RE: NAV detected a virus in a document you authored

[Michael Olds]

Several hundred members? Is that all? I feel much better about the list now.
On the other hand only several hundred producing this much traffic?

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Jamin W.Collins [mailto:[EMAIL PROTECTED]]
Sent: Saturday, October 05, 2002 4:27 PM
To: [EMAIL PROTECTED]
Subject: Re: NAV detected a virus in a document you authored


You might want to consider not sending messages like this out, especially
for viruses like the one found.  The virus your application indicates it
has found is known to spoof the sending address.  So, you can be
relatively certain if you got it, it didn't come from the address
indicated as the sender.  Thus, your notification is actually causing more
harm than good.  In this particular case, your notification was sent to a
mailing list with several hundred subscribers.  For more information about
the virus listed below, see the following site:

http://www.commandsoftware.com/virus/klez.html

On Sat, 5 Oct 2002 19:06:17 -0400 [EMAIL PROTECTED] wrote:

> Please contact your system administrator
>
>
> The scanned document was QUARANTINED.
>
>
> Virus Information:
> The attachment rock.exe contained the virus W32.Klez.H@mm and could NOT
> be repaired.
>
>
>
> --
> To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> with a subject of "unsubscribe". Trouble? Contact
> [EMAIL PROTECTED]
>


--
Jamin W. Collins


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Uninstalling Exim uninstalls Apache

[Michael Olds]

Hello,

Before I break down and ask for help again setting up email which I have
thoroughly messed up once more...can someone explain to me why when I
uninstall exim apache also gets uninstalled? I am using Synaptic as my front
end for the package manager, and it is not too clear about what it intends
to install or delete before it gets done (still I find it easier to use than
dselect or aptitude).

Thanks,

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Uninstalling Exim uninstalls Apache

[Michael Olds]

Nevermind the direct answer to this, exim and apache share a couple of other
packages that removing exim tries to remove. The real question is, why am I
not being allowed to remove exim by itself? When I tell it to not remove the
two packages needed for apache, it won't remove exim. Why should it care if
exim is removed in broken condition?

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Michael Olds [mailto:[EMAIL PROTECTED]]
Sent: Monday, October 07, 2002 3:17 PM
To: Debian-User@Lists. Debian. Org
Subject: Uninstalling Exim uninstalls Apache


Hello,

Before I break down and ask for help again setting up email which I have
thoroughly messed up once more...can someone explain to me why when I
uninstall exim apache also gets uninstalled? I am using Synaptic as my front
end for the package manager, and it is not too clear about what it intends
to install or delete before it gets done (still I find it easier to use than
dselect or aptitude).

Thanks,

Best Wishes!
Mike Olds www.buddhadust.org


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Uninstalling Exim uninstalls Apache

[Michael Olds]

Thanks Colin and Corrin,

I got as far as seeing the interdependancies. Removing Exim removes two
packages on which Apache depends. But what was bothering me was why I was
unable to Remove Exim without removing those two other packages...as though
I was not able to UNINSTALL a broken program, not that I would be leaving
Apache broken. It was not clear that Apache also depended on exim proper.
The problem relates to my ages long saga trying to set up email. I think I
had better start a new thread with a more directly related header for
that...

Best Wishes!
Mike Olds www.buddhadust.org





-Original Message-
From: Colin Watson [mailto:[EMAIL PROTECTED]]On Behalf Of Colin
Watson
Sent: Monday, October 07, 2002 3:51 PM
To: [EMAIL PROTECTED]
Subject: Re: Uninstalling Exim uninstalls Apache


On Mon, Oct 07, 2002 at 03:17:09PM -0700, Michael Olds wrote:
> Before I break down and ask for help again setting up email which I have
> thoroughly messed up once more...can someone explain to me why when I
> uninstall exim apache also gets uninstalled?

apache depends on logrotate which depends on cron which recommends a
mail-transport-agent (of which exim is one). If you're trying to remove
exim in order to replace it with a different MTA, just install the other
MTA and exim will automatically be removed. If not, well, honestly you'd
be better off leaving some kind of MTA there, even if it's only there to
forward mail on somewhere else. Maybe if you tell us a bit more about
your requirements someone can help.

Cheers,

--
Colin Watson  [[EMAIL PROTECTED]]


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Uninstalling Exim uninstalls Apache

[Michael Olds]

OK, here is the problem, now solved.

I had managed to get Exim up and running, but was only able to get e-mail
delivered to root when root set up an account with my username. No matter
what I did with permissions nothing would change it. I uninstalled Exim and
Kmail alternatively and together and altogether (went back to Sendmail, and
tried Postfix), etc. nothing would deliver mail to the user. In addition to
noticing that uninstalling the MTA uninstalled Apache which confounded me, I
also noticed that when I uninstalled and re-installed Kmail no matter how
thoroughly I thought I had taken out all the left-behind Kmail files, up it
would pop with all my settings intact. (I thought these settings were the
real source of the problem. Ok now that I'm at the end of the story, I will
make it short: what appears to have worked is uninstalling and reinstalling
Procmail.

Thanks for the help folks! NOW I'm on to POP3 setup

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: how to install just one package from some other distro (likewoody)?

[Michael Olds]

This is great, except that someone asking this question is obviously a very
beginner, and will download the package and then, using a terminal, type in
"dpkg -i package.deb" or even, maybe "dpkg -i
name_of_downloaded_package_etc.deb" which will not help them.

1. Download the package
2. In a terminal: type: "cd /directory where downloaded/filename.deb"
3. Then type: "dpkg -i nameofpackage.deb"

4. When it tells you there are errors and it cannot be installed because of
dependancies which are not met: go back to the download source and research
the various other packages that are needed in order to make this package
work. Try Sourceforge for a good resource in this connection.
5. Download all the packages on which your desired package is dependant...to
the same directory as your package.
6. In a terminal: type: "cd /directory where downloaded, etc"
7. "dpkg -i name_of_first.deb name_of_second.deb etc.deb
8. cross fingers

-Original Message-
From: Cristi Banciu [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, October 09, 2002 4:53 AM
To: [EMAIL PROTECTED]
Cc: Debian User List
Subject: Re: how to install just one package from some other distro
(likewoody)?


On Wed, 2002-10-09 at 14:08, J.S.Sahambi wrote:
> Hi
>   I wanted to install avidemux but my "apt-cache search avidemux" did
> not give me anything. My sourses.lst file is:
>
> deb http://202.141.80.72/debian testing main non-free contrib
> deb http://202.141.80.72/debian-non-US testing/non-US main non-free
contrib
download package
dpkg -i package.deb


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

inode missing

[Michael Olds]

I always suspected as much...but what does this mean?

I have downloaded and installed the partimage package, and this is the first
screen to greet me: Missing Inode...shall it create?

I am running Debian Woody with Ext3 fs. This is the downloaded package not
the one that comes with stable.

I have done a Google search and looked through my books and I have a vague
idea what an inode relates to, but would like to know precisely...and also
is it safe to let Partition Image go ahead and create these?

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Mount or disable devfs, was: inode missing

[Michael Olds]

Mount or disable devfs?

Device File System

For a good discussion of devfs see:
http://www.atnf.csiro.au/people/rgooch/linux/docs/devfs.html#what-is-it

In attempting to set up and run Partition Image, I got the message:
"Inode Missing" /dev/ide/host0/bus0/target0/lun0/disc with an offer to
create it. I contacted the PI board and asked: "Can someone explain what
this is and what it means and whether or not it is safe to allow PI to
create the inodes?" and was told:

"In your case not (safe)."

"ext2/3 inode is almost like clusters in FAT file system."

"I think you enabled devfs support in kernel but you haven't mounted
devfs partition into /dev. Partimage can't work in such situation.
You'll have to remove devfs support from kernel or mount devfs in /dev."

"Cat your /proc/partitions file. If you read
/dev/ide/host0/bus0/target0/lun0/partXXX you enabled devfs support."

I have apparently enabled devfs support as I have a fair number of these
listed, apparently one for each partition and extended partition.

"Now if you ls /dev and you only see hdaXX without ide subdirectory, you
haven't mounted devfs in /dev."

And I have no subdirectory /dev/ide

So the questions to the Debian folk are:

[As far as I can tell, this devfs thing is primarily of use to those with
external devices such as scanners and so forth. I have only one external
device connected to this computer: a printer.]

What are the advantages/disadvantages in my situation?

1. Should I mount devfs?
2. If so: how do I go about mounting it? I have read some of the
discussions, but they are generally on a higher level than my own. Is it the
same as mounting any other device? Do I need to mount each (one for each
partition) individually, or will mount /devfs mount what is need? If this is
an advisable setup, can it be set up to happen automatically on boot?
3. If not: how do I go about removing devfs support? Is this going to
require a Kernel Rebuild?
4. How did I get into this mess in the first place? i.e. What did I say yes
to that I didn't need?

A lot of questions; I am still reading, but I could use any thoughts anyone
here has on the matter.

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Kernel Panic was: System is too Big; son of make menuconfig

[Michael Olds]

Thanks Russell and Torsten,
Last night I did a re-install...not a huge loss in that this is still in the
learning/setup phase...and all of this was in the service of setting up some
kind of backup routine.

Suggestion for the KDE folk, by the way: 1. Give us a way to use just the
desktop without installing all the related programs...kate, konquoror,
kword, konsole...in fact the whole Linux setup should be different as far as
these desktop things are concerned...we should be being given a choice of
desktop setups, then for the one we choose, a second set of choices as to
what to include. I don't mean that the options are there, scattered around
throughout the available packages, but that we should start with a dialog
about what to include. 2. In KDE you can save a scheme, but it isn't saved
for everyone and it should be.

This is the thing that had me perplexed: while I did not use the original
configuration as the basis for the kernel rebuild, I did follow it...if
something was built in, I built it in, if something was a module, I made it
a module...so it seems it must have been the --initrd option.

Best Wishes!
Mike Olds www.buddhadust.org



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Kernel Panic~: Rebuild Kernel: miniHOWTO draft

[Michael Olds]

Russell, Torsten, et al:

This is the "method" I have written up in my notes for a Debian Style Kernel
Rebuild. My feeling after all this is that this is something that as hair
raising as it might seem for beginners, should be one of the first things we
learn, not to be delayed.

It would be good if a step by step like this were made for recovering using
the suggested recovery disk (including link to source)

RE: Kernel Panic~: Rebuild Kernel: miniHOWTO draft .001

[Michael Olds]

Russell, et al

Still a draft! Please note questions and also corrected typo in 2 "libc6"



[ ]1. Download the latest kernel to /usr/src/linux/kernel/ (if no
/usr/src/linux/kernel dir exists, make it. Question: the kernel source
package downloaded directly to /usr/src for me...is it safe to move the
source tar.bz2 to /usr/src/linux (for order under src)? before unpacking it)
and unpack it.
[ ]2. Make sure
   [ ]"bin86"
   [ ]"libc6-dev"
   [ ]"debianutils"
   [ ]"make"
   [ ]"bzip2" and
   [ ]"kernel-package" are installed.
[ ]3. (on 2nd and subsequent rebuilds) Copy configuration file to
/usr/src/linux/kernel_version_source_directory/
[ ]4. # cd to /usr/src/linux/kernel_version_source_directory
   # make menuconfig
and make your selections
[ ]5. # make-kpkg clean
   # make-kpkg -revision= From: Michael Olds [mailto:[EMAIL PROTECTED]]
> Sent: Sunday, October 13, 2002 6:53 AM
> To: Russell; [EMAIL PROTECTED]
> Cc: [EMAIL PROTECTED]
> Subject: RE: Kernel Panic was: System is too Big; son of make menuconfig
>
> Thanks Russell and Torsten,
> Last night I did a re-install...not a huge loss in that this is still in
the
> learning/setup phase...and all of this was in the service of setting up
some
> kind of backup routine.
>
> Suggestion for the KDE folk, by the way: 1. Give us a way to use just the
> desktop without installing all the related programs...kate, konquoror,
> kword, konsole...in fact the whole Linux setup should be different as far
as
> these desktop things are concerned...we should be being given a choice of
> desktop setups, then for the one we choose, a second set of choices as to
> what to include. I don't mean that the options are there, scattered around
> throughout the available packages, but that we should start with a dialog
> about what to include. 2. In KDE you can save a scheme, but it isn't saved
> for everyone and it should be.
>
> This is the thing that had me perplexed: while I did not use the original
> configuration as the basis for the kernel rebuild, I did follow it...if
> something was built in, I built it in, if something was a module, I made
it
> a module...so it seems it must have been the --initrd option.

read:
  /usr/src/linux/Documentation/initrd.txt
  /usr/src/linux/Documentation/modules.txt
  /usr/src/linux/README
  /usr/src/linux/README.Debian

http://qref.sourceforge.net/Debian/reference/ch-kernel.en.html

http://infocom.cqu.edu.au/Units/aut99/85321/Resources/Print_Resources/Textbo
ok/chap13/

http://www.google.com/search?q=%22kernel+headers%22+linus+quote&hl=en&lr=&ie
=UTF-8&start=10&sa=N

http://www.linuxfromscratch.org/faq/#AEN334
http://www.uwsg.iu.edu/hypermail/linux/kernel/0007.3/0587.html

Bear in mind that debian does things a bit differently than other
distros. I haven't done enough programming to figure out the
specifics of kernel headers with debian yet. The main thing is
that a new kernel can be used, but source being built only needs
the old headers that the system libraries like glibc? were
built with.


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Kernel Panic~: Rebuild Kernel: miniHOWTO draft .002

[Michael Olds]

Still a draft: Note: need sources for the Linux Documentation listed.
Clarification needed: one source (sourceforge) does not mention a
/usr/src/linux directory in it's method, the other source (linuxorbit)
suggests using a /usr/src/linux directory and creating a symbolic link from
/usr/src/kernel-source... to /usr/src/linux is this important, which is
correct, are these two separate techniques, can one blend the techniques as
I suggest here?



[ ]A. Before 1: Download and create a recovery floppy from:
http://www.toms.net/rb/

[ ]1.a. Download the latest kernel source package (use a Debian source) to
/usr/src (this is where it will be installed if you use dselect, a package
manager, synaptic or aptitude)
[ ]1.b. if no /usr/src/linux dir exists, make it
  $ cd /usr/src
  $ mkdir linux
[ ]1.c. change directory to /usr/src and unpack the source package -- use
the unpacking tools with your file manager, or:
  $ su #enter root password
  # cd /usr/src
  # tar --bzip2 -xvf kernel-source-2.4.18.tar.bz2 (or latest)
[ ]1.d create a symbolic link to /usr/src/linux
  # ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux

[ ]2. Just to be professional: go to "/etc/kernel.pkg.conf" and modify the
file by inserting your "Kernel Source Maintainer Name" and e-mail address.

