Linux-Setup Digest #723, Volume #20 Wed, 28 Feb 01 07:13:12 EST
Contents:
Re: glibc-2.2-9.i686.rpm (David Cecere)
Re: Info on getting Qwest DSL working ("Ben")
Re: linux on a notebook ("michael.fengler")
Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1 ("Pavan")
Re: Embarrassingly simple newbie question.... (Julian Midgley)
Re: How to install multiple distros ("Mark L. Kahnt")
Re: ssh on Debian, not connecting to OpenBSD 2.7 ["Disconnecting: Bad packet length
1349676916"] (Markus Friedl)
Re: glibc-2.2-9.i686.rpm (=?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?=)
Re: graphics messed up on recompiled kernel (RH7, 2.2-16)
(=?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?=)
Re: ntfs and linux-2.4.0 (=?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?=)
Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1
(=?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?=)
Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1
(=?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?=)
Help with LILO and ATA 100 on dual (Luigi Cavallo)
Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1 ("Pavan")
Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1
(=?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?=)
Setting File/Directory Permissions - newbie question. ("Greg Hains")
----------------------------------------------------------------------------
From: David Cecere <[EMAIL PROTECTED]>
Subject: Re: glibc-2.2-9.i686.rpm
Date: Wed, 28 Feb 2001 01:31:18 -0500
Reply-To: [EMAIL PROTECTED]
The latest version is glibc-2.2-12. Try Red Hat's site again using the
latest version number.
On Wed, 28 Feb 2001 02:42:53 GMT, "Paul Folbrecht" <[EMAIL PROTECTED]>
wrote:
}I want to patch my RH 7 install with the glibc fix, and the link to the
file
}is broken on Red Hat's site! I tried both Netscape under Linux and IE in
}W2K.
}
}Does anyone know another place I can get it? The file is
}glibc-2.2-9.i686.rpm. Thanks.
}
}
_____________________________________________________________________________
"Thinking is, or ought to be, a coolness and a calmness; and our poor hearts
throb, and our poor brains beat too much for that."--Herman Melville
My r�sum� and etc.: http://4.34.161.50/
My patent: http://4.34.161.50/widget/
------------------------------
From: "Ben" <[EMAIL PROTECTED]>
Subject: Re: Info on getting Qwest DSL working
Date: Wed, 28 Feb 2001 00:39:43 -0700
Thanks for the info, though I can honestly say I'm dissapointed. I've
already got the modem and the external one is lots of money. I appreciate
the reply.
Ben
------------------------------
From: "michael.fengler" <[EMAIL PROTECTED]>
Subject: Re: linux on a notebook
Date: Wed, 28 Feb 2001 08:12:25 +0100
Reply-To: Michael Fengler <[EMAIL PROTECTED]>
On Tue, 27 Feb 2001, H Dziardziel wrote:
>On Tue, 27 Feb 2001 10:31:45 +0100, "Ove Hvam Andersen"
><[EMAIL PROTECTED]> wrote:
>
>>Hi All
>>I'm planning to install Linux on my HP Omnibook xe2, with a Xircom realport
>>Cardbus ethernet an modem card.
>>Are there a prefered ver/distribution that works best on a notebook ?
>>Are there things I should be aware off ?
>>
>>MVH
>>Ove Hvam Andersen
>>
>>
>You will need the latest pcmcia version with the tulip driver and it
>may be very flakey still apparently. The modem works so far on mine
>but I have not verified the ethernet yet. Anyway in general the card
>looks nice with the leds and the realport is great but speeds stink.
Xircom RealPort is also supported by the kernel directly (has been since
2.3.99 at least), so you might be able to do without pcmcia-cs. This
gives me rock-steady ethernet but I cannot use the modem (gets detected
when booting, but for the life of me I cannot find the interface to
use it).
