Linux-Setup Digest #703, Volume #19              Tue, 26 Sep 00 14:13:14 EDT

Contents:
  Re: Debian + PHP4 + Sybase (Colin Watson)
  2 x RTL 8139 Cards = doesn't work (Paul Sullivan)
  Re: Kernel panic: No init found. Try the init= option... (H Dziardziel)
  Re: honeypot + syslogd (Damian Menscher)
  Re: Shared memory (shmget) problem on v2.3.51 (Colin Watson)
  Re: Number of groups associated with one user id ... (Colin Watson)
  NEED HELP w/ Linux install and X Server setup on Laptop... (Brandon Hoppe)
  Re: Sharing swap space between windows and linux (Michael Kelly)
  Re: boot linux from lan (Juergen Sauer)
  Re: Second display (John Janmaat)
  Re: upgrade mandrake 6.1 to 7.1 (Kingsley)
  Re: Linux/W98 to Linux/W2000 (eric)
  Re: expired root password (Emiliano)

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

From: [EMAIL PROTECTED] (Colin Watson)
Subject: Re: Debian + PHP4 + Sybase
Date: 26 Sep 2000 16:03:43 GMT

Jari Kohvakka <[EMAIL PROTECTED]> wrote:
>All right. I got the php4 with apt-get source. Then I run ./configure
>--with-sybase=/opt/sybase-11.9.2 --with-odbc
>
>Now it doesn't work with Apache (no, I'm not going to install Apache by
>hand), and the php-binary still doesn't know the
>sybase_connect()-function... Any ideas?

Look in debian/rules (the main Debian build script) in the php4 source
package and you'll see that it passes *loads* of arguments to
./configure, so it would make sense that missing out things like
--with-apxs would break it.

If you just want to build a package quickly, with the standard Debian
options plus a few of your own, editing debian/rules and typing
'debian/rules build' from the top-level source directory is the fastest
way to do it, assuming you've got all the build dependencies installed.
This won't give you a proper package, and won't even install the package
into a temporary directory; it's the moral equivalent of './configure;
make' without the 'make install'. You can copy the files yourself to
wherever they're supposed to go.

(Stop reading here if you only want a quick fix. :))

On the other hand, you might want to do it the full-blown proper way and
get yourself a Debian package with your own customizations. The way to
rebuild a Debian package with your own options and have it work at least
as well with your system as the original is as follows.

  * Make sure you have all the build-essential packages installed (see
    /usr/share/doc/build-essential/list in the package actually called
    'build-essential', or
    http://www.debian.org/doc/packaging-manuals/build-essential).

  * Make sure you have all the packages listed in the Build-Depends:
    line in debian/control in the relevant source package.

  * Install 'fakeroot' as well: it means you don't have to build as
    root.

  * Edit debian/rules and customize it whatever way you want.

  * Edit debian/changelog and add a new entry at the top with the same
    syntax as the earlier ones. Use your own name, today's date, and a
    newer version number - adding .0 to the end, say, will make it
    unlikely to clash with any future official packages, although you'll
    have to rebuild each time new fixes come out from Debian or
    upstream.

  * Type 'dpkg-buildpackage -rfakeroot' in the top-level source
    directory.

  * Wait, or debug if you need to. Eventually it should spit out a .deb
    in the parent directory; check it with 'dpkg-deb -I foo.deb' and
    'dpkg-deb -c foo.deb', then try installing it.

-- 
Colin Watson                                     [[EMAIL PROTECTED]]
"Choose mnemonic identifiers. If you can't remember what
 mnemonic means, you've got a problem." - perlstyle(1p)

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

From: Paul Sullivan <[EMAIL PROTECTED]>
Subject: 2 x RTL 8139 Cards = doesn't work
Date: Tue, 26 Sep 2000 18:20:17 +0200

Hi,

I'm trying to get a PC with 2 100/10 RTL 8139 based cards to recognised
and configure both correctly.

I currently have the kernel compiled to include support for the  '8139,
it detects both eth0 and eth1 and reports their MAC addresses at
startup.

My ifcfg-eth0 and 'eth1 scripts are excecuted and if config shows both
cards to work.  If I try to ping a known host on the subnet of eth0 I
get a situation where the network requests queue up and aren't sent
*ever*.