[ ]3. Make sure
   [ ]"expectk" if you want to use the user friendly: make xconfig
   [ ]"debhelper"
   [ ]"modutils"
   [ ]"libncurses5-dev"
   [ ]"bin86"
   [ ]"libc6-dev"
   [ ]"debianutils"
   [ ]"make"
   [ ]"bzip2" and
   [ ]"kernel-package" are installed.

[ ]4. Remove PCMCIA support if not needed

$ cd /usr/src/kernel-source-2.4.18 # if this is your kernel source
 $ rm -rf */pcmcia
 # [OPTIONAL] if one wants to use modules from pcmcia-cs or no
pcmcia (see NOTES below)

[ ]5. (on 2nd and subsequent rebuilds) Copy old configuration file to
/usr/src/kernel-source-2.4.18
  $ cd /usr/src/kernel-source-2.4.18
  $ cp /boot/config-2.4.18-386 .config # get current config as
default

[ ]6. # cd to /usr/src/linux

  $ make menuconfig or make xconfig

and make your selections

   To do this you need to have some familiarity with your hardware and what
you are intending to use the system for. This you need to work out yourself.
Once this is figured out, it might be helpful, on the first rebuild, to make
your updates with reference to the original configuration, (called
config-2.4.18bf24 (or your kernel version)) which you should find under
/root using the "M" (module) option where a module was used and building the
feature into the kernel (*) or [y] option where it was built into the
kernel. Use the HELP option for brief explanations. This at least until you
understand which can be safely changed under your setup.

[ ]5.a.  [ ]# make-kpkg clean
 [ ]# make-kpkg -revision=

NOTES: make-kpkg kernel_image actually does make oldconfig and make dep. Do
not use --initrd if initrd is not used. The current Debian initrd boot
script requires a cramfs kernel patch, if you obtain kernel source from
non-Debian archives; see http://bugs.debian.org/149236.

One can avoid rm -rf */pcmcia by selecting "General setup -->" to
"PCMCIA/CardBus support -->" in make menuconfig and setting the
configuration to "< > PCMCIA/CardBus support" (i.e., uncheck the box).

On an SMP machine, set CONCURRENCY_LEVEL according to kernel-pkg.conf(5).

Next mini: how to recover when nothing said here works.

RE: Kernel Panic~: Rebuild Kernel: miniHOWTO draft .001

[Michael Olds]

Russel...noted and I will do a revision shortly...however I got as far as
making the package with the --initrd option and managed to control my
fingers long enough to read the caution about editing lilo at the end...so I
have not installed my package. I am using ext3, and I am not sure if it was
you, but someone this morning said I needed to go initrd. But I see my
current setup using the bf24 kernel does not use initrd (or, at least it is
not mentioned in lilo.conf and I can't find the image in /boot)(but the new
kernel is not bf24 it is 2.4.18) But I do remember using it in the setup
before the most recent install (very confused at this point). I am reading
round that on a one-box system it shouldn't be needed...I do have SCSI
support enabled (for scsi emulation) as a module. So the question is: do I
need it, and if so, how do I go about getting it in there?

Best Wishes!
Mike Olds www.buddhadust.org



-Original Message-
From: Russell [mailto:[EMAIL PROTECTED]]
Sent: Sunday, October 13, 2002 4:38 PM
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: Kernel Panic~: Rebuild Kernel: miniHOWTO draft .001


Michael Olds wrote:
>
> Russell, et al
>
> Still a draft! Please note questions and also corrected typo in 2 "libc6"
>
> 
>
> [ ]1. Download the latest kernel to /usr/src/linux/kernel/ (if no
> /usr/src/linux/kernel dir exists, make it.

In the links above, Linus says it's bad to do builds in /usr/src,
whereas i think the debian manual says to do it that way. An
explanation for this would avoid confusion.

> Question: the kernel source
> package downloaded directly to /usr/src for me...is it safe to move the
> source tar.bz2 to /usr/src/linux (for order under src)? before unpacking
it)
> and unpack it.

I've been unpacking the source into /usr/src/kernel-source-2.4.18 and
symlinking /usr/src/linux to /usr/src/kernel-source-2.4.18. That way,
you can have multiple source trees in /usr/src and just change the
/usr/src/linux symlink to the new tree.

> [ ]2. Make sure
>[ ]"bin86"
>[ ]"libc6-dev"
>[ ]"debianutils"
>[ ]"make"
>[ ]"bzip2" and
>[ ]"kernel-package" are installed.

IIRC, don't use gcc 3.x, but something like gcc 2.95.

> [ ]3. (on 2nd and subsequent rebuilds) Copy configuration file to
> /usr/src/linux/kernel_version_source_directory/

Don't think that's needed.

> [ ]4. # cd to /usr/src/linux/kernel_version_source_directory
># make menuconfig
> and make your selections

If you're using --initrd, make sure "Compressed ROM file system
support" is built in, or you'll get a kernel panic (need CramFS).

> [ ]5. # make-kpkg clean
># make-kpkg -revision=

RE: Kernel Panic~: Rebuild Kernel: miniHOWTO draft .003

[Michael Olds]

Russell, et al,

Third draft. A couple of questions asked below still remain:





[ ]A. Before 1: Download and create a recovery floppy from:
http://www.toms.net/rb/

[ ]1.a. Download the latest kernel source package (use a Debian source) to
/usr/src (this is where it will be installed if you use dselect, a package
manager, synaptic or aptitude). Best practices seem to indicate that the
package be built outside of /usr/src so creating a sym-link from /linux to
/usr/src/kernel~ as in step 1.d below seems to be the way to handle this.
This way, you can have multiple source trees in /usr/src and just change the
/usr/src/linux symlink to the new tree.)
[ ]1.b. if no /usr/src/linux dir exists, make it
  $ cd /usr/src
  $ mkdir linux
[ ]1.c. change directory to /usr/src and unpack the source package -- use
the unpacking tools with your file manager, or:
  $ su #enter root password
  # cd /usr/src
  # tar --bzip2 -xvf kernel-source-2.4.18.tar.bz2 (or latest)
[ ]1.d create a symbolic link to /usr/src/linux
  # ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux

[ ]2. Just to be professional: go to "/etc/kernel.pkg.conf" and modify the
file by inserting your "Kernel Source Maintainer Name" and e-mail address.

[ ]3. Make sure
   [ ]"expectk" if you want to use the user friendly: make xconfig
   [ ]"debhelper"
   [ ]"modutils"
   [ ]"libncurses5-dev"
   [ ]"bin86"
   [ ]"libc6-dev"
   [ ]"debianutils"
   [ ]"make"
   [ ]"bzip2" and
   [ ]"kernel-package" are installed.

   R Notes: don't use gcc 3.x, but something like gcc 2.95.
 Which I do not understand. Clarification?
==
REMOVING THIS STEP: This step breaks make xconfig for people who want to go
that way. The Sourceforge method states: One can avoid rm -rf */pcmcia by
selecting "General setup -->" to "PCMCIA/CardBus support -->" in make
menuconfig and setting the configuration to "< > PCMCIA/CardBus support"
(i.e., uncheck the box). Which I will keep in the NOTES section

[ ]4. Remove PCMCIA support if not needed

$ cd /usr/src/kernel-source-2.4.18 # if this is your kernel source
 $ rm -rf */pcmcia
 # [OPTIONAL] if one wants to use modules from pcmcia-cs or no
pcmcia (see NOTES below)



[ ]5. (on 2nd and subsequent rebuilds) Copy old configuration file to
/usr/src/kernel-source-2.4.18
  $ cd /usr/src/kernel-source-2.4.18
  $ cp /boot/config-2.4.18-386 .config # get current config as
default

R: notes (and the Sorceforge docs confirm) that this is not necessary as the
script will pick up the oldconfig)...eliminating this step.
==

[ ]6. # cd to /usr/src/linux

  $ make menuconfig or make xconfig

and make your selections

   To do this you need to have some familiarity with your hardware and what
you are intending to use the system for. This you need to work out yourself.
Once this is figured out, it might be helpful, on the first rebuild, to make
your updates with reference to the original configuration, (called
config-2.4.18bf24 (or your kernel version)) which you should find under
/root using the "M" (module) option where a module was used and building the
feature into the kernel (*) or [y] option where it was built into the
kernel. Use the HELP option for brief explanations. This at least until you
understand which can be safely changed under your setup.

  Note: If you're using --initrd, make sure "Compressed ROM file system
support" is built in, or you'll get a kernel panic (need CramFS).

  (Side note: I see that answers my question about how to get it in
there...sorry!)


[ ]5.a.  [ ]# make-kpkg clean
 [ ]# make-kpkg -revision=

NOTES: make-kpkg kernel_image actually does make oldconfig and make dep. Do
not use --initrd if initrd is not used. The current Debian initrd boot
script requires a cramfs kernel patch, if you obtain kernel source from
non-Debian archives; see http://bugs.debian.org/149236.

On an SMP machine, set CONCURRENCY_LEVEL according to kernel-pkg.conf(5).

Next mini: how to recover when nothing said here works.

RE: Kernel Panic~: Rebuild Kernel: miniHOWTO draft .001

[Michael Olds]

Russell,

Ok...I am going to remake my package...without initrd as I thought it not
such a good idea to have my main file system as a module and so included it
in the kernel.

You say: Backup your HDD. Easy for you to say: I come from Windows where I
press two buttons on Drive Image to backup my whole system. Here I have been
working for a month just to try and set up so I can use partimage!

Best Wishes!
Mike Olds www.buddhadust.org



-Original Message-
From: Russell [mailto:[EMAIL PROTECTED]]
Sent: Sunday, October 13, 2002 6:12 PM
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: Kernel Panic~: Rebuild Kernel: miniHOWTO draft .001


Michael Olds wrote:
>
> Russel...noted and I will do a revision shortly...however I got as far as
> making the package with the --initrd option and managed to control my
> fingers long enough to read the caution about editing lilo at the end...so
I
> have not installed my package. I am using ext3, and I am not sure if it
was
> you, but someone this morning said I needed to go initrd. But I see my
> current setup using the bf24 kernel does not use initrd (or, at least it
is
> not mentioned in lilo.conf and I can't find the image in /boot)(but the
new
> kernel is not bf24 it is 2.4.18) But I do remember using it in the setup
> before the most recent install (very confused at this point). I am reading
> round that on a one-box system it shouldn't be needed...I do have SCSI
> support enabled (for scsi emulation) as a module. So the question is: do I
> need it, and if so, how do I go about getting it in there?

read: http://qref.sourceforge.net/Debian/reference/ch-kernel.en.html
it says: mv /usr/src/linux/arch/i386/boot/bzImage to /boot/
I have /usr/src/linux as a symlink to /usr/src/kernel-source-2.4.18

I'm using ext3, and you don't need initrd. IIRC, you only need initrd
if your filesystem is a module and not compiled into the kernel, so
at least CramFS needs to be built in if nothing else is. Backup your
HDD. If you get a kernel panic because of lilo, you can usually still
boot in using the downloaded debian rescue disk.


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE:backup script was Kernel Panic~

[Michael Olds]

Russell, Thanks, I've been trying to get a grasp of this, I'm sure this will
help!

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Russell [mailto:[EMAIL PROTECTED]]
Sent: Sunday, October 13, 2002 6:55 PM
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: Kernel Panic~: Rebuild Kernel: miniHOWTO draft .001


Michael Olds wrote:
>
> Russell,
>
> Ok...I am going to remake my package...without initrd as I thought it not
> such a good idea to have my main file system as a module and so included
it
> in the kernel.
>
> You say: Backup your HDD. Easy for you to say: I come from Windows where I
> press two buttons on Drive Image to backup my whole system. Here I have
been
> working for a month just to try and set up so I can use partimage!

I learnt some bash scripting and made this to backup onto a removeable
hard-disk (had some help from a previous poster on this group). Don't
use it until you understand how it works and what it assumes. I use
it from /root, and just type ./backup

#!/bin/sh

# Copy everything (almost) from the system to a backup removeable hard-disk
# Assumptions:
#  The spare hard-disk has enough capacity
#  The spare disk is suitably partitioned and set up.
#
# The spare disk won't be bootable unless lilo is run separately,
#  which i haven't got around to doing yet

# The backup /boot partition:
HDD_BOOT=hdc1

# The backup root / partition:
HDD_MAIN=hdc3

# Mount /boot partition
if [ ! -d /mnt/bootpart ]
then
  mkdir /mnt/bootpart
fi

if ! grep -qs /dev/$HDD_BOOT /proc/mounts
then
  if ! mount -t ext2 /dev/$HDD_BOOT /mnt/bootpart
  then
echo "Removeable /dev/$HDD_BOOT not working"
exit 1
  fi
fi

# Mount main partition
if [ ! -d /mnt/mainpart ]
then
  mkdir /mnt/mainpart
fi

if ! grep -qs /dev/$HDD_MAIN /proc/mounts
then
  if ! mount -t ext2 /dev/$HDD_MAIN /mnt/mainpart
  then
echo "Removeable $HDD_MAIN not working"
exit 1
  fi
fi

# Copy everything on main partition
# Strip some top level directories we don't want to copy
FILES=$( chroot / ls )

for i in $FILES
do
  if [ -d /$i ]
  then
echo "Updating /$i"
  else
echo "Updating $i"
  fi

  case "$i" in
  boot)
if ! rsync -a --delete /boot /mnt/bootpart
then
  echo "Failed to update /boot"
  exit 1
fi;;
  cdrom)
echo /$i | cpio -pdm /mnt/mainpart
continue;;
  proc)
echo /$i | cpio -pdm /mnt/mainpart
continue;;
  floppy)
echo /$i | cpio -pdm /mnt/mainpart
continue;;
  mnt)
echo /$i | cpio -pdm /mnt/mainpart
continue;;
  esac

  if ! rsync -a --delete "/$i" /mnt/mainpart
  then
echo "Failed to update /$i"
exit 1
  fi
done

echo "System backup successful"
exit 0


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Rebuilding the Kernel Mini HOW TO

[Michael Olds]

Hello,

This is still a draft. This is a request for advanced users to take a look
see and add comments or suggestions. The idea is a step by step instruction
for Fresh New to Linux/Debian users, with the hope that they will be
encouraged to attempt a rebuild of their kernel almost first thing after
initial install. I use a technique which is a blend of using a file manager
along with a terminal so I can see where I am at and yet do what I need to
do in the terminal. This is a good way to work into using the terminal.



