Linux-Misc Digest #753, Volume #27 Mon, 30 Apr 01 15:13:03 EDT
Contents:
Re: Could Linux be used in this factory environment ? ("Erik Funkenbusch")
Re: What recent distributions for an old 486? (John)
Re: Disc Partitioning & / ( root ) (Chris Leahy)
startx -- error ("Jeffrey J. Bacon")
Extended partition types (05, 0f & 85?) (Richard Lewis)
Re: Extended partition types (05, 0f & 85?) (John)
KDE editors problem (zak)
Re: killing a process ("Dr. Aldo Medina")
Re: Disc Partitioning & / ( root ) ("Peter T. Breuer")
Re: startx -- error (Lew Pitcher)
Re: Extended partition types (05, 0f & 85?) (Richard Lewis)
Re: Extended partition types (05, 0f & 85?) (John)
Re: Disc Partitioning & / ( root ) (John)
Re: Best way to manage source code? (Michael Heiming)
Re: Disc Partitioning & / ( root ) (Sony Anthony)
Re: killing a process (Jean-David Beyer)
----------------------------------------------------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.hardware
Subject: Re: Could Linux be used in this factory environment ?
Date: Mon, 30 Apr 2001 12:01:56 -0500
"hac" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Erik Funkenbusch wrote:
> >
> > "hac" <[EMAIL PROTECTED]> wrote in message
> > news:[EMAIL PROTECTED]...
> > > Erik Funkenbusch wrote:
> > > >
> > > > "Jonadab the Unsightly One" <[EMAIL PROTECTED]> wrote in message
> > > > news:[EMAIL PROTECTED]...
> > > > > "Erik Funkenbusch" <[EMAIL PROTECTED]> wrote:
> > > > >
> > > > > > > That's silly. All you need is to queue the upgrade to
> > > > > > > any given page until nobody's looking at it.
> > > > > >
> > > > > > And you magically know when nobody is looking at it, how?
> > > > >
> > > > > The OS should know that.
> > > >
> > > > It should? I didn't realize they had optical sensors that allowed
the
> > > > terminals to notice when someone was looking at the terminal or not.
> > > >
> > > Even an idiot should realize that no one is looking at a page that
> > > isn't displayed. Ergo, pages not displayed may be updated. Which is
> > > what he clearly intended to convey, before you went out of your way to
> > > be obtuse and argumentative. Unless, of course, you really are that
> > > stupid. You seem bent on convincing everyone that you are.
> >
> > How exactly do you "update the page" of a standalone application?
> >
> > His argument was about how a single app on a server accessed by
terminals is
> > easier to maintain than an HTML based system.
> >
> If he was talking about a single app that encapsulates the data, then
> you have a point. I was thinking of a single application that is
> delivers the data, where the data is stored as separate files,
> records, whatever. You can certainly update records in a database
> without restarting it. The key point is that the data is stored one
> and only place. Browsers store a local copy, which complicates
> updates.
The point here is what happens when you need to update the APP, not the data
(though frequently they coincide).
> The drawback to HTTP for this application is that when a page is
> requested, a connection is opened, the page downloaded, and the
> connection is then dropped. It's designed to be stateless, which is
> good for what it was designed for. But in this application, you would
> like to know what each screen is displaying. With HTTP, you have to
> add another layer, with the server keeping track of what page was last
> downloaded to which screen. It's easier if you use a protocol that
> keeps the connection open, automatically locking the
> page/screen/file/record. And an open connection simplifies displaying
> dynamic data.
Nonsense, you need only provide state information in the URL, or via a
cookie.
------------------------------
Date: Mon, 30 Apr 2001 18:02:45 +0100
From: John <[EMAIL PROTECTED]>
Subject: Re: What recent distributions for an old 486?
Slackware should work fine.
------------------------------
From: Chris Leahy <[EMAIL PROTECTED]>
Subject: Re: Disc Partitioning & / ( root )
Date: Mon, 30 Apr 2001 13:08:11 -0400
==============3A8CCBE53A8217DA0B531952
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sony Anthony wrote:
> Hi:
> following are couple of my understanding about disc partitioning ( Pl
> correct me if I m wrong ) ...
