Linux-Setup Digest #452, Volume #20 Fri, 19 Jan 01 05:13:07 EST
Contents:
Re: New Hard drive install... HELP!! ("James Ayton, Jr.")
Re: recompiling kernel and PCMCIA support - how? (David Hinds)
Re: Kickstart creation (Josef Moellers)
Missing partitions on second hard disk (=?iso-8859-1?Q?J=FCrgen?= Simonsen)
Re: ISO image copies ([EMAIL PROTECTED])
Re: partitioning failing during install of RH7.0 (Andrew McLean)
Re: recompiling kernel and PCMCIA support - how? ([EMAIL PROTECTED])
Re: DiskImage of newly installed Linux-Box (media factory)
Re: Why no real-support for ext2fs by DriveImage et al. ? (media factory)
RH7 upgrade problem ("Henning Pl�ger")
Re: partitioning failing during install of RH7.0 ([EMAIL PROTECTED])
Re: Help needed with installing new 2.4 kernel ([EMAIL PROTECTED])
system logs? ("Greg Gailer")
Re: What is dnetc?? ("Robin")
----------------------------------------------------------------------------
From: "James Ayton, Jr." <[EMAIL PROTECTED]>
Subject: Re: New Hard drive install... HELP!!
Date: Fri, 19 Jan 2001 02:18:06 -0500
well, I did 'fdisk /mbr' in Win98 command prompt, and everything worked fine
thanks for the help
"David Efflandt" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> On Tue, 16 Jan 2001, James Ayton, Jr. <[EMAIL PROTECTED]>
wrote:
> >My set up is as follows...
> >
> >I have an 18 gig hard drive with Windows 2000 and Windows 98 dual
booting,
> >this drive is C:\, and is connected to my Ultra ATA66 controller.
> >
> >I have a 4 gig drive with Corel Linux that is connected to the IDE
> >connector on the mother board. The Linux controls the MBR, as Corel's OS
> >loader pops up on start up, and I can select either Linux or Windows. If
I
> >select Windows, it goes to Windows 2000 OS Loader to select between 2000
and
> >98.
> >
> >I want to add a new 45 Gig drive as D: to the Ultra ATA66 controller.
> >However, all drives connected to this controller are assigned drive
letters
> >before the IDE connectors on the mother board are. So, if I have them
all
> >connected, all I see are the letters 'LI' when I start up. Is there a
way
> >to tell the MBR that the Corel's Linux OS Loader is on drive E: now?
>
> Sometimes the BIOS will move a drive down from what it actually is to fill
> a vacant slot. The first thing you need to tell us is if Linux thinks it
> is currently booting onto hdb or hdc.
>
> If it does think it is on hdb it is possible to duplicate that section in
> lilo.conf with a different label and with root=/dev/hdc(whatever) instead
> of hdb(whatever) and add an 'unsafe' line (w/o quotes) at the end of that
> section so lilo does not try to find it there when you run /sbin/lilo.
> You would also need to change the /etc/fstab entries from hdb to hdc just
> before connecting the other drive.
>
> But if Linux already thinks it is on hdc, I don't know what to tell you.
>
> --
> David Efflandt [EMAIL PROTECTED] http://www.de-srv.com/
> http://www.autox.chicago.il.us/ http://www.berniesfloral.net/
> http://cgi-help.virtualave.net/ http://hammer.prohosting.com/~cgi-wiz/
------------------------------
From: David Hinds <[EMAIL PROTECTED]>
Subject: Re: recompiling kernel and PCMCIA support - how?
Crossposted-To: comp.os.linux.portable
Date: Fri, 19 Jan 2001 07:27:41 GMT
In comp.os.linux.portable Michal Szymanski <[EMAIL PROTECTED]> wrote:
> 1. Is the order (kernel, pcmcia) important?
Yes, you must do kernel then pcmcia.
> 2. Will the 'make install' in pcmcia directory replace any vital
> configuration files I'm using now (somewhere in /etc/, for example)?