[ ]A. Before 1: Download and create a recovery floppy from:
http://www.toms.net/rb/

[ ]1.a. Download the latest kernel source package (use a Debian source; or
make sure you have read all the documentation above) to /usr/src (this is
where it will be installed if you use dselect, a package manager, synaptic
or aptitude). Best practices seem to indicate that the package be built
outside of /usr/src so creating /usr/src/linux and creating a sym-link from
/linux to /usr/src/kernel~ as in step 1.d below seems to be the way to
handle this.)

[ ]1.b. if no /usr/src/linux dir exists, make it
  $ cd /usr/src
  $ mkdir linux
[ ]1.c. change directory to /usr/src and unpack the source package -- use
the unpacking tools with your file manager, or:
  $ su #enter root password
  $ cd /usr/src
  $ tar --bzip2 -xvf kernel-source-2.4.18.tar.bz2 (or latest)
[ ]1.d create a symbolic link to /usr/src/linux
  $ ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux

[ ]2. Just to be professional: go to "/etc/kernel.pkg.conf" and modify the
file by inserting your "Kernel Source Maintainer Name" and e-mail address.

[ ]3. Make sure
   [ ]"expectk" if you want to use the user friendly: make xconfig
   [ ]"debhelper"
   [ ]"modutils"
   [ ]"libncurses5-dev"
   [ ]"bin86"
   [ ]"libc6-dev"
   [ ]"debianutils"
   [ ]"make"
   [ ]"bzip2" and
   [ ]"kernel-package" are installed.

   R Notes: don't use gcc 3.x, but something like gcc 2.95. Which I do not
understand. Clarification?

[ ]4. (on 2nd and subsequent rebuilds) Copy old configuration file to
/usr/src/kernel-source-2.4.18
  $ cd /usr/src/kernel-source-2.4.18
  $ cp /boot/config-2.4.18-386 .config # get current config as
default

[ ]5. # cd to /usr/src/linux

  $ make menuconfig or make xconfig

and make your selections

   To do this you need to have some familiarity with your hardware and what
you are intending to use the system for. This you need to work out yourself.
Once this is figured out, it might be helpful, on the first rebuild, to make
your updates with reference to the original configuration, (called
config-2.4.18bf24 (or your kernel version)) which you should find under
/root using the "M" (module) option where a module was used and building the
feature into the kernel (*) or [y] option where it was built into the
kernel. Use the HELP option for brief explanations. This at least until you
understand which can be safely changed under your setup.

  Note: If you're using --initrd, make sure "Compressed ROM file system
support" is built in, or you'll get a kernel panic (need CramFS).

  (Side note: I see that answers my question about how to get it in
there...sorry!)


[ ]6.a.  [ ]# make-kpkg clean
   6.b.  [ ]# make-kpkg -revision=

NOTES: make-kpkg kernel_image actually does make oldconfig and make dep. Do
not use --initrd if initrd is not used. The current Debian initrd boot
script requires a cramfs kernel patch, if you obtain kernel source from
non-Debian archives; see http://bugs.debian.org/149236. As I understand it,
initrd is used in situations where many computers use a highly modularized
kernel, or where an individual has his file system loaded as a module. This
requires the use of a virtual RAM disk in order to run smoothly, and that is
what initrd enables (Experts are requested to make this explanation more
lucid.)

On an SMP machine, set CONCURRENCY_LEVEL according to kernel-pkg.conf(5).

Next mini: how to recover when nothing said here works.

RE: Rebuilding the Kernel Mini HOW TO

[Michael Olds]

Thank you,

I am still in a fog in terms of simple things like symlinks. What you want
is to end up with a link in the linux directory called kernel-source-2.4.18
that links to /usr/src/kernel-source-2.4.18. So how should that be said?

[ ]1.d create a symbolic link to kernel-source-2.4.18 from /usr/src/linux
$ ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux

Best Wishes!
Mike Olds www.buddhadust.org


-Original Message-
From: Jamin W.Collins [mailto:[EMAIL PROTECTED]]
Sent: Monday, October 14, 2002 2:47 PM
To: [EMAIL PROTECTED]
Subject: Re: Rebuilding the Kernel Mini HOW TO


On Mon, 14 Oct 2002 14:19:48 -0700 Michael Olds <[EMAIL PROTECTED]>
wrote:

> [ ]1.b. if no /usr/src/linux dir exists, make it
>   $ cd /usr/src
>   $ mkdir linux
(snip)
> [ ]1.d create a symbolic link to /usr/src/linux
>   $ ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux

Doing 1.b. followed by 1.d. will _not_ work.  A directory with the name
"linux" exists, as it was created in 1.b. and thus a symlink with the same
name can not be made at the same level.  Instead you will wind up with a
symlink in the linux directory called kernel-source-2.4.18.

--
Jamin W. Collins


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Rebuilding the Kernel Mini HOW TO

[Michael Olds]

We're getting there.

Yes, I have it incorrectly, what is wanted is that step five should read:

[ ]5. # cd to /usr/src/linux/kernel-source-2.4.18

Corrected, thanks.

Best Wishes!
Mike Olds www.buddhadust.org



-Original Message-
From: Jamin W.Collins [mailto:[EMAIL PROTECTED]]
Sent: Monday, October 14, 2002 4:09 PM
To: [EMAIL PROTECTED]
Subject: Re: Rebuilding the Kernel Mini HOW TO


On Mon, 14 Oct 2002 15:25:07 -0700 Michael Olds <[EMAIL PROTECTED]>
wrote:

> I am still in a fog in terms of simple things like symlinks. What you
> want is to end up with a link in the linux directory called
> kernel-source-2.4.18 that links to /usr/src/kernel-source-2.4.18. So how
> should that be said?

Why would you want that?  To make sure we are talking about the same
thing, you're saying that you want to end up with
   /usr/src/linux/kernel-source-2.4.18
that points to
   /usr/src/kernel-source-2.4.18

If this is correct, then your instructions in step 5 are wrong.

|[ ]5. # cd to /usr/src/linux
|
|  $ make menuconfig or make xconfig

The "make menuconfig" needs to be done at the top of the kernel source
tree.  Which, would now be /usr/src/linux/kernel-source-2.4.18.  Perhaps
what you intended is that
   /usr/src/linux
point to
   /usr/src/kernel-source-2.4.18

If that's the case, drop 1.b. altogether.

--
Jamin W. Collins


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Rebuilding the Kernel Mini HOW TO

[Michael Olds]

Thanks Joyce,

Added a line: Why would I want to rebuild the kernel, and will just add
quotes from anyone who responds here. For me "My kernel went from an
original 1,263,333KB to 861,587KB," with the exact same setup.

I also included your link.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Joyce, Matthew [mailto:[EMAIL PROTECTED]]
Sent: Monday, October 14, 2002 4:41 PM
To: 'Michael Olds'; Debian-User@Lists. Debian. Org
Subject: RE: Rebuilding the Kernel Mini HOW TO


I found this document bery useful.

http://newbiedoc.sourceforge.net/system/kernel-pkg.en.html

I did not follow it to the letter, but I did manage to rebuild the kernel
first time with no problems.

Perhaps also a section on why someone might want to rebuild, I found the
pcmcia so much easier in 2.4.18.

Other reasons, for me were, enabling sound, video/frame buffering, and
removing stuff I just did not need, like scsi drivers.

Matthew Joyce



-Original Message-
From: Michael Olds [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, 15 October 2002 7:21 AM
To: Debian-User@Lists. Debian. Org
Subject: Rebuilding the Kernel Mini HOW TO


Hello,

This is still a draft. This is a request for advanced users to take a look
see and add comments or suggestions. The idea is a step by step instruction
for Fresh New to Linux/Debian users, with the hope that they will be
encouraged to attempt a rebuild of their kernel almost first thing after
initial install. I use a technique which is a blend of using a file manager
along with a terminal so I can see where I am at and yet do what I need to
do in the terminal. This is a good way to work into using the terminal.



[ ]A. Before 1: Download and create a recovery floppy from:
http://www.toms.net/rb/

[ ]1.a. Download the latest kernel source package (use a Debian source; or
make sure you have read all the documentation above) to /usr/src (this is
where it will be installed if you use dselect, a package manager, synaptic
or aptitude). Best practices seem to indicate that the package be built
outside of /usr/src so creating /usr/src/linux and creating a sym-link from
/linux to /usr/src/kernel~ as in step 1.d below seems to be the way to
handle this.)

[ ]1.b. if no /usr/src/linux dir exists, make it
  $ cd /usr/src
  $ mkdir linux
[ ]1.c. change directory to /usr/src and unpack the source package -- use
the unpacking tools with your file manager, or:
  $ su #enter root password
  $ cd /usr/src
  $ tar --bzip2 -xvf kernel-source-2.4.18.tar.bz2 (or latest) [ ]1.d
create a symbolic link to /usr/src/linux
  $ ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux

[ ]2. Just to be professional: go to "/etc/kernel.pkg.conf" and modify the
file by inserting your "Kernel Source Maintainer Name" and e-mail address.

[ ]3. Make sure
   [ ]"expectk" if you want to use the user friendly: make xconfig
   [ ]"debhelper"
   [ ]"modutils"
   [ ]"libncurses5-dev"
   [ ]"bin86"
   [ ]"libc6-dev"
   [ ]"debianutils"
   [ ]"make"
   [ ]"bzip2" and
   [ ]"kernel-package" are installed.

   R Notes: don't use gcc 3.x, but something like gcc 2.95. Which I do not
understand. Clarification?

[ ]4. (on 2nd and subsequent rebuilds) Copy old configuration file to
/usr/src/kernel-source-2.4.18
  $ cd /usr/src/kernel-source-2.4.18
  $ cp /boot/config-2.4.18-386 .config # get current config as
default

[ ]5. # cd to /usr/src/linux

  $ make menuconfig or make xconfig

and make your selections

   To do this you need to have some familiarity with your hardware and what
you are intending to use the system for. This you need to work out yourself.
Once this is figured out, it might be helpful, on the first rebuild, to make
your updates with reference to the original configuration, (called
config-2.4.18bf24 (or your kernel version)) which you should find under
/root using the "M" (module) option where a module was used and building the
feature into the kernel (*) or [y] option where it was built into the
kernel. Use the HELP option for brief explanations. This at least until you
understand which can be safely changed under your setup.

  Note: If you're using --initrd, make sure "Compressed ROM file system
support" is built in, or you'll get a kernel panic (need CramFS).

  (Side note: I see that answers my question about how to get it in
there...sorry!)


[ ]6.a.  [ ]# make-kpkg clean
   6.b.  [ ]# make-kpkg -revision=

NOTES: make-kpkg kernel_image actually does make oldconfig and make dep. Do
not use --initrd if initrd is not used. The current Debian initrd boot
script requires a cramfs kernel patch, if you obtain kernel source from
non-Debian archives; see http://bugs.debian.org/149236. As I understand it,
initrd is used in situations where many computers use a highly modularized
kernel, or

RE: Rebuilding the Kernel Mini HOW TO

[Michael Olds]

Russell,

No wonder none of the documentation around is exactly right or up to date!
For me this needs to be translated into Window-speak. In windows, a link is
like a file that when it is clicked, opens another file. I do not know what
is needed here, but what worked, and the only thing that worked was a "link"
in /usr/src/linux/ (in Windows there could not be such a thing as a link
that was /usr/src/linux) that was named kernel-source-2.4.18 and to "make
xconfig" you needed to cd to /usr/src/linux/kernel-source-2.4.18

I am not sure either that this is necessary, and it sure looks useless, but
I was following the instructions in one of those references...the one that
used the xconfig GUI.

And doing it as it is written here worked (I was able to rebuild and use my
kernel this way).

Alternatively, I think cd to usr/src/kernel-source-2.4.18/ (not a link)
would also work for xconfig, but that would not follow the suggestion from
elsewhere that one should not build the package in usr/src the way I read
it.

Madness! All I can say is that it seems this effort to get this straight
will be worth it in the end for other Windows converts.

Best Wishes!
Mike Olds www.buddhadust.org


-Original Message-
From: Russell [mailto:[EMAIL PROTECTED]]
Sent: Monday, October 14, 2002 6:09 PM
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: Rebuilding the Kernel Mini HOW TO


Michael Olds wrote:
>
> Thank you,
>
> I am still in a fog in terms of simple things like symlinks. What you want
> is to end up with a link in the linux directory called
kernel-source-2.4.18
> that links to /usr/src/kernel-source-2.4.18. So how should that be said?
>
> [ ]1.d create a symbolic link to kernel-source-2.4.18 from /usr/src/linux
> $ ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux

No. I don't know if you even *need* the linux symlink. But if you
do, the symlink is /usr/src/linux and it points to
/usr/src/kernel-source-2.4.18.


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Rebuilding the Kernel Mini HOW TO

[Michael Olds]

Thanks Bruce, please see my response to Russell...there is some kind of
misunderstanding (no doubt in my head) about the meaning of link and whether
a link is to or from, etc. What I described actually worked, so I need to
figure out how to say it so everyone will understand.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Bruce [mailto:[EMAIL PROTECTED]]
Sent: Monday, October 14, 2002 6:17 PM
To: [EMAIL PROTECTED]
Subject: RE: Rebuilding the Kernel Mini HOW TO


> We're getting there.
>
> Yes, I have it incorrectly, what is wanted is that step five should read:
>
>[ ]5. # cd to /usr/src/linux/kernel-source-2.4.18
>
> Corrected, thanks.