>
> 1. All bootable partitions should be in the first hard disk
Not true
>
> 2. All bootable partitions should be a primary partition
True
>
> 3. All bootable partitions should start below 1024 MB
Depends on your system
>
>
> Then it turns out that the root ( / ) partition for linux should be
> below 1024 MB. Since this is true for the C: drive for windows also,
> I always end up with the following partition strategy all the time..
> 1. / partition sitting at 0 - 500 MB
> 2. C: drive starts at 500MB and spans 5 GB or something
> 3. The remaining hard disk partitioned as a huge extended partition
> ( that contains both windows logical drives and logical partitions for
> linux ).
> 4. LILO sitting in /dev/hda MBR
This just isnt true. I'm currently booting / from well above 1024
hde1=FreeBSD 4.3 size = 5000 MB
hde2=Redhat 7.1
and my hardware is a good 2 years old.
As you can see, I'm booting from the 5th
(not really but the OS looks at it as first drive on third IDE controller.
Hence hde)
drive and above 1024 cyl as well
This is a Promise ATA100 card that hde is connected to.
I have an ATAPI CD-ROM drive connected to the primary on-board IDE
controller
as master. There is no slave. The secondary on-board controller is
disabled.
>
>
> I never thought about this till recently, I know I m wrong somewhere
> since this will prevent me from adding one more OS, since it s root
> drive has to be in the first 1024 MB.
>
> Could somebody clarify the confusion. Also does anybody have any better
> partitioning strategy.
>
> Thanks
>
> sony
--
==========================================================
Christopher Leahy | [EMAIL PROTECTED]
UNIX Systems Administrator | http://www.zoltanium.com
Zoltanium (aka /dev/null) | Voice (610)408-0151
==========================================================
==============3A8CCBE53A8217DA0B531952
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Sony Anthony wrote:
<blockquote TYPE=CITE>Hi:
<br> following are couple of my understanding about disc partitioning
( Pl
<br>correct me if I m wrong ) ...
<p>1. All bootable partitions should be in the first hard disk</blockquote>
Not true
<blockquote TYPE=CITE>
<br>2. All bootable partitions should be a primary partition</blockquote>
True
<blockquote TYPE=CITE>
<br>3. All bootable partitions should start below 1024 MB</blockquote>
Depends on your system
<blockquote TYPE=CITE>
<p>Then it turns out that the root ( / ) partition for linux should be
<br>below 1024 MB. Since this is true for the C: drive for windows also,
<br>I always end up with the following partition strategy all the time..
<br>1. / partition sitting at 0 - 500 MB
<br>2. C: drive starts at 500MB and spans 5 GB or something
<br>3. The remaining hard disk partitioned as a huge extended partition
<br>( that contains both windows logical drives and logical partitions
for
<br>linux ).
<br>4. LILO sitting in /dev/hda MBR</blockquote>
This just isnt true. I'm currently booting / from well above 1024
<p>hde1=FreeBSD 4.3 size = 5000 MB
<br>hde2=Redhat 7.1
<br>and my hardware is a good 2 years old.
<p>As you can see, I'm booting from the 5th
<br>(not really but the OS looks at it as first drive on third IDE controller.
Hence hde)
<br>drive and above 1024 cyl as well
<br>This is a Promise ATA100 card that hde is connected to.
<br>I have an ATAPI CD-ROM drive connected to the primary on-board IDE
controller
<br>as master. There is no slave. The secondary on-board controller is
disabled.
<blockquote TYPE=CITE>
<p>I never thought about this till recently, I know I m wrong somewhere
<br>since this will prevent me from adding one more OS, since it s root
<br>drive has to be in the first 1024 MB.
<p>Could somebody clarify the confusion. Also does anybody have any better
<br>partitioning strategy.
