foobar.dll.  But where did foobar.dll come from?  Tried 
re-installing IE, but that didn't help.  The ending of this story is the 
same: re-install the whole system because you don't know where one file 
comes from.

: Keep in mind, all these options are menu selectable and changeable at 
: any time.

Helpful.  Where's the feature that tells me what package a file came 
from?  That's the one I actually need.

: Also keep in mind that while MSI is not currently widely used - most 
: major apps are getting prepped to roll out MSI-based installs sometime 
: shortly after Win2K finally ships - the differences between this and 
: exising methods aren't all that significant, except that existing 
: installs typically only have an "uninstall" option from the control 
: panel.

Oh, so it's a "new inovation."  And it's still the same old shit, right?  
You still can't find out what package a file came from, can you?

<snippage>
: The problem with your view of things isn't that command lines are 
: inherently bad, it's that you don't seem to comprehend that *not* every 
: user is also a developer.  Many users want to *use* the machine, not 
: have to understand it - that's *our* job, as developers.  We're supposed 
: to make the machine as nearly invisible to the user as we possibly can.

Ramble rambe.  See "GUI frontends exist" above.

: The only time I've ever had that problem was when the dll file was, in 
: fact, corrupt - usually because I had a drive crash.  Mind you, that 
: hasn't happened in, oh, 4 years.  So how does Linux handle that?  Does 
: Linux contain code that can magically rebuild dynamicly-loaded modules, 
: presumably by analyzing what the calling application thinks they should 
: do, then re-writing them on the fly?

Were they magically created in the first place?  Or were they just copied 
from somewhere?  Maybe if Windows package management could tell me from 
WHERE they were copied, I could just replace the broken files.  Not the 
whole system.

BTW, I think you should take out a patent on that magical rebuilding 
idea.  I would expect that, rather than fixing the problem, Microsoft is 
likely to add MagicRebuildWizard to it's next verion of the OS, Windows 
2013, and I think you should at least make some money of it, since it was 
your idea.

: Oh, I see.  Linux *doesn't* do this any better than Windows, so you're 
: simply spewing for the sake of making noise.  :)

If you'd used RPM, you would have seen the obvious solution to the 
problem that I mentioned above, you wouldn't have made this comment, and 
now you wouldn't look like a twit.

Expand your mind.  Try Linux.

------------------------------

From: Leonard Evens <[EMAIL PROTECTED]>
Crossposted-To: jaring.os.linux
Subject: Re: Help: RH6 Installation Problem
Date: Mon, 06 Sep 1999 17:00:27 -0500

Tan wrote:
> 
> I don't know where is the packages to be stored, I thought all of
> these are going to be stored under /?

Packages are stored in a variety of directories. Any given
package may have components in several directories.  But the
bulk of everything goes in /usr.   If you just make a / partition
and no /usr or /var partition, you won't have to worry about
any of this.   Recall that partitions are not the same as
directories.  The basic file system has a tree like structure.
Some nodes on that tree may have filesystems mounted on them,
but that is only one way of organizing space.  Thus you may
have a separate partition with is own filesystem mounted on
/usr or you may just have subdirectory called usr of the /
filesystem.

> So, how should I separate my HDD space for these partitions?

I've already made some recommendations in a separate communication
so I won't repeat myself.

> Is there any other ways that I can install LILO beside install it into
> MBR or diskette? I don't like to mess with my MBR.

Except for NT+Linux dual boot systems, I've never seen a case
where putting the lilo boot loader in the MBR of the first drive
caused a problem.  But you can put the lilo boot loader in
any active primary Linux partition on the first drive.  Since
this is more complicated and you don't yet appear to have
mastered the installation process, I wouldn't recommend
trying it.

> Which package I should select for my installation?

Go with the default packages.  Later you can add or remove
packages as you learn more about what they can do.  But
it wouldn't hurt to select the option which lets you go through
and select individual packages.   As you go through them, the
F1 key gives you a brief description of what each does.  It
is probably best to err on the safe side and just accept everything
that is there by default.  But you may find a few other packages
you think would be helpful.   If there are unresolved dependencies,
you will be notified at the end and allowed to add other needed
packages.

The long list of packages can be intimidating, but in practice
it is not so difficult to work with them.   It is rather
trivial to add additional packages, so nothing is written
in stone by your choices at installation.


-- 

Leonard Evens      [EMAIL PROTECTED]      847-491-5537
Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208

------------------------------

From: [EMAIL PROTECTED] (Avery Earle)
Subject: procps funny under smp
Date: 6 Sep 1999 22:58:48 GMT

I find that the START column of the output of "ps ux" is
way off under smp. Huh? 

======= example =========
{10}$ who
ae       pts/2    Sep  6 20:04
{11}$ uptime
  8:15pm  up 1 day,  1:35,  2 users,  load average: 0.32, 0.33, 0.27
{12}$ date
Mon Sep  6 20:15:27 NDT 1999
{13}$ ps ux
USER       PID %CPU %MEM   VSZ  RSS TTY      STAT START   TIME COMMAND
ae       30408  0.0  0.2  2492  868 pts/2    R    07:27   0:00 ps ux

======= hardware =======
dual pii-233, asus
RTC set approx. to gmt, date set via rdate.

======= software =======
rhl 6.0, 6.0 patches to date, except kernel.
Current kernel from rawhide; trouble dates back to forst use of 2.2

procps-2.0.2-2
procps-X11-2.0.2-2
kernel-smp-2.2.12-4

-- 
Compromise would be complicity.    -- Maurice Duplessis

------------------------------


** FOR YOUR REFERENCE **

The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:

    Internet: [EMAIL PROTECTED]

You can send mail to the entire list (and comp.os.linux.misc) via:

    Internet: [EMAIL PROTECTED]

Linux may be obtained via one of these FTP sites:
    ftp.funet.fi                                pub/Linux
    tsx-11.mit.edu                              pub/linux
    sunsite.unc.edu                             pub/Linux

End of Linux-Misc Digest
******************************

Reply via email to