I'm not sure that is what you want to do; what you want is your kernel
source
in a folder that clearly identifies it by its number, i.e.:

/usr/src/kernel-source-2.4.18

Then, to allow that to be found easily, make a symlink from that folder, and
call the symlink "linux"

# cd /usr/src
# ln -s kernel-source-2.4.18 linux

Now,  typing "cd /usr/src/linux" will take you into
"/usr/src/kernel-source-2.4.18"


Good idea to put the basic requirements in a simple howto, and thanks for
taking that on. I would not mesh the "generic" instructions with Debian
instructions, though. For Debian, my mini-kernel howto would be brief (along
with your desctiptions and explanations):

# apt-get install kernel-package
# apt-get install kernel-source-2.4.18
# apt-get install 
# ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux
# cd /usr/src/linux
# make xconfig
# make-kpkg clean
# make-kpkg -rev mypc.1 --append-to-version mypc kernel_image
# cd ..
# dpkg -i kernel-image-2.4.18mypc_mypc.1_i386.deb
reboot

Bruce


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Rebuilding the Kernel Mini HOW TO

[Michael Olds]

Osamu,

Yes I did read those sources, however they created problems...for those
using xconfig taking out the pcmcia support in this way made it impossible
to use. (The what to do with old kernel headers problem needs to be spelled
out, the issue of where to make the package needs to be spelled out) Not
really trying to reinvent the wheel, trying to get the wheel on the ground
for people like myself who know absolutely next to nothing about what they
are doing in Linux, but who still need to do a couple of things before they
start investing a lot of time in setting things up and customizing:
rebuilding the kernel and learning how to do backups. And second these
instructions need to be geared towards the individual user, not the
corporate sysadmin. The instructions out there look like they are step by
step to the old hands (I know how this works) but they really do not resolve
doubts that occur between steps for the newbie. That is what I am trying to
do with this, not trying to step on anyone's toes. You see your discussion
of certain issues below makes no sense whatsoever to me and I believe it
probably makes perfect sense.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Osamu Aoki [mailto:[EMAIL PROTECTED]]On Behalf Of Osamu
Aoki
Sent: Monday, October 14, 2002 6:55 PM
To: [EMAIL PROTECTED]
Cc: Michael Olds
Subject: Re: Rebuilding the Kernel Mini HOW TO


Hi,

Did you read make-kpkg documentation especially by Kent West?
I think instead of reinventing wheel, suppliment existing good document
as patch bug report.

 /usr/share/doc/make-kpkg/README.gz (last section is by Kent)

For this, use unstable version.

Also my Debian Reference has a (good) summary.  Bug report is welcome.

  http://qref.sf.net/Debian/reference/ch-kernel.en.html or
  http://www.debian.org/doc/manuals/reference/ch-kernel.en.html

Lastly, are you aware of following facts (Content from recent post by
prominent Debian person. I am worried about mine ...):

* kernel header files for user application

Some may still say:
>  The problem under Debian and SuSE is that the Linux kernel is
>  not installed under /usr/src/linux - so you just need to ensure
>  that /usr/src/linux links to the directory containing the Linux
>  kernel sources. ...

Colin Watson <[EMAIL PROTECTED]> said:
That is absolutely terrible advice. It's a very good way to break a Debian
system, and you should ignore it. Linus doesn't recommend this old way of
doing
things nowadays either.

Instead, if you *need* kernel headers for something, alter the Makefiles so
that their include path points to
/include/linux, etc.

Most "normal" programs don't need kernel headers and in fact may break if
you
use them directly; instead they should be compiled against the headers with
which glibc was built, which are the versions in /usr/include/linux and
/usr/include/asm.

 END OF QUOTE ---

Do not ask me but it looks like using /usr/src/linux seems to be bad
idea.  (OOps, I do that in 2nd classic method.  Just do not compile
any application, this may be OK.. I need to check why this is bad...)

On Mon, Oct 14, 2002 at 02:19:48PM -0700, Michael Olds wrote:
> Hello,
>
> This is still a draft. This is a request for advanced users to take a look
> see and add comments or suggestions. The idea is a step by step
instruction
> for Fresh New to Linux/Debian users, with the hope that they will be
> encouraged to attempt a rebuild of their kernel almost first thing after
> initial install. I use a technique which is a blend of using a file
manager
> along with a terminal so I can see where I am at and yet do what I need to
> do in the terminal. This is a good way to work into using the terminal.
>
> 
>
> [ ]A. Before 1: Download and create a recovery floppy from:
> http://www.toms.net/rb/
>
> [ ]1.a. Download the latest kernel source package (use a Debian source; or
> make sure you have read all the documentation above) to /usr/src (this is
> where it will be installed if you use dselect, a package manager, synaptic
> or aptitude). Best practices seem to indicate that the package be built
> outside of /usr/src so creating /usr/src/linux and creating a sym-link
from
> /linux to /usr/src/kernel~ as in step 1.d below seems to be the way to
> handle this.)
>
> [ ]1.b. if no /usr/src/linux dir exists, make it
>   $ cd /usr/src
>   $ mkdir linux
> [ ]1.c. change directory to /usr/src and unpack the source package -- use
> the unpacking tools with your file manager, or:
>   $ su #enter root password
>   $ cd /usr/src
>   $ tar --bzip2 -xvf kernel-source-2.4.18.tar.bz2 (or latest)
> [ ]1.d create a symbolic link to /usr/src/linux
>   $ ln -s /usr/src/kernel-source-2.4.18 /usr/src/linux
>
> [ ]2. Just to be professional: go to "/etc/kernel.pkg.conf" and modify the
> file by insertin

Anonymous Proftp setup problems

[Michael Olds]

I am unable to get Proftp Anonymous ftp working (download only; "real" use
on the internal network is working fine), and am hoping that someone here
can help me out with this. I have read all the Proftp docs faqs etc, and
subscribe to the list where I have not received any help. I have checked,
hosts_allow and deny, ftpaccess, I have user ftp, and nogroup, I have
uninstalled the files set up by wu-ftp which I had on originally, Proftp was
setup as a Debian package, and I have uninstalled it, reinstalled it as
initrd, reinstalled it as standalone; this is on a Debian Woody box. I am
running in standalone mode, I have checked the syntax, and I have followed
the other suggested remedies I have seen.

If I run proftpd -nd5 debugging I get:

The error log reads: failed binding to 0.0.0.0, port 21; address already in
use
Check server type directive to ensure you are configured correctly:
I have

ServerType standalone
Port 21

I am not aware of anything else that might be using Port 21...and if it is,
it is out of service most of the time because I do not usually have that
port open on the router.

So first, let me ask a fundamental: Is this setup of the Anonymous Ftp
directive correct:

 (this is a partition /ftp user: me,rwx; group: ftp-data,rx;
world, rx)

  Userftp
  Group   nogroup
  UserAlias   anonymous ftp
  RequireValidShell off
  # Limit the maximum number of anonymous logins
  MaxClients 10
  DisplayLogin   welcome.msg
  DisplayFirstChdir .message

  

  DenyAll

  


I have now reached the stage where it looks like I am being logged in, but
it a) takes a very long time, and b) times out before it gives me a
directory listing

This is the error message I get from the ftp client latest message first:

  *** CuteFTP Pro 2.0 - build Dec  4 2001 *** see notations within

STATUS:>Getting listing ""...
STATUS:>Resolving host name ftp.mydomain.net...
STATUS:>Host name ftp.mydomain.net resolved: ip = 00.000.00.00.
STATUS:>Connecting to ftp server ftp.mydomain.net:21 (ip =
00.000.00.00.)...
STATUS:>Socket connected. Waiting for welcome message...
  220 ProFTPD 1.2.5rc1 Server (mydomain.net) [mydomain]
STATUS:>Connected. Authenticating...
COMMAND:>   USER anonymous
  331 Anonymous login ok, send your complete email address as your
password.
COMMAND:>   PASS *
  230 Anonymous access granted, restrictions apply.
STATUS:>Login successful.
COMMAND:>   PWD
  257 "/" is current directory.
STATUS:>Home directory: /
COMMAND:>   FEAT
  500 FEAT not understood.
STATUS:>This site doesn't support the 'features' command.
COMMAND:>   REST 100
  350 Restarting at 100. Send STORE or RETRIEVE to initiate transfer.
STATUS:>This site can resume broken downloads.
COMMAND:>   TYPE A
  200 Type set to A.
COMMAND:>   REST 0
  350 Restarting at 0. Send STORE or RETRIEVE to initiate transfer.
COMMAND:>   PASV
  227 Entering Passive Mode (000,000,000,00,128,14). (NOTE: IP of Local
server)
COMMAND:>   LIST
STATUS:>Connecting ftp data socket 000.000.000.00:32782...(also IP of
local server...does 32782 need to be open on the router...and I did try
that, with no change)
ERROR:> Timeout.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Anonymous Proftp setup problems

[Michael Olds]

Nate, thank you for this response,

I am behind my router's firewall, but set up with static external and
internal IPs:

In Proftp (global) I Set up:
UseReverseDNS   off
IdentLookupsoff
MasqueradeAddress   000.000.000.00 of the servers internal IP
PassivePorts 6-65535

and set up my router to listen on 21 and 6-65535

and I still get timed out:

COMMAND:>   PASV
227 Entering Passive Mode (internal.server.ip.no,251,62).
COMMAND:>   LIST
STATUS:>Connecting ftp data socket internal.server.ip.no:64318...
ERROR:> Timeout.

I did #fuser -n tcp 21 and got:
21/tcp: 230

I really don't know what that means. I do not have a user 230

I also did netstat -an | grep 21 and got a bunch of stuff, but didn't see
anything that rang any bells. But here too I really do not know what I am
looking for.

Best Wishes!
Mike Olds www.buddhadust.org


-Original Message-
From: nate [mailto:debian-user@;aphroland.org]
Sent: Thursday, October 24, 2002 1:13 PM
To: [EMAIL PROTECTED]
Subject: Re: Anonymous Proftp setup problems


Michael Olds said:
> I am unable to get Proftp Anonymous ftp working (download only; "real"
> use on the internal network is working fine), and am hoping that someone

you mention "internal" network. to me this makes me think you
are running it behind NAT. passive ftp does not work well with NAT.
try your client in ACTIVE ftp mode or put the ftp server infront
of your NAT box.

>
> If I run proftpd -nd5 debugging I get:
>
> The error log reads: failed binding to 0.0.0.0, port 21; address already
> in use

run fuser -n tcp 21

if that doesn't turn up anything then check

netstat -an | grep 21

sounds like your problem is NAT though, serving FTP data behind
dynamic NAT is not a good idea. FTP works fine if you use static
NAT(1 external IP maps to 1 internal ip on all ports). I have only
done this on cisco routers though.

nate




--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Anonymous Proftp setup problems

[Michael Olds]

nate,

First, you are correct,

The information I gave you about the Masquerade directive should have been:
MasqueradeAddress external.static.ip#

I gave you and used internal.static.ip#, but internal or external, global or
within the  directive, it still doesn't work. With the
external.static.ip masquerade I get this error:

COMMAND:>   PASV
  227 Entering Passive Mode (external,static,ip,234,160).
COMMAND:>   LIST
STATUS:>Connecting ftp data socket external.static.ip:60064...
ERROR:> Can't connect to remote server. Socket error = #10060.
ERROR:> Failed to establish data socket.

and, additionally, (naturally enough) internal network ftp no longer works.
(If using the real external IP had worked for anonymous, not for the
internal network, then I would have thought that assigning another IP
exclusively to the ftp server might have solved the problem, but not now).

I am behind a USRobotics broadband router which has a built in firewall. I
have two boxes, a linux server and a windows 2000 workstation each with
their own what I call "internal" static IP numbers. I am fresh enough to
doubt my understanding of Network Address Translation to need to describe
what I know and let you guess: My external IP number connected to the
Internet connected to the Router is Static. Each box has it's own static
number (no DHCP) connected to the other side of the router. That sounds like
NAT to me, but maybe not. The Router is set up to Exclude All incoming
traffic except when specifically allowed or when in response to a request
coming from within (I think that's how to describe it).

There is no "config" for this router, it's handled via browser window
GUI...you click buttons, but I am reasonably sure it is set up properly for
the situation as the tech support guy walked me through it and I asked
enough questions to get it straight in my head what to do. (And, as it looks
to me, I am being logged in.)

I checked inetd and ftp is  (If I didn't say it before, proftpd is set
up as standalone.)

To start and stop I have just been rebooting (this is a small personal setup
and it doesn't matter if I am disconected for a few minutes)...but I would
like to know if there were something like  # apachectl stop

Best Wishes!
Mike Olds www.buddhadust.org



-Original Message-
From: nate [mailto:debian-user@;aphroland.org]
Sent: Thursday, October 24, 2002 2:56 PM
To: [EMAIL PROTECTED]
Subject: RE: Anonymous Proftp setup problems


Michael Olds said:
> Nate, thank you for this response,
>
> I am behind my router's firewall, but set up with static external and
> internal IPs:

what is the configuraiton? you sure it is setup for static NAT?
are you using NAT at all?(static or dynamic?)

>
> In Proftp (global) I Set up:
> UseReverseDNS off
> IdentLookups  off
> MasqueradeAddress 000.000.000.00 of the servers internal IP
> PassivePorts 6-65535
>
> and set up my router to listen on 21 and 6-65535
>
> and I still get timed out:

I've never tried these options before...but it looks like from
the doc the MasqueradeAddress needs to be set to the REAL ip
not the NAT'd ip. from your description above it sounds as if you
put the NAT IP in that field.

> I did #fuser -n tcp 21 and got:
> 21/tcp:   230
>
> I really don't know what that means. I do not have a user 230

that means process id 230 is using tcp port 21

i just said run that because you mentioned running a copy of proftpd
and it spit back cannot bind(address already in use).

it is possible that proftpd is configured to run through inetd, check
/etc/inetd.conf and /etc/xinetd.conf(if you have xinetd) to be sure.

I am not famillar with the kind of router you have which allows
it to 'listen' on those ports. a public ftp server that I setup
for my former company is setup using static 1:1 NAT behind a
cisco 2500 series router, the config directive:
ip nat inside source static 10.115.17.195 65.115.17.195

the ftp server is ftp.graphon.com you can try to ftp to it and
grab a file to see, as far as the ftp server is concerned it
has a real IP, the router handles everything, no special
configuration needed on the server itself, its totally transparent.

change the proftpd directive to the real ip of the system if your
currently using the NAT'd ip, and see what happens. If your not sure
proftpd is restarted, shutdown all copies of it, ftp to localhost
to be sure it is not running, run nestat -an | grep 21 to be sure
nothing is using port 21(you may see some TIME_WAIT messages, if
so, wait until they are gone, can take up to 15-20 minutes) then
start proftpd again.


nate




--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Apache handling SSI

[Michael Olds]

Hello,

I know I could just do this, but I am a little gun shy: can an Apache server
that now runs SSI using an .shtml mime type and an .shtml AddHandler be made
to run SSI on .htm files by just including
AddHandler server-parsed .htm ?

Not sure I said that correctly: I want to include includes in plain old .htm
files and forget about .shtml.

And is there some reason for not doing this?