------------------------------
From: "Pavan" <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1
Date: Wed, 28 Feb 2001 13:32:50 +0530
> So, I 'shift-PageUp' and look at the kernel
initialization with the hard
> drive detection and I see something out of the
ordinary that I've not seen
> before, and definately don't see when I boot
7.2:
> Partition Check:
> /dev/ide/host0/bus0/target0/lun0/P1 P2 <P5
P6.....P11>
>
Looks like you compiled with the new devfs
support(under filesystems, if I remember) & *no*
old-style dev supported. May be installing the
package devfsd will solve this.
I'm not too sure about all this, but there has
been a change in the device files standard in
kernel 2.4. I'm pretty sure your problem is
something to do with this. Compile with the
old-style support enabled.(don't remember what it
is called but I think it is under filesystems)
-Pavan
------------------------------
Subject: Re: Embarrassingly simple newbie question....
From: [EMAIL PROTECTED] (Julian Midgley)
Date: Wed, 28 Feb 2001 08:33:01 GMT
In article <[EMAIL PROTECTED]>,
Scott Gardner <[EMAIL PROTECTED]> wrote:
>On 28 Feb 2001 00:28:45 GMT, [EMAIL PROTECTED] (DTi4565459)
>wrote:
>
>>>What is the best way from the command line to find out where a
>>>particular file resides on my hard drive? I've noticed that "ls"
>>>seems to have some fundamental differences from the DOS "dir" command.
>>
>>try locate filename*
>> dave
>>
>>http://www.columbia.edu/~mdt1/
>>
>>(1 = one, not little L; and don't forget the trailing / )
>
>Thanks. Found "locate" a little after I posted my question. Thanks
>for all the replies.
It's important to note the difference between 'locate' and 'find'.
'locate' searches an index of the files on your system which is
typically updated nightly by a cron job (running a program called
updatedb). 'locate' will not, therefore, be able to find files which
have been created since updatedb was last run, and similarly, may
find files which don't exist, if they've been deleted since the last update.
In addition, the updatedb index won't necessarily contain entries for
all files on your system. /etc/updatedb.conf configures which paths
are excluded from the index.
That said, since 'locate' makes use of a pre-computed index, it will
complete very quickly.
'find', on the other hand, actually looks at the filesystem, stat()ing
files and directories in order to do it's work. It is, therefore,
slower, but should always be accurate (modulo race conditions with
processes deleting/moving files).
All the best,
Julian Midgley
--
Julian Midgley
Principal Consultant [EMAIL PROTECTED]
Zeus Technology Ltd http://www.zeus.com
------------------------------
From: "Mark L. Kahnt" <[EMAIL PROTECTED]>
Crossposted-To: linux.debian.user
Subject: Re: How to install multiple distros
Date: Wed, 28 Feb 2001 03:53:22 -0500
Reply-To: [EMAIL PROTECTED]
Rod Smith wrote:
>
> In article <[EMAIL PROTECTED]>,
> "Mark L. Kahnt" <[EMAIL PROTECTED]> writes:
> >
> > I'm honestly not sure if /boot could be shared
>
> I've done it. The main problem is that some of the LILO-related files,
> like /boot/chain.b, may be for different versions. This means you may
> only be able to run the lilo program from one distribution (LILO can
> still boot both distributions, though; you can only INSTALL it from
> one). I believe there was also some issue with the System.map file, but
> IIRC I resolved that by using symbolic links to the System.map files in
> the /usr/src/linux directory tree. I suppose a similar solution would
> work for chain.b and other LILO files, but that would require mucking
> with stuff installed via RPM or Debian packages.
>
> On the whole, it's not worth the bother to share /boot. That partition,
> if it's separate at all, is so small that it's not worth the hassle. The
> only reason I can think of to try it would be if you somehow need
> primary partitions for both distributions' /boot partitions, and you're
> running out of them.