I reboot (removing) the 2nd card, all works fine.  (eth1's
initialisation is just delated).

Does anyone have a similar experience or perhaps know why I can't get
both working simultaneuosly?

Regards
Paul


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

From: [EMAIL PROTECTED] (H Dziardziel)
Subject: Re: Kernel panic: No init found. Try the init= option...
Date: Tue, 26 Sep 2000 16:21:39 GMT

On Mon, 25 Sep 2000 08:58:42 GMT, jmsalvo <[EMAIL PROTECTED]>
wrote:

>
>
>Kernel 2.2.16
>
>
>Before I rebooted and got the message above, all files seems to have
>"disappared" (but not really. More on later/below). For example, 'ls' or
>'sh' gives a "No such file or directory".
>
>Unfortunately, I did not try '/bin/sh'.( I am saying this as this has
>impotance below)
>
>However, I can pwd, cd, set.
>
>I rebooted, and got the kernel panic specified in the subject of this
>post. The kernel panic shows after mounting the root fs in read-only
>mode and after the "Freeing unused kernel memory: 44kb freed" message.
>
>To get around the kernel panic, I tried, from within the LILO prompt,
>like:
>
>LILO: linux-latest init=/bin/sh debug root=/dev/hda7 panic=10
>
>...without any luck.
>
>
>Now the strange part:
>
>I can mount and see the entire contents of my root partititon (and
>everything else for that matter), if I:
>
>* Boot from CD using the RedHat installation CD, then
>* Select Install
>* Tell the installation to use fdisk to partition hda (I did this only
>so I can mount the root fs), ...
>* Then while the installation is in fdisk ( but not changing anything on
>the partititon tables ), I switch to a console (Alt+F2) for the bash
>prompt and I mount the root fs:
>
>mount -t ext2 hda7 /tmp/disk1
>
>.... and there is my root fs.
>
>* sbin/init is there (I tried executing /tmp/disk1/sbin/init to test if
>it is corrupted. It worked so it is not corrupted)
>* bin/sh is there ( also tried executing /tmp/disk1/bin/sh. Worked. No
>file corruption )
>* etc/inittab is there and I can read it using 'cat
>/tmp/disk1/etc/inittab' )
>
>Verified that the root fs specified in the kernel is /dev/hda7, so I ran
>:
>
>/tmp/disk1/usr/sbin/rdev /tmp/disk1/boot/vmlinuz-latest
>
>and it gave me:
>
>Root 0x0307
>
>...which is correct.
>
>
>All in all:
>
>* Not a LILO problem
>Otherwise, the kernel would not have loaded in the first place and the
>kernel would not have been able to mount the root fs in read-only mode.
>
>* Not a partition table problem.
>Otherwise, I would not have been able to mount and read/execute files
>from my root fs using the procedure above.
>
>* Not a kernel problem.
>Otherwise, rdev would have shown a different value for the root
>partition
>
>
>
>
>Any ideas out there?
>
>What is unique about a root partition / fs that allows it to be mounted
>but not have the kernel read or execute anything?
>
>
>
>Regards,
>
>John Salvo
>
Hi, is the root / under the magic 1024 cylinder so lilo
can find it? The /boot does not have a file system.

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

From: Damian Menscher <[EMAIL PROTECTED]>
Subject: Re: honeypot + syslogd
Crossposted-To: 
alt.computer.security,alt.linux,alt.os.security,comp.os.linux.x,comp.security.firewalls,comp.security.misc,comp.security.unix,linux.redhat.install,linux.redhat.misc
Date: Tue, 26 Sep 2000 16:24:09 GMT

In comp.security.misc David Martin <[EMAIL PROTECTED]> wrote:
> Rasputin wrote:
>> [EMAIL PROTECTED] <Damian Menscher> wrote:
>> >In comp.security.misc jhuman <[EMAIL PROTECTED]> wrote:
>> >> I was reading the article by lance spitzner on builing a honeypot and he
>> >> mentioned something about "recompiling syslogd" to read from a different
>> >> configuration file other than syslog.conf like /var/tmp/conf.  I was
>> >> wondering how that is done?  Any help would be great.
>> >
>> >First off, you _could_ just run it as syslogd -f /var/tmp/conf.  If
>> >you want to be a bit more sneaky, you could leave it running as usual,
>> >but have /etc/syslog.conf point to a different file when it is
>> >starting than what it is during normal operation.
>>
>> Do you mean a symlink? How is that secure?