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Apache handling SSI

[Michael Olds]

OK Thanks something to think about

-Original Message-
From: nate [mailto:debian-user@;aphroland.org]
Sent: Thursday, October 24, 2002 7:24 PM
To: [EMAIL PROTECTED]
Subject: Re: Apache handling SSI


Michael Olds said:
> Hello,
>
> I know I could just do this, but I am a little gun shy: can an Apache
> server that now runs SSI using an .shtml mime type and an .shtml
> AddHandler be made to run SSI on .htm files by just including
> AddHandler server-parsed .htm ?
>
> Not sure I said that correctly: I want to include includes in plain old
> .htm files and forget about .shtml.
>
> And is there some reason for not doing this?


it should work fine yes. as for why not, probably a bit less load
on the system if it doesn't have to inspect every single .html file
for includes.

nate




--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Proftp behind firewall

[Michael Olds]

Greetings once again,

I have requested help on this issue before, but the problem was never
resolved, so here goes one more try (a little different this time; I have
given up trying to set up anonymous ftp):

I need to get plain-ol' ftp working.

I am using Proftp. My setup is:
Intel 850 P4 1.9G; 1GbRAM; Win2000;
Intel 845 P4 1.6G; 768MbRAM; Debian Linux (3.0 Woody);
Connected to a static DSL IP through a US Robotics Broadband Router.

I have given up on Anonymous ftp; I just need to provide a couple of users
secure access to websites I want to host for them.

Currently I am using sftp (SSH2) to access my own websites from the Windows
box behind the router. When I try to use the Internet I cannot make the
connection (various error messages...tell me what setup you want me to try
and I will post the error message).

I have read the documentation and nothing quite deals with my situation
which surely cannot be unique. And I have read each of the following:
Slow logins:
  http://www.proftpd.org/docs/faq/faq_full.html#AEN414
Networking/firewall/NAT issue:
http://www.castaglia.org/proftpd/doc/contrib/ProFTPD-mini-HOWTO-NAT.txt
http://slacksite.com/other/ftp.html (See extract below)
Debugging:http://www.castaglia.org/proftpd/doc/contrib/ProFTPD-mini-HOWTO-De
bugging.html

I need to know:
What ports should be opened through the router's firewall? I have tried 21
20 and those plus the range 6-65000.

What should my proftpd.conf look like:

I have tried variations on the folowing in the proftpd.conf:

# UseReverseDNS off
# IdentLookups  off
#
# MasqueradeAddress  both ftp.mydomain.net and http://www~ and
00.000.00.00 the static IP Address used to connect to the Internet
# PassivePorts  6 65535
# AllowForeignAddress on
#

Neither ftp nor nogroup are in /etc/ftpusers

What I have not changed is anything to do with configuring the NAT as I am
not sure that is what I need to do (or what I need to do to change it)...so
I need to know if it is what I need to do.

I have reconfigured this endless times over the last couple of months and I
am tuckered! Anyone can guide me through this or point me in the right
direction, I would appreciate it much.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Proftp behind firewall problem solved

[Michael Olds]

Hello again,

I should be recognizing this by now...whenever I am about to send a message
to a help list I am about five minutes from solving the problem...and of
course if I actually get as far as describing the problem in great detail
the solution is sure to make me look like an idiot.

In this case: opening port 21 or 20, or setting up a range of thousands of
open ports for PASV mode for SFTP won't do it. Gotta open port 22 in the
firewall. That was all it took.

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

EXIM relay for specific users

[Michael Olds]

EXIM relay for specific users

Hello again,
I cannot get EXIM to let me relay email. I have read Hazel, EXIM (O'Reilly),
and numbers of other documents but I am missing something.

The setup:
Intel 845 P4 1.6G; 768MbRAM; Debian Linux (3.0 Woody)
behind a US Robotics Firewall connected to a Static IP#
The two computers on the LAN have static IPs

I would like to allow valid users to relay mail from their computers to
anywhere through my server.

Right now everything but that appears to work fine. I can mail to my account
on the server and download the mail from the server via POP3 and I can mail
from the users desktop to anywhere.


host_auth_accept_relay = * (also tried mydomain.net)

I have both "login" and "plain" uncommented in the exim.conf and I have
created an /etc/exim/passwd file

I always get the following error: Server response 550 relaying prohibited by
administrator.

Can someone tell me what I need to do to get this working?

Thanks!

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: EXIM relay for specific users

[Michael Olds]

OK, Paul I don't understand this, but I did find your archive solution...I
think...about how to use PAM for authentication. I will check into this in
the morning.

Thanks,

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Paul Johnson [mailto:[EMAIL PROTECTED]]
Sent: Monday, December 02, 2002 8:12 PM
To: debian-user List
Subject: Re: EXIM relay for specific users


On Mon, Dec 02, 2002 at 10:00:35PM -0500, P. M. Wright wrote:
> > > EXIM relay for specific users
> >
> > Check the archives.  Keywords to look for are "ASMTP" and "exim."
> >
> > I posted the appropriate fix a while back.
>
> RTFM?!? YFA!

This isn't RTFM.  This is an already recently answered question,
readily found in the archive.  Don't be an ingrate.

--
 .''`. Baloo <[EMAIL PROTECTED]>
: :'  :proud Debian admin and user
`. `'`
  `-  Debian - when you have better things to do than to fix a system


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: EXIM relay for specific users

[Michael Olds]

I guess I didn't see the post with the RTFM and YFA. I said I read O'Reilly.
I have read the manuals. I have searched Google and have a good collection
of links. I have played with the configuration for more than a month
now...first began (look in the archives...) in October. I work at it until I
can't think of anything more to do and then ask. I don't think I am abusing
the system. Either the answer to this question is not spelled out in the
documentation or I am reading in incorrectly in some way that needs to be
changed, either way I am stuck. (I did not see the series of answers Paul
gave because I have not been connected to the list for the last month or
more.)

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Paul Johnson [mailto:[EMAIL PROTECTED]]
Sent: Monday, December 02, 2002 8:12 PM
To: debian-user List
Subject: Re: EXIM relay for specific users


On Mon, Dec 02, 2002 at 10:00:35PM -0500, P. M. Wright wrote:
> > > EXIM relay for specific users
> >
> > Check the archives.  Keywords to look for are "ASMTP" and "exim."
> >
> > I posted the appropriate fix a while back.
>
> RTFM?!? YFA!

This isn't RTFM.  This is an already recently answered question,
readily found in the archive.  Don't be an ingrate.

--
 .''`. Baloo <[EMAIL PROTECTED]>
: :'  :proud Debian admin and user
`. `'`
  `-  Debian - when you have better things to do than to fix a system


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: EXIM relay for specific users

[Michael Olds]

Andrew,

I also couldn't find it using the search tool, kept bringing me back to
alternative ways to search. I used the author index. It's in November under
Paul's name.

But I copied the relevant posts very much fits what you are after methinks:

I want to setup a mailserver, like the servers an ISP has. I have exim
installed and it delivers and recieves mail just fine. I have a static
IP and a domain. I also have uw-imapd-ssl running , which is really
neat. I can read email over ssl links from anywhere on the net with
mozilla-mail, outlook express, etc.

But this is only half of the deal. I can't send email from my email
clients because exim rejects the mail, saying it won't relay email. So
what I want is to have exim ask for a login password, and if possible do
this over an ssl connection too.

At first I tried using PAM, but nothing I did effected exim, then I
noticed in my exim.conf I found some lines about setting up
authentication but I can't quite get it working. Are there any guides on
this? I really want to set this up using SSL since most of the email
logins will also be shell logins, so sending those logins unencrypted
seems like a bad idea.

Is there anykind of a guide I can read? I tried searching google, but I
get lots of stuff not related to what I want to do, and the NAG has some
info, but not on having smpt password login.
On Mon, Nov 04, 2002 at 08:56:10AM -0500, ZZ wrote:
> At first I tried using PAM, but nothing I did effected exim, then I
> noticed in my exim.conf I found some lines about setting up
> authentication but I can't quite get it working. Are there any guides on
> this? I really want to set this up using SSL since most of the email
> logins will also be shell logins, so sending those logins unencrypted
> seems like a bad idea.

I don't know about the SSL part, but I'll save you the royal pain in
the ass associated with exim and pam...

In /etc/exim/exim.conf...
In Main Configuration Settings add...

# Fix problems with auth.

exim_user=mail
exim_group=shadow

In Authentication Configuration, make it read something like...

##
#   AUTHENTICATION CONFIGURATION #
##

# Look in the documentation (in package exim-doc or exim-doc-html for
# information on how to set up authenticated connections.

plain:
  driver = plaintext
  public_name = PLAIN
  server_condition = "${if pam{$2:$3}{1}{0}}"
  server_set_id = $2

login:
  driver = plaintext
  public_name = LOGIN
  server_prompts = "Username:::Password::"
  server_condition = ${if pam {$1:${sg{$2}{:}{::}}}{yes}{no}}
  server_set_id = $1

# End of Exim configuration file
Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Holiday Wish list

[Michael Olds]

oops, sorry for the off-list post Bill, here's what I said:

There is definately something very inefficient about these mailing lists for
sure. I think a better approach is the Bulletin Board with an active
moderator, so that a permanent thread could be opened on a topic and
additions and questions asked eventually building up a valuable database of
information. Here as you say, things come up over and over with maybe slight
variations, and get lost very easily. The problem with old, out of date
information (expecially books you don't find out about until you've paid
your $40 bucks) is another bad one...a price we pay I suppose for not being
a huge market. Still more energy is wasted I'd bet than it would take to do
it right the first time.

Best Wishes!
Mike Olds www.buddhadust.org



-Original Message-
From: Bill Moseley [mailto:[EMAIL PROTECTED]]
Sent: Thursday, December 05, 2002 8:49 AM
To: [EMAIL PROTECTED]
Subject: Holiday Wish list


Ok, so this is coffee induced, but my one wish is that all HOWTOs include
a last modified date so I don't spend hours learning and installing the
old way, and that all HOWTOs explain *why* each step is done instead of
just the steps.  (Is that two wishes?)

The main reason I pick Debian is that I want to install just what I need
and use, and understand *why* it's installed and *how* the parts fit
together. I've often been tempted by Linux from Scratch for this reason,
but I love apt-get too much.

And if I can ask for anything, I'd ask for Debian specific HOWTO wikis.
I think most all debian setup and config options are discussed on this
list - often over and over.  I'd love to see individual topic HOWTOs
specific to debian that can be annotated/updated as questions and answers
come up on the list.

And then perhaps a CGI interface where a new user can be asked a series of
questions about what kind of system they want and get a list of steps that
are pointers to the individual HOWTO wikis.  Kind of like tasksel, but in
documentation.

It would also have to be a rather smart system to detect that addition of,
say ide-scsi, might need a kernel recompile and include those docs.

Ok, it would never work as TIMTOWTDI when setting up a machine.  But for
newcomers some time there's too many ways, and one good way would be a
nice start.

The other reason to reject such an idea is that there already is a ton of
really good docs available.  Still, I like the wiki idea to update docs as
things come up on this list.

I say all this because I have been trying to keep good notes when I set up
a new machine.  I install a base system and then try and apt-get only what
I need.  My notes try and explain why I need to install something, some
basics of how it works (e.g. "/etc/init.d/foo runs, reads
/etc/default/foo.conf, and foo updates something else" kind of thing).
Rarely perfect but takes some of the mystery out of it.  And I cut-n-paste
the actual commands I type into my notes which make re-using them really
easy (as I can then cut-n-paste from the notes).

I reinstalled a machine a few days ago and I picked sections of what I
wanted to install from the notes of three other installs and it was so
easy to set up that new machine for me.  That's what made me think about a
way to select the things I want setup and have the notes compiled together
for me.

Ok, enough of that.  Time for another cup of coffee.

--
Bill Moseley [EMAIL PROTECTED]


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: EXIM relay for specific users

[Michael Olds]

Andrew,

Glad you replied to this! This was Paul Johnson's solution not mine. I used
the other suggestion that was made (simpler, so I tried it first thinking it
would probably not work as I thought I had already tried it) that is to put
the public IP number in the
hosts_accept_relay=
directive.

Unfortunately when I saved this I clipped off the name of the person who
gave me that answer...thanks to you whoeva you is!

Now...how come it's working so slowly? I commented out the lookup directive
to no effect. Maybe I should try Paul's solution.

I'm afraid I am not up to speed on SSL...I'm listening if anyone wants to
jump in.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Andrew R Reid [mailto:[EMAIL PROTECTED]]
Sent: Thursday, December 05, 2002 9:52 PM
To: Debian User
Subject: RE: EXIM relay for specific users



> -Original Message-
> From: Michael Olds [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, 4 December 2002 12:10 AM
> To: debian-user List
> Subject: RE: EXIM relay for specific users

> I don't know about the SSL part, but I'll save you the royal pain in
> the ass associated with exim and pam...
>
> In /etc/exim/exim.conf...
> In Main Configuration Settings add...
>
> # Fix problems with auth.
>
> exim_user=mail
> exim_group=shadow
>
> In Authentication Configuration, make it read something like...
>
> ##
> #   AUTHENTICATION CONFIGURATION #
> ##
>
> # Look in the documentation (in package exim-doc or exim-doc-html for
> # information on how to set up authenticated connections.
>
> plain:
>   driver = plaintext
>   public_name = PLAIN
>   server_condition = "${if pam{$2:$3}{1}{0}}"
>   server_set_id = $2
>
> login:
>   driver = plaintext
>   public_name = LOGIN
>   server_prompts = "Username:::Password::"
>   server_condition = ${if pam {$1:${sg{$2}{:}{::}}}{yes}{no}}
>   server_set_id = $1
>
> # End of Exim configuration file
> Best Wishes!
> Mike Olds www.buddhadust.org
>

Thank you very much for this Michael.  It is working perfectly.  How do I do
this with ssl?

There are a lot of parts of my exim.conf that are commented out relating to
tls conections.  Is this similar to ssl?


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Holiday Wish list

[Michael Olds]

Thanks Karsten, I'll check it out...Wiki's are just entering my
consciousness at this point...

-Original Message-
From: Karsten Self [mailto:[EMAIL PROTECTED]]On Behalf Of
Karsten M. Self
Sent: Friday, December 06, 2002 2:25 AM
To: [EMAIL PROTECTED]
Subject: Re: Holiday Wish list


on Thu, Dec 05, 2002 at 09:11:49AM -0800, Michael Olds
([EMAIL PROTECTED]) wrote:
> oops, sorry for the off-list post Bill, here's what I said:
>
> There is definately something very inefficient about these mailing
> lists for sure. I think a better approach is the Bulletin Board with
> an active moderator, so that a permanent thread could be opened on a
> topic and additions and questions asked eventually building up a
> valuable database of information. Here as you say, things come up over
> and over with maybe slight variations, and get lost very easily.