No, it won't touch the *.opts files, which are the only ones that you
should ever modify.
> 3. Do I have to remake pcmcia every time I reconfigure kernel (even if
> the changes do not regards PCMCIA devices)?
It is not always necessary, but it is necessary any time you change a
kernel configuration option that changes any kernel data structures or
interfaces that are used by the PCMCIA modules. So, there is some
subset of kernel configuration options that you can change without
needing to rebuild PCMCIA, but I don't know exactly what that subset
is. Just do the rebuild.
-- Dave
------------------------------
From: Josef Moellers <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.system,comp.os.linux.misc
Subject: Re: Kickstart creation
Date: Fri, 19 Jan 2001 08:38:29 +0100
bill davidsen wrote:
> =
> I have to install a bunch of servers, and for various reasons will be
> using Redhat. I see how to install using a kickstart file, but how do I=
> create the kickstart file? There are references to a utility which take=
s
> the current installation and creates the ks.cfg file based on the
> current install. Obviously there is such a thing, but I sure don't see
> the name of it.
> =
> I have several 600+ page Redhat books, but they all want to tell me how=
> to use a ks.cfg, not how to create one in some way better than by hand.=
There used to be an mkkickstart utility (and package), at least I'v
found one on a 6.1 and a 6.2 CD, but not on the 7.0 CD. It was a shell
script which generated the kickstart file for you. When I did my RHCE
with 6.0, there were some bugs in it (among others, partition sizes and
mouse directive were incorrect). I've never used it, so I can't give any
further details.
-- =
Josef M=F6llers (Pinguinpfleger bei FSC)
If failure had no penalty success would not be a prize (T. Pratchett)
------------------------------
Date: Fri, 19 Jan 2001 08:53:38 +0100
From: =?iso-8859-1?Q?J=FCrgen?= Simonsen <[EMAIL PROTECTED]>
Subject: Missing partitions on second hard disk
Hello all,
I've redhat7 on a 30GB master disk (Western Digital) and NT on a
10GB-slave-disk (Conner). I've installed lilo to swap the disks
before booting NT (swap 80/81) and everything worked well.
But after backup of the NT-partitions on the 30GB hd with
partition magic 5.0 I've got a problem to access a dos-partition
on the master disk from NT. This constellation worked well before the
image copy.
If I start the NT-harddisk manager I don't see the master hard disk
anymore. I could see the the master hard disk before I've copied the
partition.
Has anyone an idea what is going wrong?
Thanks
J�rgen
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: ISO image copies
Date: Fri, 19 Jan 2001 08:25:29 GMT
In article <948b08$f4n$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> Anyone have experience getting one of the .iso images burned onto a
cd? I
> have a HP with cd writer delux, so if anyone can point me into the
general
> area I'd appreciate it. Thanks in advance...
>
> Sent via Deja.com
> http://www.deja.com/
>
Hi,
First of all, read the 'CD-Writing-HowTo'.
Even for my first burn I had only to follow what Winfried Tr"umper
wrote. Really!
As a resume, you have to get (at least) cdrecord (mkisofs forthe time
you will wante to make your own iso) and eventually a graphical front-
end to cdrecord like xcdroast, and follow these steps:
- make sur you have right devices: (/dev/sg* and /dev/loop*) else
(Thanx again to Winfried!):
test `whoami` = 'root' || echo "You must be root to execute the
commands."
cd /dev/
umask -S u=rwx,g=rwx,o-rwx
[ -f loop0 ] \
|| ./MAKEDEV loop \
|| for i in 0 1 2 3 4 5 6 7; do mknod loop$i b 7 $i; done
[ -f sg0 -o -f sga ] \
|| ./MAKEDEV sg \
|| for i in 0 1 2 3 4 5 6 7; do mknod sg$i c 21 $i; done
- if your cd-burner is not scsi, you have to load the compatibility
driver ide-scsi. (and recompile your kernel to get support for it).