<p>Thanks
<p>sony</blockquote>
<pre>--
==========================================================
Christopher
Leahy |
[EMAIL PROTECTED]
UNIX Systems Administrator | <A
HREF="http://www.zoltanium.com";>http://www.zoltanium.com</A>
Zoltanium (aka /dev/null) | Voice (610)408-0151
==========================================================</pre>
</html>
==============3A8CCBE53A8217DA0B531952==
------------------------------
From: "Jeffrey J. Bacon" <[EMAIL PROTECTED]>
Subject: startx -- error
Date: Mon, 30 Apr 2001 17:36:11 GMT
Authentication failed - cannot start X server
Perhaps you do not have console ownership?
the above error appear when calling startx from a user account (root
works fine).
I'm on RedHat7, I have just wiped out KDE from my system and put it back
on (ONLY: libs, libs-sound, support, base) as root. I started it as
root and copied root's .Xclients, .Xclients-default to the user's home
directories but the above error occurs when trying to start from user's
directories.
--
================================
Jeffrey Bacon
================================
Administrator, Breakfast.ca
Student, Carleton U.
Java Programmer, Extrordinaire!
================================
[EMAIL PROTECTED]
http://www.breakfast.ca/~jjbacon
------------------------------
From: Richard Lewis <[EMAIL PROTECTED]>
Subject: Extended partition types (05, 0f & 85?)
Date: Mon, 30 Apr 2001 18:36:58 +0100
I have two hard disks both with extended partitions. These extended
partitions are currently of the type '05'. Should I change them to
'85', the 'Linux extended' partition type. I only run linux on this
box (Redhat 7.1), and I have no intention of ever ever installing
anything even remotely Micro$oft related.
Basically I just wanna know if there are any advantages to the '85'
partition type, and whether its safe to change it using sfdisk (or
something else?). Is there any chance I could lose data?
TIA
Rich Lewis
------------------------------
Date: Mon, 30 Apr 2001 18:41:13 +0100
From: John <[EMAIL PROTECTED]>
Subject: Re: Extended partition types (05, 0f & 85?)
If you're running Linux exclusively, why not use only Linux Native (83)
partitions.
Richard Lewis wrote:
>
> I have two hard disks both with extended partitions. These extended
> partitions are currently of the type '05'. Should I change them to
> '85', the 'Linux extended' partition type. I only run linux on this
> box (Redhat 7.1), and I have no intention of ever ever installing
> anything even remotely Micro$oft related.
>
> Basically I just wanna know if there are any advantages to the '85'
> partition type, and whether its safe to change it using sfdisk (or
> something else?). Is there any chance I could lose data?
>
> TIA
> Rich Lewis
------------------------------
Date: Mon, 30 Apr 2001 10:51:07 -0400
From: zak <[EMAIL PROTECTED]>
Subject: KDE editors problem
Hi,
Im using Mandrake 8.0 and KDE 2.1.1, when i use 'text editor' ,
'advanced text editor' or the kde viewer i dont get a good resolution
for the fonts, its barely recognizable. I also get the same result when
i use Konqeror browser, if i want to compose an email or type in the
search field in a search engine i get the same low quality fonts....what
can be the problem?
(I remember i had this problem with redhat 6.1 but only when i use the
simple text editor not the advanced text editor so i never
bothered)...thanks for your help.
------------------------------
From: "Dr. Aldo Medina" <[EMAIL PROTECTED]>
Subject: Re: killing a process
Date: Mon, 30 Apr 2001 11:41:16 -0600
Jean-David Beyer wrote:
>
> Mladen Gavrilovic wrote:
> >
> > Jean-David Beyer wrote:
> > > Look at it in top and see if its STAT is D. If so, the only way to
> > > kill it is to reboot, unless you can make whatever device it is
> > > waiting for let go. Sometimes waiting an hour or so will do it,
> > > because there are some very long timeouts in there for some devices.
> >
> > It's STAT is R. Here's the complete line:
> > 1747 root 16 0 80 4 4 R 94.9 0.0 6:38 cdda2wav
>
> I see no reason why kill -9 1747, run by root, would not terminate
> that process. Since I assume you have tried that, I am as mystified as
> you are.