> I suspect he means that your syslog startup script points /etc/syslog.conf at
> the real configuration file, then starts the syslog daemon, then points
> /etc/syslog.conf at a fake file all using links.  That way the process table
> does not show any giveaway arguments and a vi of syslog.conf shows fake info.
> Mind you, if I saw that the /etc/syslog.conf file was a link I would probably
> suspect something was awry anyway...

Um, no, I never said anything about a symlink.  Your startup could
do something like:

mv /etc/.backup/syslog.real /etc/syslog.conf
syslogd
mv /etc/.backup/syslog.fake /etc/syslog.conf

Quite safe and secure, especially if you do it by hand instead of
automating it in a script.  And no links to give it away.

HTH,

Damian Menscher
-- 
--==## Grad. student & Sys. Admin. @ U. Illinois at Urbana-Champaign ##==--
--==## <[EMAIL PROTECTED]> www.uiuc.edu/~menscher/ Ofc:(217)333-0038 ##==--
--==## Physics Dept, 1110 W Green, Urbana IL 61801 Fax:(217)333-9819 ##==--

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

From: [EMAIL PROTECTED] (Colin Watson)
Subject: Re: Shared memory (shmget) problem on v2.3.51
Date: 26 Sep 2000 16:15:35 GMT

John <[EMAIL PROTECTED]> wrote:
>Tariq Hussain <[EMAIL PROTECTED]> wrote:
>> I've been trying to use kernel v2.3.51, it recompiles fine. The previous
>> version was v2.2.13. I've got all the correct version of additional
>> packages. My problem is I get the following error when I do dmesg:
>
>Why would you not go to the latest 2.4.0 kernel? They're not perfect,
>but I'm sure they're better than 2.3.51 (which I think is one I couldn't run).

Somewhere in the 2.3.99 to 2.4.0-test series, some nasty memory
management bugs were introduced, and it's not clear that they've all
been worked out yet. They don't bite everyone, so some people are happy
enough with 2.4 kernels - I'm using 2.4.0-test2 without too much
trouble, though earlier kernels had serious performance problems for me.
But, for the people they do bite, I can see why there might be a good
reason to stay back at 2.3.51 for the time being.

Of course, if you're going to run development kernels, you might as well
at least try the latest ones.

-- 
Colin Watson                                     [[EMAIL PROTECTED]]
"A recommended acceptance test might be to experiment with subjects
 whose skulls are only at partial vacuum, such as Vice-Presidents of
 Marketing." - RFC 1437

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

From: [EMAIL PROTECTED] (Colin Watson)
Crossposted-To: linux.redhat.install,alt.os.linux,linux.redhat.misc,linux.redhat.rpm
Subject: Re: Number of groups associated with one user id ...
Date: 26 Sep 2000 16:27:07 GMT

[Follow-ups set away from linux.redhat.rpm; my guess is that that's for
development of the RPM tool. Posting to too many newsgroups is a bad
idea anyway.]

Tajendra Vir Singh <[EMAIL PROTECTED]> wrote:
>we have a system running Redhat 6.2 ersion. There are many around 50 user
>group added to the system. The situation is that one user has to be in all
>user groups in order to do some special task. I saw that one can't have
>more then 32 groups associated with a userid. Wanted to modify it to 64.

Well, you could try changing the NGROUPS #define in
linux/include/asm-i386/param.h [1] [2] and the NGROUPS_MAX #define in
linux/include/linux/limits.h to the figure you want, and recompiling
your kernel.

[1] ... where linux/ is wherever you've got your kernel source tree.
[2] asm-i386 might be asm-sparc, asm-alpha, or whatever your processor
    architecture is.

However, some userspace tools - like your C library, for instance -
might well be compiled against these header files too, and you'll have
to rebuild all of those. Trust me, you really do not want to have to
rebuild your C library if you can possibly help it.

Instead, do it a different way, and forget about the limit on
supplementary groups. Have a look at 'sudo', which can almost certainly
do what you want in a much cleaner and easier way.