Read the CD-Writing-HowTo (specific to cd-writing options) and Kernel-
HowTO (general concepts) to get success in.
- then you're ready to burn:
1> check your burner is seen by your os:
cdrecord -scanbus
2> make a dummy burn to see if your OS support the maximum speed of
your burner:
shell> SCSI_BUS=0 # taken from cdrecord -scanbus if your burner
# is on bus #0
shell> SCSI_ID=6 # if your burner has target #6
shell> SCSI_LUN=0 # in general if you don't hve LUN it's 0,
# else take the one indicated by cdrecord -scanbus.
shell> cdrecord -v speed=2 dev=$SCSI_BUS,$SCSI_ID,$SCSI_LUN \
-data -dummy your_cd_image
Of course, you can write it shorter:
shell> cdrecord -v speed=2 dev=0,6,0 -data -dummy your_cd_image
Once successed, you can burn for reall:
shell> cdrecord -v speed=2 dev=$SCSI_BUS,$SCSI_ID,$SCSI_LUN \
-data your_cd_image
or
shell> cdrecord -v speed=2 dev=0,6,0 -data your_cd_image
Not difficult isn't it?
Good luck!
The Die
Sent via Deja.com
http://www.deja.com/
------------------------------
From: Andrew McLean <[EMAIL PROTECTED]>
Subject: Re: partitioning failing during install of RH7.0
Date: Fri, 19 Jan 2001 08:57:30 +0000
"C. Jarvis" wrote:
> When I try to install RH 7.0, during the partitioning step for a custom
> install...
>
> I am given two choices: use DiskDruid or fdisk
>
> When I choose DiskDruid, my harddrive info is displayed. I can make
> changes to the displayed info, but when I press 'next' to implement these
> changes, the install crashes with a signal 11.
>
> When I choose fdisk, the install crashes right away with a signal 11.
>
> Any ideas? BIOS settings?
>
> I previously had a working install of RH 6.0, but removed it. LILO is
> still installed.
I had this exact problem. I then created the required partitions using
Partition Magic,
and this time the RH7 installation worked OK. When it got to the HDD page, I
chose 'disk druid', it displayed the new partitions, and I just assigned the
mount
points, then hit 'Next'.
It seems that the installer can't create the partitioning itself, at least on
two systems (yours and mine).
AM
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.portable
Subject: Re: recompiling kernel and PCMCIA support - how?
Date: Fri, 19 Jan 2001 09:17:52 GMT
Try the new kernel 2.4.0 it have inside a good support for most pcmcia
type.
www.kernael.org
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Michal Szymanski) wrote:
> Hi all,
>
> I've done kernel recompiling and installing many times already but now
> it is the first time I need to do it on a laptop, so I need PCMCIA
support.
> I didn't even know before that it is not compiled by default.
>
> It's a RedHat 6.2 machine. Kernel sources RPM contains pcmcia sources
> in /usr/src/linux tree. After "plain" kernel recompiling there are no
> pcmcia modules in /lib/modules/... So I guess I should manually enter
> pcmcia... directory and proceed with "make config; make; make
install".
>
> Few simple questions:
>
> 1. Is the order (kernel, pcmcia) important?
>
> 2. Will the 'make install' in pcmcia directory replace any vital
> configuration files I'm using now (somewhere in /etc/, for example)?
> Should I backup anything and restore afterwards?
>
> 3. Do I have to remake pcmcia every time I reconfigure kernel (even if
> the changes do not regards PCMCIA devices)? It seems to be so, because
> when I tried to just copy old /lib/modules/..../pcmcia to the new
kernel
> modules tree, 'depmod -a' resulted in a bunch of "unresolved symbols"
> messages (although both kernels are 2.2.16).
>
> any hints would be appreciated,
>
> regards, Michal.