> >
> > > I have that problem with a tape drive, and the MT command alludes to
> > > changing the timeouts to more realistic values. Unfortunately, the MT
> > > command does not accept the arguments required.
> >
> > What is the MT command? Which package is it it a part of?
> >
>
> I hope I did not confuse you by typing mt in upper case. I just wanted
> it to stand out a little more. I would have used bold, but html
> posting on this newsgroup is deprecated. In any case, if you run a Red
> Hat based system, it is in:
>
> valinux:jdbeyer[~]$ rpm -qf /bin/mt
> mt-st-0.5b-7
> valinux:jdbeyer[~]$
>
> There is a manual page for it. man mt
>
> In my case, the process does not hog the CPU; it just gets into state
> D and never progresses, never unlocks the file, etc. I must reboot to
> regain control of the tape drive. Since it also happens with hard
> drives on a different controller when I run rpm sometimes, I have
> reduced my belief that it is a tape drive (or its SCSI controller)
> problem. I have run lots of memory tests, including several hours of
> memtest86, I suspect memory errors less and less. The memory is 100MHz
> ECC SDRAM.
>
> .~. Jean-David Beyer
Why don't you try ps -f to see the PPID?. Maybe xcdroast called a
process which in turn called cdda2wav. Then you should try to kill -9
this PPID
--
Linux User #98419 -o) | Similitudes entre mujeres y
http://counter.li.org /\ | computadoras: 8. Si tienes una
ICQ 94335020 _\_v | buena, te tienen envidia 9. Las dos
Si quieres ayudarme, ponme de | acaban volvi�ndote loco
referencia en www.puntosclub.com |
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Disc Partitioning & / ( root )
Date: Mon, 30 Apr 2001 19:26:16 +0200
Chris Leahy <[EMAIL PROTECTED]> wrote:
> Sony Anthony wrote:
>> 2. All bootable partitions should be a primary partition
> True
False.
>> 3. All bootable partitions should start below 1024 MB
> Depends on your system
False. Depends on the bios plus the bootloader.
> This just isnt true. I'm currently booting / from well above 1024
True ;-)
Peter
------------------------------
From: [EMAIL PROTECTED] (Lew Pitcher)
Subject: Re: startx -- error
Reply-To: [EMAIL PROTECTED]
Date: Mon, 30 Apr 2001 17:53:51 GMT
<editorial type="comment" voice="sardonic">
I'm surprised that you are even contemplating allowing your users to
run X. You _are_ aware that it is more of a security risk than
letting your users cd through your filesystem (a risk you
found exceedingly unacceptable a couple of weeks ago).
</editorial>
On Mon, 30 Apr 2001 17:36:11 GMT, "Jeffrey J. Bacon"
<[EMAIL PROTECTED]> wrote:
>Authentication failed - cannot start X server
>Perhaps you do not have console ownership?
So, how are you starting X? Are you
a) running 'startx' from a commandline,
b) running 'X' from a commandline,
c) running 'gdm', or
d) running 'kdm'
?
Remember, the user process (X) has to be able to open the local
display (/dev/console) for read/write access. If you are using gdm,
then take a look at /etc/X11/gdm/GiveConsole and
/etc/X11/gdm/TakeConsole
>the above error appear when calling startx from a user account (root
>works fine).
>
>I'm on RedHat7, I have just wiped out KDE from my system and put it back
>on (ONLY: libs, libs-sound, support, base) as root. I started it as
>root and copied root's .Xclients, .Xclients-default to the user's home
>directories but the above error occurs when trying to start from user's
>directories.
Lew Pitcher, Information Technology Consultant, Toronto Dominion Bank Financial Group
([EMAIL PROTECTED])
(Opinions expressed are my own, not my employer's.)
------------------------------
From: Richard Lewis <[EMAIL PROTECTED]>
Subject: Re: Extended partition types (05, 0f & 85?)