Which is one of the goals of a Wiki.  I'm running one myself, with a
focus on free software tools:

http://twiki.iwethey.org/twiki/bin/view/Main

...and it's aimed at filling in just this gap between mailing lists and
Usenet on the one hand and HOWTOs and hardcopy on the other.

Peace.

--
Karsten M. Self <[EMAIL PROTECTED]>http://kmself.home.netcom.com/
 What Part of "Gestalt" don't you understand?
   Geek for hire:  http://kmself.home.netcom.com/resume.html


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Proftp behind firewall problem solved

[Michael Olds]

Thanks for the response Rob,

I should be recognizing this by now...whenever I shoot off my mouth about
having solved a problem, it isn't really solved at all: Mistaking the fact
that it works for the solution of the problem.

1. This is a one-man 2-computer setup, so I generally ftp from behind the
firewall, but I would like it to work just to make it work, and I will
eventually want it to work for a couple of friends whose sites I would like
to host.

2. I am probably using the wrong terminology. I use Cute FTP. The settings
for this connection are:
SFTP using SSH2 (Secure Shell) port 22
Use Global Settings
and I have SSH2 set up on the Linux box and it works (meaning I connect and
I can up and download files).

My (thin-film level) understanding was that SSH2 actually pretended to be
whatever open ports were necessary for the communication, but all traffic
went through 22. At this time the firewall is actually closed on 20 and 21
and everything else except 22, 80, 110, 25.

The address I connect to is not the internal IP but the domain name; the
client log shows it is going through the Internet and is initializing SFTP
module.

So I don't know what to say here.

Answering your second response here, I do not know what more I can do to
look into it further except to let my understanding mature to the level
where I can see what I am overlooking. I bought O'Reilly's TCP/IP Network
Administration and a half dozen other books and have read the various
manuals (and Google how to pages where this very problem seems to be being
solved all the time) and tried a huge number of variations on that setup
with no success.

Best Wishes!
Mike Olds www.buddhadust.org




-Original Message-
From: Rob Weir [mailto:[EMAIL PROTECTED]]
Sent: Monday, December 09, 2002 12:39 AM
To: Debian-User
Subject: Re: Proftp behind firewall problem solved


On Mon, Dec 02, 2002 at 02:51:50PM -0800, Michael Olds wrote:
> Hello again,
>
> I should be recognizing this by now...whenever I am about to send a
message
> to a help list I am about five minutes from solving the problem...and of
> course if I actually get as far as describing the problem in great detail
> the solution is sure to make me look like an idiot.
>
> In this case: opening port 21 or 20, or setting up a range of thousands of
> open ports for PASV mode for SFTP won't do it. Gotta open port 22 in the
> firewall. That was all it took.

Uh, that's weird.  I don't think this can be very robust, since the data
port is semi-randomly chosen...Plus, it's the SSH port, so how are you
going to SSH/sftp/scp into that machine?

-rob


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: WSFTP problems from Win2000 to Debian--PLEASE HELP!

[Michael Olds]

Check 
your permissions. 
 
 
 
-Original Message-From: Sasha Karlik 
[mailto:[EMAIL PROTECTED]]Sent: Tuesday, December 10, 2002 9:01 
AMTo: [EMAIL PROTECTED]Subject: WSFTP problems 
from Win2000 to Debian--PLEASE HELP!
 
I have a network connection between a Debian box and 
a win2000 box. 
My WSFTP client browses to the file sytem on the debian 
box.
When I try to put files to the debian box, it locks and fails 
with this message: 
! Receive error: Blocking call cancelled
I tried with passive transfer mode both on and 
off.
The same error message appears, regardless.
Does anyone know what I can do to make my FTP 
transfers work from the Windows side?

RE: WSFTP problems from Win2000 to Debian--PLEASE HELP!

[Michael Olds]

Hello.
Responding to the subject message, I received the following from [EMAIL PROTECTED]
 
(in 
plain text) 
 
1. is 
this part of the Debian Users list setup? and I would like to be able to do this 
for myself, can anyone explain how this is done?
 

This is an automatic reply. Your message
HAS BEEN REGULARLY DELIVERED.
However, some MIME parts have been stripped from it.
YOU DO NOT NEED TO RESEND THIS MESSAGE.
The MIME type(s) that have been removed are listed at the bottom of this 
message. Please refrain from sending further messages with these MIME 
types
to this address.
If your original message was sent to a mailing list, you may want to 
seriously reconsider sending the MIME type(s) listed below to the list.
If you find something unclear about this notification, feel free to reply 
to
this message.
This is a list of the MIME types of every stripped part:
A. text/html
-Original Message-From: Michael Olds 
[mailto:[EMAIL PROTECTED]]Sent: Tuesday, December 10, 2002 9:20 
AMTo: Sasha Karlik; [EMAIL PROTECTED]Subject: 
RE: WSFTP problems from Win2000 to Debian--PLEASE HELP!
Check 
your permissions. 
 
 
 
-Original Message-From: Sasha Karlik 
[mailto:[EMAIL PROTECTED]]Sent: Tuesday, December 10, 2002 9:01 
AMTo: [EMAIL PROTECTED]Subject: WSFTP problems 
from Win2000 to Debian--PLEASE HELP!
 
I have a network connection between a Debian box and 
a win2000 box. 
My WSFTP client browses to the file sytem on the debian 
box.
When I try to put files to the debian box, it locks and fails 
with this message: 
! Receive error: Blocking call cancelled
I tried with passive transfer mode both on and 
off.
The same error message appears, regardless.
Does anyone know what I can do to make my FTP 
transfers work from the Windows side?

RE: EXIM relay for specific users

[Michael Olds]

On Mon, Dec 02, 2002 at 08:28:56PM -0800, Michael Olds wrote:
> I guess I didn't see the post with the RTFM and YFA. I said I read
O'Reilly.
> I have read the manuals...

I don't think Paul was too harsh...

...

First, I don't believe it was Paul that sent the RTFM and YFA post...at
least I hope not as I was grateful for the information he did provide.

Second for ANYONE to say something like RTFM and YFA is out of line in my
book. Given the state of the documentation, and considering that none of us
know the circumstances of the people writing for help the only reasonable
thing is to either answer the question or keepeeupeepeeclo. This applies
even moreso when someone states outright the fact that they HAVE read the
documentation and still did not find the answer.

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Proftp behind firewall problem solved

[Michael Olds]

p.s. It's standard practice to not CC anyone on the list, unless their
Mail-Followup-To header says to or they explictely ask.


Sorry if my reply went directly to you. I am not really in the "list"
habit...what usually happens is that I hit "Reply" in stead of "Reply to
All"



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Cracked cracker?

[Michael Olds]

This is a small sample from my access log. Can someone explain to me why
this person would repeatedly attempt access to my computer using the same IP
and the same requests over and over? This isn't to the point of being a DOS
attack; can't they see I don't have any of these things that they think will
enable them to crack my machine? Or is there something else going on here?


63.205.213.16 - - [11/Dec/2002:13:16:07 -0800] "GET /scripts/root.exe?/c+dir
HTTP/1.0" 404 270 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:07 -0800] "GET /MSADC/root.exe?/c+dir
HTTP/1.0" 404 268 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:07 -0800] "GET
/c/winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 278 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:07 -0800] "GET
/d/winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 278 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:08 -0800] "GET
/scripts/..%255c../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 292 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:08 -0800] "GET
/_vti_bin/..%255c../..%255c../..%255c../winnt/system32/cmd.exe?/c+dir
HTTP/1.0" 404 309 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:08 -0800] "GET
/_mem_bin/..%255c../..%255c../..%255c../winnt/system32/cmd.exe?/c+dir
HTTP/1.0" 404 309 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:08 -0800] "GET
/msadc/..%255c../..%255c../..%255c/..%c1%1c../..%c1%1c../..%c1%1c../winnt/sy
stem32/cmd.exe?/c+dir HTTP/1.0" 404 325 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:08 -0800] "GET
/scripts/..%c1%1c../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 291 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:08 -0800] "GET
/scripts/..%c0%2f../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 291 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:08 -0800] "GET
/scripts/..%c0%af../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 291 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:09 -0800] "GET
/scripts/..%c1%9c../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 291 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:09 -0800] "GET
/scripts/..%%35%63../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 400 275 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:09 -0800] "GET
/scripts/..%%35c../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 400 275 "-" "-"
63.205.213.16 - - [11/Dec/2002:13:16:09 -0800] "GET
/scripts/..%25%35%63../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 292 "-"
"-"
63.205.213.16 - - [11/Dec/2002:13:16:09 -0800] "GET
/scripts/..%252f../winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 292 "-" "-"


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Cracked cracker?

[Michael Olds]

This same guy is at it again today. The stupid thing to my mind is using the
same IP. How exactly do I go about reporting him to his ISP? Anyone
know...or should I just let him be?



-Original Message-
From: Antoine Jacoutot [mailto:[EMAIL PROTECTED]]
Sent: Thursday, December 12, 2002 8:55 AM
To: Michael Olds
Cc: Debian-User
Subject: Re: Cracked cracker?


On Thu, 2002-12-12 at 16:55, Michael Olds wrote:
> This is a small sample from my access log. Can someone explain to me why
> this person would repeatedly attempt access to my computer using the same
IP
> and the same requests over and over? This isn't to the point of being a
DOS
> attack; can't they see I don't have any of these things that they think
will
> enable them to crack my machine? Or is there something else going on here?

My apache access.log is full of those too...
I guess that must be some king of automated attack against IIS servers,
so I never worried.

Antoine




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Cracked cracker?

[Michael Olds]

Thanks everyone. I did know it was the code red virus or worm or whatever
(you can see the depth of my knowledge...haha)...I have seen the logs from
my hosted website before. This was on my own server, and what was
interesting this time was that this same IP kept showing up. Usually you get
a string of hits then the same or similar string with another IP. This (ok
not person, box) was going at it every ten minutes most of the afternoon
yesterday and again today (I shut down during the night). I am new enough
with my own setup to not have seen this aspect of the code red attack...web
server up and running for only about a month.

I emailed the ISP to tell them about it, they may have shut them (it) down.


Best Wishes!
Mike Olds www.buddhadust.org



-Original Message-
From: Craig Dickson [mailto:[EMAIL PROTECTED]]
Sent: Thursday, December 12, 2002 9:12 AM
To: Debian-User
Subject: Re: Cracked cracker?


Michael Olds wrote:

> This is a small sample from my access log. Can someone explain to me why
> this person

It's not a person.

> would repeatedly attempt access to my computer using the same IP
> and the same requests over and over? This isn't to the point of being a
DOS
> attack; can't they see I don't have any of these things that they think
will
> enable them to crack my machine? Or is there something else going on here?
>
>
> 63.205.213.16 - - [11/Dec/2002:13:16:07 -0800] "GET
/scripts/root.exe?/c+dir
> HTTP/1.0" 404 270 "-" "-"
> 63.205.213.16 - - [11/Dec/2002:13:16:07 -0800] "GET /MSADC/root.exe?/c+dir
> HTTP/1.0" 404 268 "-" "-"
> 63.205.213.16 - - [11/Dec/2002:13:16:07 -0800] "GET
> /c/winnt/system32/cmd.exe?/c+dir HTTP/1.0" 404 278 "-" "-"

[etc.]

I have to admit I'm amazed that anyone running a mailserver wouldn't know
what that was. Have you just set up one for the first time?

Anyway, this is the Windows Nimda virus trying to break into a
vulnerable installation of the Microsoft IIS server, not realizing that
you're not what it's looking for. Nimda has been doing this for at least
a year now, I think, and it got quite a lot of press when it first came
out.

Unsurprisingly, 63.205.213.16 appears to belong to a local cable company
that offers digital TV and cable modem services. So the machine in
question is probably just a Windows 2000 machine belonging to one of
their customers, who, typical of the average Windows users, has no idea
that he is running a web server on his computer, no idea that his
computer is infected with a virus, and no idea what the phrase "security
update" means.

Although, interestingly, telnetting to 63.205.213.16 on the www and smtp
ports gets "connection refused". So either the user has belatedly got a
clue, or they've simply gone offline and someone else now has their DHCP
address.

Craig


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Cracked cracker?

[Michael Olds]

Run the 'host' command on the IP address...

Craig

The Debian User listserver isn't all too healthy today...big lag. I did
finally run a tracert and e-mail the ISP and the attacks have stopped, so
maybe it got through.

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Removing devfsd

[Michael Olds]

I've been through this one. You need to recompile the kernel and change the
supported files.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Bob Hilliard [mailto:[EMAIL PROTECTED]]
Sent: Saturday, December 14, 2002 11:20 AM
To: [EMAIL PROTECTED]
Subject: Removing devfsd


 How can devfsd be removed safely?  I have purged the package,
cleaned out /dev and reconstructed the device files with `MAKEDEV
generic', but the partition is then unbootable.  On boot, fsck
reports: "fsck.ext3: No such file or directory while trying to open
/dev/hda6".

 To recover, I have booted another partition, mounted hda6 on
 /mnt, chrooted into hda6 and re-installed devfsd.  This is
 reproducible - I have done it three times.

Regards,

Bob
--
   _
  |_)  _  |_Robert D. Hilliard<[EMAIL PROTECTED]>
  |_) (_) |_)   1294 S.W. Seagull Way <[EMAIL PROTECTED]>
Palm City, FL 34990 USA   GPG Key ID: 390D6559



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: continuing email problems (fwd)

[Michael Olds]

Is it possible those undeliverables are just out? Here in California we are
having some very bad weather that has boinked cable, power, telephone, etc
for the last couple of days.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Cheryl Homiak [mailto:[EMAIL PROTECTED]]
Sent: Sunday, December 15, 2002 6:39 AM
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: continuing email problems (fwd)


Re-sending. It appears that my Shellword character set was wrong. sorry!