>
> --
> Michal Szymanski ([EMAIL PROTECTED])
> Warsaw University Observatory, Warszawa, POLAND
>
Sent via Deja.com
http://www.deja.com/
------------------------------
From: media factory <[EMAIL PROTECTED]>
Subject: Re: DiskImage of newly installed Linux-Box
Date: Fri, 19 Jan 2001 10:28:58 +0100
>> You could try something along the lines of dd if=/dev/hdxy bs=4096 |
>> gzip -9c >filename.gz
>>
>> Restoring would be a matter of just cat filename.gz | gzip -dc | dd
>> of=/dev/hdxy bs=4096
>>
>> MST
Thanks, nice idea.
>
> Will this work if restoring to a larger/smaller disk/partition?
>
> --Cameron
It seemd it only works for identical partitions.
Made a quick test with different partition sizes and the result was an
unmountable destination partition after restore. Linux didn't recognize it
as an ext2fs partition any more.
The second thread "Why no real-support for ext2fs by DriveImage et al. ?" is
also quite interesting although no perfect suggestion has been made so far.
Peter
------------------------------
From: media factory <[EMAIL PROTECTED]>
Subject: Re: Why no real-support for ext2fs by DriveImage et al. ?
Date: Fri, 19 Jan 2001 10:28:59 +0100
> On Fri, 19 Jan 2001 00:46:27 +1300, Steve Withers
> <[EMAIL PROTECTED]> wrote:
>> I put the new disk in as master and the old disk in as slave. I instaled
>> the same level of Linux onto the new drive as i was running on the old
>> drive.
>
> Precisely what I do NOT want to do :-) The goal is to have a generic
> Linux image on a file server, and use it to deploy hosts. I guess this
> could be done by booting with eg., and download an image made
> with dump.
That's it, exactly what I'm looking for too!
Could you please explain what TomsRtBt is?
>
> BTW, what about the MBR and the partition table: how come I can
> restore an image on a new hard disk whose geography is different from
> the HD that I used for cloning the original image? Would a whole image
> made with dump also save the MBR?
As far as I understand, diskimages cannot include more than one partition.
That means, that restoring a disk image does not touch the MBR and partition
table. Partitioning has to be setup before. And MBR must reflect where your
kernel (in case of Linux) can be found.
A problem might be with the file system of the partition, in case its size
differs from that of the image. I don't no the details of ext2fs but
normally every filesystem has same kind of FAT which corresponds to the
number of blocks of the partition. This has to be tweaked to the actual size
of the partion by the DiskImage-Software.
An alternative not so elegant solution might be to make a new drive easily
attachable to Linux-Box, partition and format it probably with a shell
script and restore a previously created tar-archive onto the drive. Finally
install a predefined lilo-booter in the MBR. Most things could be made
automatically with shell scripts.
See also my parallel thread DiskImage of newly installed Linux-Box
Peter
------------------------------
From: "Henning Pl�ger" <[EMAIL PROTECTED]>
Subject: RH7 upgrade problem
Date: Fri, 19 Jan 2001 10:39:18 +0100
Every time I try to upgrade from RH6.2 to RH7 I get the error message that
the partitions of the RH6.2 system were not unmounted properly. I should
boot the old system and shut it down so that all partitions were unmounted
correctly. But when I have done all that and I try to upgrade, I get this
error message again.
What's the problem?
Is it possible to deactivate this partition check?
Henning
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: partitioning failing during install of RH7.0
Date: Fri, 19 Jan 2001 09:21:25 GMT
First Try to build partitions with partitions magic ;)
In article <Pine.SUN.3.96.1010118170653.1088A-100000@thetis>,
"C. Jarvis" <[EMAIL PROTECTED]> wrote:
> When I try to install RH 7.0, during the partitioning step for a
custom
> install...
>
> I am given two choices: use DiskDruid or fdisk
>
> When I choose DiskDruid, my harddrive info is displayed. I can make
> changes to the displayed info, but when I press 'next' to implement
these
> changes, the install crashes with a signal 11.