Date: Mon, 30 Apr 2001 18:55:36 +0100
My data partitions (i.e. another primary partition on each drive and
the logical partitions within the extended partitions) are Linux
Native (83)
I just want to know if its advantageous to change the extended
partitions themselves to 'Linux Extended' (type 85) or if I should
just keep them as 'Normal Extended' (type 05).
On Mon, 30 Apr 2001 18:41:13 +0100, John <[EMAIL PROTECTED]> wrote:
>If you're running Linux exclusively, why not use only Linux Native (83)
>partitions.
>
>Richard Lewis wrote:
>>
>> I have two hard disks both with extended partitions. These extended
>> partitions are currently of the type '05'. Should I change them to
>> '85', the 'Linux extended' partition type. I only run linux on this
>> box (Redhat 7.1), and I have no intention of ever ever installing
>> anything even remotely Micro$oft related.
>>
>> Basically I just wanna know if there are any advantages to the '85'
>> partition type, and whether its safe to change it using sfdisk (or
>> something else?). Is there any chance I could lose data?
>>
>> TIA
>> Rich Lewis
------------------------------
Date: Mon, 30 Apr 2001 19:23:48 +0100
From: John <[EMAIL PROTECTED]>
Subject: Re: Extended partition types (05, 0f & 85?)
I could be wrong, but I think that type 05 partitions are supposed to
start on a cylinder boundary, I.E. head 0, whereas 85's can start on,
for example, cylinder 123, head 5.
Unless somebody has very obscure requirements, (e.g. an laptop with very
limited hard disk options), there seems to be no practical reason to use
one over another.
> I just want to know if its advantageous to change the extended
> partitions themselves to 'Linux Extended' (type 85) or if I should
> just keep them as 'Normal Extended' (type 05).
------------------------------
Date: Mon, 30 Apr 2001 19:26:48 +0100
From: John <[EMAIL PROTECTED]>
Subject: Re: Disc Partitioning & / ( root )
Don't forget that although most hard disks have a 512 byte sector size,
there are some 2K sector size devices around, (although I've only ever
seen that with SCSI devices). Therefore it's best to think in terms of
cylinders, heads and sectors, or blocks, rather than "the first 8 gigs"
or whatever.
> >> 3. All bootable partitions should start below 1024 MB
>
> > Depends on your system
>
> False. Depends on the bios plus the bootloader.
------------------------------
Date: Mon, 30 Apr 2001 20:35:59 +0200
From: Michael Heiming <[EMAIL PROTECTED]>
Subject: Re: Best way to manage source code?
MH wrote:
>
> Aside from compiling kernels, I generally don't compile from source.
> Generally, I rely on RPM or DEB packages. Originally, I did this because I
> wasn't familiar with the process of compiling source. Also, I wasn't aware
> of some of the benefits of compiling from source--namely, customization. I
> also wasn't aware that it was generally possible to uninstall compiled
> packages as easily as it was RPM or DEB packages.
>
> I have since learned more about compiling, installing, and uninstalling
> source packages, but effective maintenance, and ease of maintenance, seems
> to depend greatly on how you manage source packages. For example, there is
> the issue of space. There is also the issue of complexity. By that I
> mean that some packages (KDE or GNOME, e.g.) actually consist of dozens of
> individual source packages, some of which may have to be compiled in a
> specific order?
>
> I am sure there are other issues, for example compiling for specific
> kernels or libraries? Is there a good guide available on how to manage
> source packages? Ideas, comments, pointers will be appreciated.
Don't know of any, I use to unpack source tarballs in /usr/local/src and
install (mostly) to /usr/local/...
If you choose to install from source, it's IMHO presumed, that you
know what you do, read the docs that come with the sources, check the
configure/Makefile for anything curious or unusual, perhaps
edit it and take a look at the sources. After this you decide to
get it running or not compile/install it at all...
Good luck
Michael Heiming
------------------------------
From: Sony Anthony <"ffdfptz"@[EMAIL PROTECTED]>
Subject: Re: Disc Partitioning & / ( root )
Date: Mon, 30 Apr 2001 14:58:52 -0400
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Sorry I m a bit confused here. The Mini-Howto on partitioning says every
bootable partition should be a primary.