>
> --
> Rod Smith, [EMAIL PROTECTED]
> http://www.rodsbooks.com
> Author of books on Linux & multi-OS configuration
Yeah - roughly my expectations of that - I can imagine that it would be
a horror if you were using a custom kernel and modules, sharing /boot
but not /lib/modules...
The only benefit I can see is if the kernel supported some hardware
under only one distribution due to a special add-on that Linus hadn't
yet adopted, and you wanted to try that support under other
distributions, but I think it would make more sense to re-compile that
source code against the specific distribution *just in case* and
squander, what, one whole MB for a separate /boot?
------------------------------
From: [EMAIL PROTECTED] (Markus Friedl)
Crossposted-To: comp.security.ssh,comp.unix.bsd.openbsd.misc
Subject: Re: ssh on Debian, not connecting to OpenBSD 2.7 ["Disconnecting: Bad packet
length 1349676916"]
Date: 28 Feb 2001 09:47:47 GMT
In <97h8rt$c1i$[EMAIL PROTECTED]> J Phillips <[EMAIL PROTECTED]> writes:
>In comp.os.linux.setup Richard E. Silverman <[EMAIL PROTECTED]> wrote:
>> > Disconnecting: Bad packet length 1349676916
>--------------------------------^
>> 1349676916 decimal = 50726F74 hex = "Prot" ASCII
>> Looks suspiciously like some sort of text message, perhaps beginning
>> "Protocol" or "Protection"? What do the syslog messages on the server
>> say? How about if you try the connection with the server in debug mode?
>> --
>> Richard Silverman
>> [EMAIL PROTECTED]
>I think it disconected because if assumes a 1.3GB packet is some sort
>of attack.
no, Richard is right.
apart from this, 1.3GB packets are not legal in the SSH protocol.
-m
------------------------------
From: =?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?= <[EMAIL PROTECTED]>
Subject: Re: glibc-2.2-9.i686.rpm
Date: Wed, 28 Feb 2001 11:19:07 +0100
On Wed, 28 Feb 2001, Paul Folbrecht wrote:
> I want to patch my RH 7 install with the glibc fix, and the link to the file
> is broken on Red Hat's site! I tried both Netscape under Linux and IE in
> W2K.
>
> Does anyone know another place I can get it? The file is
> glibc-2.2-9.i686.rpm. Thanks.
ftp.redhat.com/pub/redhat/updates/7.0
Btw. there is a 2.2-12...
Rasmus
------------------------------
From: =?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?= <[EMAIL PROTECTED]>
Subject: Re: graphics messed up on recompiled kernel (RH7, 2.2-16)
Date: Wed, 28 Feb 2001 11:22:55 +0100
On 28 Feb 2001, Jeremy Schneider wrote:
> When I install Redhat 7 on my system and use the kernel that is
> automatically installed, things are displayed just fine. However, when
> I recompile the kernel and run with the new kernel, the display is
> slightly (but very annoyingly) messed up: The icons at the bottom are
> displayed as garbage; The "embossed" RedHat logo seems to be smeared
> across the screen; windows with any graphics in them come up as mostly
> garbage.
/* SNIP */
> Hardware system config:
> Motherboard: ABIT KT7-Raid
> CPU: Athlon 1000 MHz,
> Graphics Card: Hercules (Guillemot) GE Force GTS Pro 64 MB
> Monitor: Viewsonic E790
> RedHat 7.0.
Are you by any chance using the nvidia drivers for your graphics card
instead of the standard nv, that comes with XFree86? Then you will have
to recompile the nvidia drivers for the new kernel.
What is in /var/log/messages when starting X? What is in
/var/log/xdm-errors (or the output from X)?
Rasmus
------------------------------
From: =?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?= <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.help
Subject: Re: ntfs and linux-2.4.0
Date: Wed, 28 Feb 2001 11:26:28 +0100
On Wed, 28 Feb 2001, it was written:
> >After compiling and installing the kernel-2.4.0 it's impossible to list
> >the files on an ntfs(win2k) filesystem. The ntfs module is compiled in
> >the kernel, it mounts the partition without any errors, but when I go to
> >the mount-point and do an "ls", there is not list of the files in that
> >directory. What am I missing? Under 2.2.17, it workd fine without a
> >problem.