-- 
Colin Watson                                     [[EMAIL PROTECTED]]
"F(x) = d(y)/d(x)"  "This isn't the equation we're looking
 for.  Move along." - [EMAIL PROTECTED], rec.arts.sf.written

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

From: Brandon Hoppe <[EMAIL PROTECTED]>
Crossposted-To: 
alt.os.linux,linux.help,alt.os.linux.mandrake,alt.uu.comp.os.linux.questions,comp.os.linux.misc,comp.os.linux.x,linux.redhat.install
Subject: NEED HELP w/ Linux install and X Server setup on Laptop...
Date: Tue, 26 Sep 2000 11:28:03 -0500


I have an HP Pavilion N3250 Notebook.

   AMD K6/2 475MHz
   6.0GB Harddrive 
   12.1 Inch SVGA TFT Display
   Trident Cyberblade I7 AGP 3D Graphics.

I am installing Mandrake 7.1 using lnx4win because I don't want to
remove Windows. I was able to install it and get it running but I wasn't
able to get the X Server running. I found a reference on how an N3270
(which has the same video chipset, just a larger display) was setup at:
http://vorpal.nebrwesleyan.edu/~mcclung/hp3270.html.

I read that I need to use the Frame Buffer as in the way the N3270 uses
it. I tried using the svga X Server but couldn't come up with any
combinations of video/monitor that worked so I then tried the Frame
Buffer. But, I didn't have a XF86_FBDev in /usr/X11R6/bin. I only had
XF86_SVGA and XF86_VGA16. So I download an XF68_FBDev (yes, it says XF68
not XF86, why?? I don't know, does anyone know why?) and tried that. But
I got this error:

        unable to open device /dev/fb0 (no such device)

I also read that XF86_FBDev comes with xfree86 3.3.6, which comes in
Mandrake 7.1 (correct me if I'm wrong). So, does this get installed
normally or do I have to specify it to get installed when I run lnx4win
(drake)?

So, to sum it up, here's my questions/needs:

  What to I need to have installed when I install Linux on a laptop
(specifically XF86_FBDev)?

  How do I setup the X Server? Do I use the Frame Buffer? How do I
install the Frame Buffer?

  What do I need in my XF86Config file? (There's an XF86Config file in
the above web page for use with FBDev)


Any help would be greatly appreciated. Thanx,
Brandon

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

From: Michael Kelly <[EMAIL PROTECTED]>
Subject: Re: Sharing swap space between windows and linux
Date: Tue, 26 Sep 2000 12:45:21 -0400

On Mon, 25 Sep 2000 14:34:48 -0400, "Space Cowboy" <[EMAIL PROTECTED]>
wrote:

>Am I correct in assuming this would make linux run a little slower since
>you're using a FAT32 drive for the swap file rather than a linux native
>partition?
>

Yup.  Going through the file system slows things down.
Also with a swap partition not only do you bypass the
file system but the pages are contiguous.  It doesn't
get fragmented the way a swap file can.

With cheap multigig drives unless you're really on a tight
budget it's probably a better use of your time just to put
in a big drive.


"I wonder if she actually had an orgasm in the
 two years we were married, or did she fake it
 that night?"

                - Woody Allen (Play It Again Sam)

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

From: Juergen Sauer <[EMAIL PROTECTED]>
Subject: Re: boot linux from lan
Date: Mon, 25 Sep 2000 16:58:15 +0000 (UTC)

robert zhang <[EMAIL PROTECTED]> schrieb
am Fri, 22 Sep 2000 08:12:53 -0500 in comp.os.linux.setup:
> Is there a way to boot linux from lan network, or how does the boot from lan
> (in bios setup) work at all?

The Network Card's bios mu�t support this, esp. the nic *has* a bios/rom.

After initialisation of the system the master bios seeks a boot device,
this may over take by a BootProm on the NIC.

After this sage the boot process may start, the NIC rom broadcasts
into the Ethernet an searches vor a aktive bootp server,
        this assigns a IP Address to the Boot-Client,
        and delivers over tftp a "Boot" image = the Kernel file.
Further initialisation is the part of the "initial Ramdisc" contents
and ther existing scripts to boot up the diskless client.