-- Forwarded message --
Date: Sun, 15 Dec 2002 08:26:36 -0600 (CST)
From: Cheryl Homiak <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Subject: continuing email problems

Having checked my exim logs on both my desktop (cable connection charter)
and laptop (ppp mailtag) computers, I find two different errors.
On the cable machine, I find in the logs when the receiving host tries to
take the email from charter: "relay prohibitted. You must authenticate
first." The email remains in my queue and is retried.
On the machine using ppp, some emails go to their destinations, but for
some hosts the mail is returned to me saying it is undeliverable and that
this is a fatal permanent error.
I appear to be able to get through to debian-user with my mailtag account
on my laptop, but at this point I am using my shellworld shell account
(knew there were good reasons to keep that).
Any HELP would be appreciated.
I think incoming mail is ok, but am not sure about that.
Also noticed that my ip address has changed radically sometime in the last
few days on my cable connection; this could have something to do with
things I suppose as I had the same ip address for a long time.

Cheryl


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Cracked cracker?

[Michael Olds]

Thanks Kenneth,

Another possibility, if you just want to stop logging the pests, is to 
put the following in the global section of your httpd.conf:


# Stop logging nimda requests, based on:
#http://lists.netfilter.org/pipermail/netfilter/2001-October/026587.html
#
SetEnvIfNoCase Request_URI /cmd.exe|/root.exe|/default.ida nimda
CustomLog /var/log/apache/access.log combined env=!nimda
# comment out next line to NOT log nimda requests
CustomLog /var/log/apache/nimda.log "%a %v %t %U" env=nimda
#
# suppress logging errors from serving 404s to the bastards:

  RedirectMatch (.*)/root.exe http://not.nimda.friendly.invalid$1
  RedirectMatch (.*)/cmd.exe http://not.nimda.friendly.invalid$1
  RedirectMatch (.*)/default.ida http://not.nimda.friendly.invalid$1



I have a line in there to log nimda requests to a seperate file, comment 
it out if you don't want it. If you're maintaining seperate logs for 
multiple VirtualHosts, repeat the CustomLog directive(s) for each one 
(adjust your paths):



 ServerName www.DOMAIN.TLD
 ServerAlias DOMAIN.TLD
 ServerAdmin [EMAIL PROTECTED]
 DocumentRoot /var/www/DOMAIN.TLD/htdocs
 UserDir disabled
 #keep nimda out of access.log
 CustomLog /var/www/DOMAIN.TLD/log/access.log combined env=!nimda
 CustomLog /var/www/DOMAIN.TLD/log/nimda.log "%a %v %t %U" env=nimda
 ErrorLog /var/www/DOMAIN.TLD/log/error.log



I am going to do this today.

Best Wishes!
Mike Olds www.buddhadust.org 


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: getting networking to work [was Re: ethernet card]

[Michael Olds]

-Original Message-
From: Andrew Hurt [mailto:[EMAIL PROTECTED]]
Sent: Monday, December 16, 2002 5:41 AM
To: [EMAIL PROTECTED]
Subject: getting networking to work [was Re: ethernet card]

For the most part, everything seems to be o.k., and tulip-diag/ifconfig/etc.
are
giving me the green light (so to speak), but I'm concerned because I can't
get
to the http://192.168.254.254/ in the router--"SpeedStream 2602" (caveat: no
modem/network connection to test, yet).


Have you tried turning everything off completely -- power off, then starting
up again?

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: cups and hpijs driver

[Michael Olds]

I'm catching up to this one late, and am being thrown off by the very start.

I have a DeskJet 960C

I am pretty sure it's hooked up right as it was running under RedHat
(meaning I have PnP turned off in the BIOS) and during one point in playing
around tonight I saw that something knew the printer was there.(don't ask,
I've been trying to find it again and can't).

I have installed the foomatic.bin and foomatic.db and cupsys and it's
depends
and have generated and saved in /usr/share/cups/model/HP-DeskJet`.ppd

But when I type lpadmin in the terminal I get unknown command and when I use
the web interface I am thrown by the first three questions (I think I
understand what follows!):
Name: what name? Is that just an I make it up name? DeskJet 960C? The Help
file says the helpful wizzard will walk you through this and you just answer
a few obvious questions! Not obvious to me.
Locatin: what location? and if this is the /dev/lp or lp0 or lp1 can I just
go and assign that?
Description I can guess is nothing important.

Thanks for any help and sorry if this was already covered. I would like to
get a printer hooked up to the rig.

Best Wishes!
Mike Olds www.buddhadust.org



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: cups and hpijs driver

[Michael Olds]

OK Donald, thanks for clearing up the obvious. I'm going to give the KDE
package way a shot (that was the way I had it set up on RedHat...I did roll
my own ppd and it was seeing it.

That's it for me tonight though.

Best Wishes!
Mike Olds www.buddhadust.org


-Original Message-
From: Donald R. Spoon [mailto:[EMAIL PROTECTED]]
Sent: Monday, December 16, 2002 7:13 PM
To: [EMAIL PROTECTED]
Subject: Re: cups and hpijs driver


Michael Olds wrote:
> I'm catching up to this one late, and am being thrown off by the very
start.
>
> I have a DeskJet 960C
>

Same printer here.  Works great with CUPS.

> I am pretty sure it's hooked up right as it was running under RedHat
> (meaning I have PnP turned off in the BIOS) and during one point in
playing
> around tonight I saw that something knew the printer was there.(don't ask,
> I've been trying to find it again and can't).
>
> I have installed the foomatic.bin and foomatic.db and cupsys and it's
> depends
> and have generated and saved in /usr/share/cups/model/HP-DeskJet`.ppd
>

If you are running KDE, install the kdelibs3-cups package and use the
KDE "Control Center"-->"System"-->"Printing Manager" to get to the setup
Wizzard.  Make sure you choose "CUPS" as your printing system.

I would also suggest the following additional packages be installed:
cupsys-client
cupsys-bsd
foomatic-db-gimpprint
hpijs

In addition, I have found the cupsys-driver-gimpprint drivers to be of
fairly high quality.  I use the "HP Deskjet 900 series" driver there.
If you want to "roll your own" PPD for the 960C, then go to
www.linuxprinting.org, look up your printer and / or the hpjis "driver"
and generate a specific PPD for it.  It offers a mildly better color
printing resolution than the gimpprint drivers and access to some extra
functions on the 960C, but it is slower here.

> But when I type lpadmin in the terminal I get unknown command and when I
use
> the web interface I am thrown by the first three questions (I think I
> understand what follows!):
> Name: what name? Is that just an I make it up name? DeskJet 960C?

Yep...whatever you want to call it!  It is nice to call it something
like "lp" so that other programs that look for "lp" will find it, but
this is not necessary.  Keep it short... reduces typing .  This
entry is MUST, since that is the "name" the system will use to identify
this specific printer.  It is not much of a concern with a single
printer, but on a network with several printers available for use it
helps distinguish them.

>The Help
> file says the helpful wizzard will walk you through this and you just
answer
> a few obvious questions! Not obvious to me.

The only "wizzard" I have found is in the KDE Printing Manager.  It is
quite helpful.  You can only get this for CUPS if you have the
kdelibs3-cups package installed.  IMHO, it is easier than the web interface.

> Locatin: what location? and if this is the /dev/lp or lp0 or lp1 can I
just
> go and assign that?

Location is where the printer is physically located.  This is mainly
useful for sharing a printer on a network.  You wouldn't want to print
something out on a printer across campus for example.  In my case the
"location" is "Basement" (of my house).  Not mandatory

> Description I can guess is nothing important.
>

Again, useful on a network, where you might want to tell a remote user
the capabilities and limitations (color, b&W, resolution, speed, etc).
Not mandatory


> Thanks for any help and sorry if this was already covered. I would like to
> get a printer hooked up to the rig.
>
> Best Wishes!
> Mike Olds www.buddhadust.org
>

Good Luck!  I struggled with this setup for quite a while, but once I
got it going, it works great.  Holler if you need more help...especially
if you want to put it on a network.  I found there were a few "gotchas"
in setting it up for a network, however setting it up for use on the
computer where it is attached is fairly straight-forward.

-Don Spoon-




--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: cups and hpijs driver

[Michael Olds]

OK,
Re: HPDeskJet 960c as local printer on Woody box

 I've used up my variables. I installed the KDE kdelips3-cups package, (and
I also have installed: cupsys; cupsys-client; cupsys-pstoraster; mpage;
foomatic-bin; foomatic-db; and hpijs -- I did not install cupsys-bsd as this
is for printing off the local linux box only; I did a custom PPD which looks
like it is installed correctly and is being used) and it knows about CUPS,
but when I start up the wizzard the first dialog box I get is all grayed out
except for "Class of Printers" and if I select that, the next one shows no
printers available.

I did not try Benedict's Cupsomatic Perl script under the thinking that this
was taken care of by my custom PPD and the functions are supplied by the
foomatic guys...yes?

In the CUPS http admin tool I appear to be able to set up the printer, and
it sees it at:
URI: parallel /dev/lp0
and it has Device: Parallel Port 1

Clicking on "Print Test Page" gets the printer rolling, and then it spits
out a blank page.

when i go to configure it it gives me: client-error-not-found
...first time out I went back and installed the CUPS client, but that didn't
change anything.

(I have deleted and re-added, and stopped and started the printer several
times).

I also took a look at the kernel and I have IEEE support, parallel printer
support, and USB support.

When I try and print a file from Nedit I get: Unable to Print sh:lpr command
not found (I also mentioned previously that when I try to enter the
administration tool via the command line lpadmin I get no such command)

Just to be sure I shut down and restarted a couple of times. Also generously
peppered the operation with /etc/init.d/cupsys restart

Just Guessing in the dark, I added the printer to the classes following the
obvious...ahum...choices in the wizzard. nada

Best Wishes!
Mike Olds www.buddhadust.org

How come after 3 rebuilds, when i run # uname -a my kernel is still listed
as #1  the date is the rebuild date, and the vmlinus version in /boot is
also dated correctly.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: RE: cups and hpijs driver

[Michael Olds]

Thanks again Benedict, well said,

I will revisit your previous suggestion and also reinstall cupsys.bsd.

The one further development is I think printer-specific. On the printer
itself I pressed the middle button and the printer put out what was a
beautifully printed test sheet...I believe an internal test sheet, but still
it shows the printer works.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, December 18, 2002 1:10 AM
To: [EMAIL PROTECTED]
Cc: Michael Olds
Subject: Re: RE: cups and hpijs driver



 Michael Olds <[EMAIL PROTECTED]> wrote:


>OK,
>Re: HPDeskJet 960c as local printer on Woody box
>
> I've used up my variables. I installed the KDE kdelips3-cups package, (and
>I also have installed: cupsys; cupsys-client; cupsys-pstoraster; mpage;
>foomatic-bin; foomatic-db; and hpijs -- I did not install cupsys-bsd as
this
>is for printing off the local linux box only;
Hm, i installed cupsys-bsd as well. Not sure if i need it though since it's
also a local printer in my setup

>I did a custom PPD which looks
>like it is installed correctly and is being used) and it knows about CUPS,
>but when I start up the wizzard the first dialog box I get is all grayed
out
>except for "Class of Printers" and if I select that, the next one shows no
>printers available.
I didn't have to bother with class of Printers. I went to linuxprinting.org,
made a ppd for my printer and saved it in de models directory.
Then i installed the cupsomatic perl script from the same site in the filter
directory and restarted cups

>
>I did not try Benedict's Cupsomatic Perl script under the thinking that
this
>was taken care of by my custom PPD and the functions are supplied by the
>foomatic guys...yes?
Foomatic is needed. I tested it once by uninstalling and see if stuff would
work but it doesn't. I'm not sure about the cupsomatic perl script either.
linuxprinting.org said to add it so i did.

>
>In the CUPS http admin tool I appear to be able to set up the printer, and
>it sees it at:
>URI: parallel /dev/lp0
>and it has Device: Parallel Port 1
>
Mine is usb: /dev/usb/lp0

>Clicking on "Print Test Page" gets the printer rolling, and then it spits
>out a blank page.
Does printing from a webpage work?

>when i go to configure it it gives me: client-error-not-found
>first time out I went back and installed the CUPS client, but that
didn't
>change anything.
I had that error too untill i installed all of the above.
I *think* that i got that error because i didn't install foomatic.bin

>
>(I have deleted and re-added, and stopped and started the printer several
>times).
>
>I also took a look at the kernel and I have IEEE support, parallel printer
>support, and USB support.
>
>When I try and print a file from Nedit I get: Unable to Print sh:lpr
command
>not found (I also mentioned previously that when I try to enter the
>administration tool via the command line lpadmin I get no such command)
Don't you need to have the cupsys-bsd stuff to manager printing via lpr?


>Just to be sure I shut down and restarted a couple of times. Also
generously
>peppered the operation with /etc/init.d/cupsys restart
>
>Just Guessing in the dark, I added the printer to the classes following the
>obvious...ahum...choices in the wizzard. nada
Like i said, i didn't need to add the printer to the classes.
>
>Best Wishes!
>Mike Olds www.buddhadust.org
>
>How come after 3 rebuilds, when i run # uname -a my kernel is still listed
>as #1  the date is the rebuild date, and the vmlinus version in /boot is
>also dated correctly.
>
>
>--
>To UNSUBSCRIBE, email to [EMAIL PROTECTED]
>with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]
>



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: cups and hpijs driver

[Michael Olds]

Re: setting up DeskJet 960c as local printer on Woody box

I have reviewed the cupsomatic perl script situation and this basically
duplicates the work being done by the custom PPD and the foomatic-bin and
foomatic-db (at least that's how I understand it). I have also installed
cupsys-bsd, gv, and qtcups per various sets of instructions. So now I have
the following installed:
cupsis-pstoraster
cupsys
cupsys-bsd
cupsys-bsd
cupsys-client
foomatic-bin
foomatic-db
gv
hpijs
mpage
qtcups

Still no joy.

I think the problem is occuring somewhere between the printer as hardware
which seems to be installed ok, and the various installed packages which
also seem to be set up ok. It's just that the two are not talking to each
other.

Except for the fact that this may indicate something else wrong with my
setup, I'm about to say "Why do I need this?" since the printer off my W2K
box works fine and I could probably set up this printer on an open USB port
and have it up and running in two minutes (the little test it did for itself
looked like a very nice job of printing, better than the multi-function
OfficeJetT65 I use on the W2K box.)

1. What would be the best way to get everything print related out of the
machine to start over? Uninstall all of the above? Anything else?

2. Can I still be missing something in my kernel?

3. How do you command-line send a job to the printer? lpr and pdq both give
me: "command not found". I looked in modutls aliases and inserted the lines:
alias /dev/printers  lp
alias /dev/lp0   lp
alias parport_lowlevel parport_pc
and rebooted, but that did nothing, so I removed them. My thinking is it is
something like this that is the problem: no communication between the
parallel port and the printing drivers.