> >
> >mount command: mount -t ntfs /dev/sda1 /win2k -o ro
> >under: Linux mandrake 7.2
> >kernel-2.4.0 + upgraded modutils, binutils,... etc.
Do you get any kernel errors ('dmesg')? What is the output from ls? Can
you list free space with df (broken for NTFS fro w2k, but should show
something). Is it mounted at all ('mount')?
Are you sure, there are no errors on the NTFS?
Rasmus
------------------------------
From: =?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?= <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1
Date: Wed, 28 Feb 2001 11:29:17 +0100
On Wed, 28 Feb 2001, Rake@Q3 wrote:
> E2fsck 1.19....etc...Superblock could not be read or doesn't describe a
> correct ext2 FS ..../dev/hda/
Yikes. I got the same thing on an updated RedHat 7.0 system and my FS
was actually OK.
I found it was the broken version of gcc on RedHat 7.0 - run a decent
stable one (gcc 2.95.2) - it solved the problem for me.
Rasmus
------------------------------
From: =?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?= <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1
Date: Wed, 28 Feb 2001 11:34:55 +0100
On Wed, 28 Feb 2001, Rasmus B�g Hansen wrote:
> On Wed, 28 Feb 2001, Rake@Q3 wrote:
>
> > E2fsck 1.19....etc...Superblock could not be read or doesn't describe a
> > correct ext2 FS ..../dev/hda/
>
> Yikes. I got the same thing on an updated RedHat 7.0 system and my FS
> was actually OK.
>
> I found it was the broken version of gcc on RedHat 7.0 - run a decent
> stable one (gcc 2.95.2) - it solved the problem for me.
Eh, forgot to mention, that you should recompile e2fsprogs - and the
kernel too, I recommend.
Rasmus
------------------------------
From: Luigi Cavallo <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Help with LILO and ATA 100 on dual
Date: Wed, 28 Feb 2001 11:41:25 +0100
Hi,
I just installed RH 7.0 on a Dual PIII with ATA 100 HDs on the ATA100
ide2.
However, when I try to reboot from the HD the box hangs before arriving
at the lilo prompt. Actually, it hangs with the first two uppercase
charcaters of lilo, as
LI
If I boot from the floppy made during installation, and by feeding linux
ide2=...
it boots properly.
I tried to install lilo both in the MBR as well as on /dev/hde, but the
box behaves the same.
Any suggestions ?
HW
MicroStar 694D MB
2 PIII CPU
2 30 GB Quantum Fireball lct20 ATA 100 HDs on ide2 and ide3
TIA
luigi
------------------------------
From: "Pavan" <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1
Date: Wed, 28 Feb 2001 16:55:36 +0530
> > I found it was the broken version of gcc on
RedHat 7.0 - run a decent
> > stable one (gcc 2.95.2) - it solved the
problem for me.
>
> Eh, forgot to mention, that you should recompile
e2fsprogs - and the
> kernel too, I recommend.
>
I managed to compile the 2.4.2 kernel without
problems & its running smoothly ( on a RH7 with
all the updates ). I did not have to recompile
e2fsprogs.
Pavan
------------------------------
From: =?ISO-8859-1?Q?Rasmus_B=F8g_Hansen?= <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.4.2 and 'all blowed up' on e2fsck check of /dev/hda1
Date: Wed, 28 Feb 2001 12:32:51 +0100
On Wed, 28 Feb 2001, Pavan wrote:
> > > I found it was the broken version of gcc on
> RedHat 7.0 - run a decent
> > > stable one (gcc 2.95.2) - it solved the
> problem for me.