>
> When I choose fdisk, the install crashes right away with a signal 11.
>
> Any ideas? BIOS settings?
>
> I previously had a working install of RH 6.0, but removed it. LILO is
> still installed.
>
>
Sent via Deja.com
http://www.deja.com/
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Help needed with installing new 2.4 kernel
Date: Fri, 19 Jan 2001 09:28:27 GMT
I have found kernel how to but it is obsolete,
Unpack the bz or bz2 file under usr/src/linux and read the "README" it
is very good and simple.
Make xconfig
Make dep
Make install (this is very good, it palce the image in /boot and run
lilo)
Make modules
Make modules_simple
There is a lot documetntation under /usr/src/linux/Documentation ;)
Ste
In article <946s15$2pp$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> Hi,
>
> I'm always facing problems when upgrading kernel on my linux box right
> now I'm upgrading from 2.2.14 (SuSE6.4) to the new 2.4.0 kernel. Till
> now I haven't found any document describing the exact procedure one
> must follow and which problems one can face doing this.
>
> Most of the time I get stuck with module loading problems for ex.
> depmod looking into the wrong directories on creating the modules.dep
> file. Modules that cannot be found during startup etc. etc.....
>
> My questions are:
>
> Is there a HOWTO describing the exact procedure on how one can upgrade
> and correctly install a new kernel ?
>
> Is there also a reference list where I can find which char-major-xx
> belongs to which device ?
>
> How is the modules.conf file created ?
>
> Anyone some info on this ?
>
> Tia,
>
> Eric
>
> Sent via Deja.com
> http://www.deja.com/
>
Sent via Deja.com
http://www.deja.com/
------------------------------
From: "Greg Gailer" <[EMAIL PROTECTED]>
Subject: system logs?
Date: Fri, 19 Jan 2001 21:14:21 +1100
Greetings All
I am using Mandrake 7.2 and I would like to set up my system to log all
pppd output, xinetd output and ipchain output to one file. I also have
autofs running which creates a lot of logs in /var/log/messages. Is there
some way to get them out of there and into another file? I have read the
man page on syslog.conf but it doesn seem to help much. If anyone has any
examples it would make things a whole lot easier.
TIA
Greg
------------------------------
From: "Robin" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.os.linux.misc,uk.comp.os.linux
Subject: Re: What is dnetc??
Date: Fri, 19 Jan 2001 10:09:17 -0000
Agree with you in terms of not being RT. The kernel is RT but user space
(applications are not). In principal it's because the needs of an
application can not pre-empt execution out of the kernel.
"Daniel Robinson" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> On Thu, 18 Jan 2001 16:39:10 GMT, Thomas Rasmussen
> >If it's your own computer and you haven't started it, then I really
> >don't know what it is... but since it is niced to 19, it shouldn't
> >affect any other processes.
>
> This is not 100% correct. Linux is not a real-time operating system.
> Even with a nice of 19 the process will be given the CPU once in a
> while. To test this you can run something like Seti@home or dnetc with
> a nice of 19, and then launch another CPU intensive job such as MP3
> encoding without any nice. On my system Seti@Home still gets around
> 6-8% of the CPU.
>
> This is due to the fact that Linux (and most Unix systems) is
> optimised for average case performance. Users expect there jobs to get
> a bit of work done even if a higher priority job is available in the
> system.
>
> If you want more information on this phenomenon look up the RTLinux
> site at www.rtlinux.org.
>
> Just my two cents (sorry for nit-picking, but as the objective of a
> newwsgroup is to share information I thought I might as well but in)
>
> :-)
>
> Daniel
>
>
> -----------------------------------------------
> "He who breaks a thing to find out what it is,
> has left the path of wisdom" - Gandalf
>
> Web page : http://student.ulb.ac.be/~drobinso
> -----------------------------------------------
------------------------------
** 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.setup.
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-Setup Digest
******************************