Search the ftp.kernel.org for "boot/etherboot" and simular environments
do setup such beast.

mfG
        Jojo

-- 
- Professionelle Linux Server,   Professioneller Support und Dienstleistungen
- AutomatiX GmbH  - Vollautomatische Kransteuerungen & SAP f�higes Lagerger�t
- J�rgen Sauer Neue Str. 11 28790 Schwanewede        mailto:[EMAIL PROTECTED]
- +49-4209-4699 +49-172-5466499  FAX  +49 4209 4644   http://www.automatix.de
- Hinweis: Nach �28 Abs.3 Bundesdatenschutzgesetz WIDERSPRECHE
- ich der Nutzung meiner Daten fuer Werbezwecke!



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

From: John Janmaat <[EMAIL PROTECTED]>
Subject: Re: Second display
Date: Tue, 26 Sep 2000 14:36:27 -0300

Antoine Martin wrote:
> 
> > I have a thinkpad A20m, and need to connect to an external monitor.
> > Ideally, I would use the Fn-F7 combination to switch to the external
> > monitor. I need some pointers on how to configure XFree86 for two
> > monitors, and switching between them.
> 
> I believe this should be handled by the laptop itself (bios?) and not
> XFree.

The BIOS is involved with the switching, but XFree has to be configured
to supply an appropriate signal that the external monitor can
understand.  Since the panel display has a sync rates that are likely to
be different from the external CRT, I need some way to switch these
things when I switch to the second display.  

John.
-- 
_________________________________________________________________
John Janmaat,  Department of Economics,  Acadia University,
Wolfville,  Nova Scotia,  Canada.  B0P-1X0

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

From: [EMAIL PROTECTED] (Kingsley )
Crossposted-To: alt.linux,alt.os.linux
Subject: Re: upgrade mandrake 6.1 to 7.1
Reply-To: [EMAIL PROTECTED]
Date: Tue, 26 Sep 2000 17:55:36 GMT

On Tue, 22 Aug 2000 21:19:13 -0400, "Dale" <[EMAIL PROTECTED]> wrote:

>You should always back up files whenever doing an upgrade.  The 7.1 disc
>offers you the option to upgrade as opposed to doing a new install.  If you
>don't have medium on which do back up your files, you should store them in
>your /home paritition.  This way, you can do an upgrade OR a reinstall
>without losing any info.  If you don't have a /home parition set up, I would
>suggest that you do a clean install and create one.  I use have used my
>/home and have never lost any info whether doing an upgrade or a clean
>install.

What sort of files do you mean when you say backup? And how do you
backup the files? 


Kingsley 
Linux Newbie Since 17 September 2000

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

From: eric <[EMAIL PROTECTED]>
Subject: Re: Linux/W98 to Linux/W2000
Date: Tue, 26 Sep 2000 14:03:10 -0400


 [EMAIL PROTECTED] wrote:
 > Has anyone upgraded a dual boot WIn98/Linux system
 > to Win200/Linux system.
I now am the proud owner of a triple boot Nt4/win2k/linux system.
The trick is to have 3 hard drives.  It works very well.  I have linux
on
hda, nt4 on hdb and win2k on sda.  Lilo boots the computer, win2k and
nt4 don't really know linux is there because lilo remaps the drives in
bios.  This makes setup of msft o/s very easy because you can set up the
drive as the boot drive, setup the msftos and then reconfigure.  To boot
you reconfigure using lilo.  
eric

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

From: Emiliano <[EMAIL PROTECTED]>
Subject: Re: expired root password
Date: Tue, 26 Sep 2000 20:07:05 +0200

SM McAllister wrote:

> I have Caldera Open Linux 2.2. The root password has expired. I have
> tried starting linux in single user mode ( lilo  - linux single ). I
> cannot change the root password. I get this error when typing
> the "passwd root" command :
>
> PAM_pwdb[260]: user (root/0) update failed; pwdb: another process has
> locked resource
>
> This system is not critical, but I have invested alot of time
> configuring it over the last year. I prefer not to lose that investment.
>
> SM McAllister
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.

I think you must start linux in single mode. Now you must erase the
password field in the file /etc/shadow

the password is crypted and is between the first and the second ':' .

Now restart linux and reinsert the password, linux ask you the confirm and
you reinsert the password.

:-)   Hi ! Emiliano


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


** 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.setup) 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-Setup Digest
******************************

Reply via email to