> >
> > Eh, forgot to mention, that you should recompile
> e2fsprogs - and the
> > kernel too, I recommend.
> >
>
> I managed to compile the 2.4.2 kernel without
> problems & its running smoothly ( on a RH7 with
> all the updates ). I did not have to recompile
> e2fsprogs.
I had gcc 2.96-69 (the latest update I think) installed and compiled
e2fsprogs 1.19. It gave the described error.
Then I compiled 1.18 - same error.
Then I installed gcc 2.95.2 and recompiled e2fsprogs 1.19 - no
problems...
gcc 2.96 is in beta and does not work with every program.
My kernel did actually compile and run with gcc 2.96 - I just don't feel
safe with that crappy compiler.
Rasmus
------------------------------
Reply-To: "Greg Hains" <[EMAIL PROTECTED]>
From: "Greg Hains" <[EMAIL PROTECTED]>
Subject: Setting File/Directory Permissions - newbie question.
Date: Wed, 28 Feb 2001 21:13:16 +0930
Hi,
Could somebody please assist me with setting security on directories and
files within Linux (Mandrake 7.2 specifically). I am new to Linux but have
been shown how to use CHMOD and CHOWN and understand the 777 (eg.) format.
I have been working with Netware and Windows NT security for quite a number
of years now, and am able to apply any kind of security of various users and
groups to the file system. My problem lies with the fact that with the above
two OSs I can grant or deny access to any number of users and/or groups
_simultaneously_ to a directory or file. Unless I misunderstand I can only
see that there are three simultaneous rights that can be assigned to a
directory/file (owner,group,all). I am sure that Linux is not limited by
this initial observation, but I am damned if I can get my head around it.
I have several Linux books and have seen some good web sites which explain
the CHMOD/CHOWN usage, but not how to apply it on a larger scale.
Lets say that I have the following users: (over and above root)
"greg", "john", "jack", "mary", "nic", "webmaster" and that they are
assigned to a primary group of the same name.
And that we have the following sample directory structure (let's ignore
rights to individual files right now and we only want to set permissions on
directories.)
Sample directory structure (there are more directories, yes, but I've no
need to changed permissions on other system directories right now)
/var
---/www
=====/html
=======/weblive
=========/nic
=======/greg
=======/client1
=======/client2
=======/john
=======/jack
=======/mary
=====/htmltest
=======/client3
=======/client4
/home
---/greg
---/john
---/jack
---/mary
---/nic
*---*
* The /var/www/html/weblive directory is our live website.
* I am "greg" (strange that!) and that I want to be able to do anything to
any file on the system as I will be maintaining the server overall. (This
includes by default \home\greg)
* Is there a situation where "root" can be locked out? I dont want this to
happen, but by assigning owners of various directories, I assume thet "root"
will always have rights.
* "john", "jack" and "mary" are colleagues of mine and need rights to their
/home/%username% directory (scuse the DOS expression).
"john" needs to give rights to "mary" so that she may see his /home/john
directory, but otherwise no other user should be able to see the contents of
the /home directory of any other user (with exception of "greg" being able
to RWX all directories.)
* "nic" only FTPs in and uploads files to the /var/www/html/weblive/nic
directory. He is only allowed rights to this directory and needs no orther
rights - especially up (/var/www/html/weblive) where he could damage the
live site.
* While others can read the live website, no-one else but "greg" (and
"root" - inferred) has rights to change/delete anything (bar "nic" in the
subdirectory below it)
* "john" and "mary" are to have RWX rights to
/var/www/html/html/test/client3, but only "john" is to have RWX rights to
/var/www/html/webtest/client4.
This is not a complex request. Once I understand the concepts behind it, I
will be able to apply it to my needs. Can somebody please explain how I use
groups and ownerships with CHMOD/CHOWN. Even just a "ls -l" sample of this
ideal tree would look like.
Thankyou.
Greg Hains
------------------------------
** 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
******************************