<br>Is that an outdated howto.
<br>Also given that I use LILO, is there any way I can look at the
BIOS and determine if my machine can boot below 1024. ( I could not
find anything that says this directly ). Or is it to be found out by trying
out on a machine to machine basis.
<br>Thanks
<br>sony
<p>"Peter T. Breuer" wrote:
<blockquote TYPE=CITE>Chris Leahy <[EMAIL PROTECTED]> wrote:
<br>> Sony Anthony wrote:
<p>>> 2. All bootable partitions should be a primary partition
<p>> True
<p>False.
<p>>> 3. All bootable partitions should start below 1024 MB
<p>> Depends on your system
<p>False. Depends on the bios plus the bootloader.
<p>> This just isnt true. I'm currently booting / from well above 1024
<p>True ;-)
<p>Peter</blockquote>
</html>
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: killing a process
Date: Mon, 30 Apr 2001 15:08:58 -0400
"Dr. Aldo Medina" wrote:
>
> Jean-David Beyer wrote:
> >
> > Mladen Gavrilovic wrote:
> > >
> > > Jean-David Beyer wrote:
> > > > Look at it in top and see if its STAT is D. If so, the only way to
> > > > kill it is to reboot, unless you can make whatever device it is
> > > > waiting for let go. Sometimes waiting an hour or so will do it,
> > > > because there are some very long timeouts in there for some devices.
> > >
> > > It's STAT is R. Here's the complete line:
> > > 1747 root 16 0 80 4 4 R 94.9 0.0 6:38 cdda2wav
> >
> > I see no reason why kill -9 1747, run by root, would not terminate
> > that process. Since I assume you have tried that, I am as mystified as
> > you are.
> > >
> > > > I have that problem with a tape drive, and the MT command alludes to
> > > > changing the timeouts to more realistic values. Unfortunately, the MT
> > > > command does not accept the arguments required.
> > >
> > > What is the MT command? Which package is it it a part of?
> > >
> >
> > I hope I did not confuse you by typing mt in upper case. I just wanted
> > it to stand out a little more. I would have used bold, but html
> > posting on this newsgroup is deprecated. In any case, if you run a Red
> > Hat based system, it is in:
> >
> > valinux:jdbeyer[~]$ rpm -qf /bin/mt
> > mt-st-0.5b-7
> > valinux:jdbeyer[~]$
> >
> > There is a manual page for it. man mt
> >
> > In my case, the process does not hog the CPU; it just gets into state
> > D and never progresses, never unlocks the file, etc. I must reboot to
> > regain control of the tape drive. Since it also happens with hard
> > drives on a different controller when I run rpm sometimes, I have
> > reduced my belief that it is a tape drive (or its SCSI controller)
> > problem. I have run lots of memory tests, including several hours of
> > memtest86, I suspect memory errors less and less. The memory is 100MHz
> > ECC SDRAM.
> >
> > .~. Jean-David Beyer
>
> Why don't you try ps -f to see the PPID?. Maybe xcdroast called a
> process which in turn called cdda2wav. Then you should try to kill -9
> this PPID
>
For me, it is not xcdroast, bug usually either bru (commercial
backup-recovery utility), IBM's DB2 UDB (commercial relational dbms),
or rpm (Red Hat's thingie). Doing ps -f is no use because it is easy
enough to see the parents, but killing them makes no difference, since
the children do not care what their parents are doing (the parents are
usually hanging on a wait(2) for their children to die, which they
never will).
Since I replaced the second HP C1599A DDS-2 tape drive with an Ecrix
VXA-1 tape drive, this has never happened, but this is not yet
statistically significant since I did that only 9 days (almost 10)
ago.
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey http://counter.li.org
^^-^^ 3:05pm up 9 days, 21:29, 3 users, load average: 0.18, 0.13, 0.10
------------------------------
** 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 by posting to comp.os.linux.misc.
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
******************************