4. I tried printing from the Web, no go, but that's no real test as Mozilla
doesn't do anything for me except see pages (I just installed the browser).

5. The error message off Nedit has changed, and is now: lpr: unable to print
file: server-error-service-unavailable


Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: cups and hpijs driver

[Michael Olds]

Donald,

Thank you for sticking with this! I havn't yet done anything on Linux that
hasn't been a major hassle (not complaining, I really like the brain
exercise and I like the end result which usually works well...I have my
webserver up and running without a problem now for a couple of
months...getting ready to set up mysql and php...anyway:)

> Using the parallel port.
OK...same here

> Do a "lsmod" as root and see if the following modules are loaded:
> paraport
> paraport_pc
> lp
> All three present and accounted for.
OK...drivers are present.

I see a message in the bootup messages indicating it "sees" the printer,
> I see the same in my messages. Names it by name.
> and also:
> lp0 using parport0


>OK...just a thought here.  Do you have your parallel port turned "on" in
the BIOS.  It shouldn't make a difference to Linux, but one never knows ;)
I Checked and it is "on".

> But... > > and when I try the KDE Printing Manager it shows up already
listed in the > "Local Port Selection" screen of the Wizzard. >>Right.. same
here.  Mine shows a "tree" of selections: Parallel Parallel Port #1
HEWLETT-PACKARD DESKJET 960C Do you high-light it (select it) so
"parallel:/dev/lp0" shows in the "URI" window at the bottom of the screen?
This is a miscommunication. I do not get this. I was quoting you. haha.

> I do not get as far as "Local Port Selection in the wizzard as it is >
demanding a "backend" In "Backend" > "Classes" are all greyed out except
"Class of Printers" which if I select, > brings me to "Class Composition"
which gives me no selections at all. End of > the line. >
>I get this as the very first screen after the KDE Wizzard starts up. All of
mine are available except the "Print to file" and "Serial Fax/Modem Printer"
selections, which are greyed-out.

I wonder where does the wizzard get this information? It must be a
configuration file somewher, no?

>>Do you have "CUPS" selected as the "Print System Currently Used" at the
bottom of the KDE Printing Manager?  The only other thing that might cause
this is that somehow your kdelibs3-cups didn't get installed
properly...Strange.
>try re-installing the kdelibs3-cups package.  What version of KDE are you
using?  I am using 2.2.2 from Woody/Testing.  I have NOT made the move to
the 3.X versions yet.

I am using 2.2.2 too.
I just uninstalled, shut down, restarted, installed ran /etc/init.d/cupsys
restart and opened control center>system and I get the same routine: backend
selections all grayed out, etc. Yes I have CUPS as the Print System
Currently Used.

>You ARE trying all this config as "root" ... right?
Yes

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: about web link

[Michael Olds]

  Can I put a http://www.htmlhelp.com/bbs/ to get
up to speed with html, this has nothing to do with apache or Debian



Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

cups and hpijs driver - follow up

[Michael Olds]

Attempting to set up HP DeskJet960c on Linux Woody box.

I have made progress which makes failure that much more frustrating at this
point.

First I got both the KDElibs3-cups utility and the CUPs administration tool
to recognize and even allow me to configure the printer by 1. removing all
the printer related packages I could find and deleting any related leftover
directories and files I could find. The only thing I left on (that I know
of) was the /usr/share/CUPS/models/ ppd. Then I reinstalled:

cupsys -pstoraster
cupsys
cupsys-bsd
cupsys-client
foomati-bin
foomatic-db
hpijs
mpage
kdelibs3-cups

Now the only problem is that although the printer does respond to being
given a test page to print (either from the KDE tool or from the CUPs admin
tool...or for that matter by attempting to print a file)...all that happens
is that a blank sheet of paper gets sent through the printer without any
printing being done.

Can anyone provide any insight as to what I am missing here?

Best Wishes!
Mike Olds www.buddhadust.org



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

PHP/MySQL problem

[Michael Olds]

OK I need a little help here, some insight into what may be going wrong.

Setting up on a Debian Woody with Apache; trying to install a PHP BB script.
I have checked in with the script folks but I need to work this from both
ends.

The PHP module is set up and working (I have uncommented the correct line in
LoadModule and the two AddType lines -- does there need to be anything like
a "script alias" for php? I still havn't figured out how scripts can be
processed in the htdocs folder safely) as is MySQL (I'm not saying nothing
can be wrong with the setup, only just that it is working as far as I can
tell). I can access the database from the box using the username and
password set up for the script. And I have run a couple of small php include
scripts and done the  thing and "mysql" was built in (but my book
says there should be a section on MySql and there was none...is there
something further that needs to be done to activate PHP/MySql interaction?).
I have double checked the spelling, and given the root directory path
properly.

What happens is I can GET the first page of the "Administration" setup
script for the BB and it appears to POST ok, and then all I get is a blank
(white) screen that tells me its done.

No error messages at all.

I checked my Norton Firewall and refers are allowed...but not recommended
while trying to set up PHP scripts...setting this up from my W2K box ftping
to the Linux box.

I uploaded the script files in ASCII and the permissions are correct.

I have checked the on-line documentation and have looked at a couple of
books but there is very little in any of them about how to set up (if that
is the problem), most are on how to use.

Any pointers, clues, hints, insights?

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: PHP/MySQL problem SOLVED

[Michael Olds]

This problem solved.

PHP was working fine, MySql was working fine, MySql support was built into
the PHP Apache module ok, but the PHP MySql module was not installed.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Michael Olds [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, December 24, 2002 2:24 PM
To: Debian-User
Subject: PHP/MySQL problem


OK I need a little help here, some insight into what may be going wrong.

Setting up on a Debian Woody with Apache; trying to install a PHP BB script.
I have checked in with the script folks but I need to work this from both
ends.

The PHP module is set up and working (I have uncommented the correct line in
LoadModule and the two AddType lines -- does there need to be anything like
a "script alias" for php? I still havn't figured out how scripts can be
processed in the htdocs folder safely) as is MySQL (I'm not saying nothing
can be wrong with the setup, only just that it is working as far as I can
tell). I can access the database from the box using the username and
password set up for the script. And I have run a couple of small php include
scripts and done the  thing and "mysql" was built in (but my book
says there should be a section on MySql and there was none...is there
something further that needs to be done to activate PHP/MySql interaction?).
I have double checked the spelling, and given the root directory path
properly.

What happens is I can GET the first page of the "Administration" setup
script for the BB and it appears to POST ok, and then all I get is a blank
(white) screen that tells me its done.

No error messages at all.

I checked my Norton Firewall and refers are allowed...but not recommended
while trying to set up PHP scripts...setting this up from my W2K box ftping
to the Linux box.

I uploaded the script files in ASCII and the permissions are correct.

I have checked the on-line documentation and have looked at a couple of
books but there is very little in any of them about how to set up (if that
is the problem), most are on how to use.

Any pointers, clues, hints, insights?

Best Wishes!
Mike Olds www.buddhadust.org


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: PHP/MySQL problem SOLVED

[Michael Olds]

Three teeth-grinding days!

-Original Message-
From: Steve [mailto:[EMAIL PROTECTED]]
Sent: Thursday, December 26, 2002 8:29 AM
To: [EMAIL PROTECTED]
Subject: Re: PHP/MySQL problem SOLVED


On 12/25/2002 6:54 PM, Michael Olds wrote:
> This problem solved.
>
> PHP was working fine, MySql was working fine, MySql support was built into
> the PHP Apache module ok, but the PHP MySql module was not installed.

That gets me every time. ;)



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Script Alias Directive in Apache

[Michael Olds]

Hello,
Can someone tell me how to configure Appache to allow two separate users to
have their own cgi-bins?

I have read two books, the Apache docs and done some googling around and I
can't find an example of how to set up two different virtual hosts with
separate cgi-bins.

Can anyone give me the layout?

I have tried it:
Like this (the comment distinctly uses the plural: Aliases) (or does this
get chained? X = Y Y=Z?):

ScriptAlias /cgi-bin/ "/www/usr/public_html/cgi-local/"
ScriptAlias /cgi-bin/ "/www/usr2/public_html/cgi-local/"

With:


AllowOverride None
Options ExecCGI
Order allow,deny
Allow from all



   AllowOverride None
   Options ExecCGI
   Order allow,deny
   Allow from all


And under the Virtual Host directives:


   UseCanonicalName off
   ServerName www.usr.com
   DocumentRoot "/www/usr2/public_html"
   ScriptAlias /cgi-bin/ "/www/usr/public_html/cgi-local"
   ServerAdmin usr@his



   UseCanonicalName off
   ServerName www.usr2.com
   DocumentRoot "/www/usr2/public_html"
   ScriptAlias /cgi-bin/ "/www/usr2/public_html/cgi-local"
   ServerAdmin usr@his


And both ways work but only for "usr"; anything asking for usr2/cgi-local
gets sent to usr/cgi-local

Best Wishes!
Mike Olds www.buddhadust.org



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: KVM locks tty

[Michael Olds]

I have this switch also. The workaround that I have used is to be sure that
all operations on the one computer are finished before going to the next,
and I use the manual button, not the keyboard.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Lance Hoffmeyer [mailto:[EMAIL PROTECTED]]
Sent: Thursday, December 26, 2002 2:46 PM
To: [EMAIL PROTECTED]
Subject: KVM locks tty


I have a kvm switch and every time I am in a console and switch
back and forth it locks up the console.  I have to be on tty8
or in X for the switch in work.  In X, a switch to another computer
and back will screw up the mouse and I have to go to a tty and back
into X for the mouse to work properly.   This is a Belkin box with
no software.  Anything I can try so that my tty's don't lock up?

Lance


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Script Alias Directive in Apache

[Michael Olds]

Thank you for this response.

> ScriptAlias /cgi-bin/ "/www/usr/public_html/cgi-local/"
> ScriptAlias /cgi-bin/ "/www/usr2/public_html/cgi-local/"

This won't work. ScriptAlias works just like Alias, doing this tries to
assign the same URI to two different directories.

Ok, these are commented out.

<-->


> 
> AllowOverride None
> Options ExecCGI
> Order allow,deny
> Allow from all
> 

If the directory is ScriptAlias'd I think you don't actually need to
specify ExecCGI, it's implicit. Not that it does any harm.

I left these as is, one for each...as long as they don't do any harm, I feel
more comfortable with them in there.
<-->
This is what I have now:

> 
>UseCanonicalName off
>ServerName www.usr.com
>DocumentRoot "/www/usr2/public_html"
>ScriptAlias /cgi-bin/ "/www/usr/public_html/cgi-local"
>ServerAdmin usr@his
> 
>
> 
>UseCanonicalName off
>ServerName www.usr2.com
>DocumentRoot "/www/usr2/public_html"
>ScriptAlias /cgi-bin/ "/www/usr2/public_html/cgi-local"
>ServerAdmin usr@his
> 

>If you have only these without the ScriptAlias lines and the your
>virtual hosts are set up correctly, it should work.

The two work fine for serving html pages correctly, and # one serves cgi
scripts fine by itself.

> If you have
>/cgi-bin/ set up in the top-level server configuration (it is by
>default) it will get in the way, so make sure all other ScriptAlias
>lines are commented out.

The lines in the  are the only ScriptAlias lines in the
file, and all (2, both me) users are set up this way.


>If you want to have a global CGI directory in
>addition to the private ones you'll have to use different paths (like
>cgi-bin for one and cgi-local or cgi-shared for the other).

No need.

>Alternatively, if you just want individual user directories (~user) to
>have CGIs instead of virtual hosts you can add ExecCGI to the options
>for the user directories and use "AddHandler cgi-script cgi" to treat
>any file ending in .cgi as a CGI.

I don't want to go this way.

<-->

OK before I send this off saying that this is the way I am set up and it
doesn't work. I will double check everything...hang on...

Nope. I get An Internal Server Error on the Browser side (trying to run a
simple script) and in the log it says "no such file or directory) (the file
doesn't exist in the first users cgi-bin).

I also know that it's using the wrong one because I have the same script in
each cgi-bin with only a different header, and when I call it up via the url
of the second user, I get the first user's script.

>From what I have read also, what I have should be working.

Could I use: Script Alias /www.firstuser.com/cgi-bin/
"/www/usr/public_html/cgi-local"
Script Alias /www.seconduser.com/cgi-bin/
"/www/usr2/public_html/cgi-local"

I guess what I don't understand is what this means. I hear this as Apache
has a thing called /cgi-bin/ that allows scripts to be executed. If user one
wants to execute scripts, he needs to have a directory of his pretend to be
that directory, so I would say that the fake name was
"/www/usr/public_html/cgi-local" and the real name was /cgi-bin/ and if I
were the little appache sitting in there sorting messages, I would be able
to recognize: This call for executing a script coming in as
"/www/usr/public_html/cgi-local" is ok because here it says x = cgi-bin, and
the same for the other.

What it looks like is actually happening is that user 1s fake cgi directory
has become the one real cgi-directoryi.e. the sorter sees cgi-bin and
throws it to #1 without any thought about the rest of the address. I would
write it: ScriptAlias /www/usr/public_html/cgi-local/ "cgi-bin"
etc. I can't seem to get my mind to turn it around.

Best Wishes!
Mike Olds www.buddhadust.org





-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Secure CGI Implementation

[Michael Olds]

If the cgi-bin is placed outside the DocumentRoot and a script creates an
html page to be served to the public, how does this work?

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Secure CGI Implementation

[Michael Olds]

Q: If the cgi-bin is placed outside the DocumentRoot and a script creates an
html page to be served to the public, how does this work?

A: Just fine.  You can map your web space any way you like.

Actually, Bill, 'Just Fine is just fine if it works...I havn't quite got it
up and running yet.' I am trying to ask short 'n sweet questions to the
point (not my natural style, which usually includes numerous subordinate
clauses, etc...you understand...)...I wouldn't object to an explanation of
the process.

Best Wishes!
Mike Olds www.buddhadust.org



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Secure CGI Implementation

[Michael Olds]

This is what I would like to set up because I would like to offer server
space to one user aside from myself that I trust, and who will need to make
modifications to his scripts via SFTP (using SSH2); but I would like to be
protected in the event of his user name and password being discovered, and
in the event of him making a mistake in his setups.

Directory Structure, Users and Groups, and Permissions

apache = apache deamon
admingroup = {me}

/www
   /users  {Server Document Root}   {apache:rwx admingrp:rw: w:--}

  /usr1/logs {apache:rwx usr1:rw: w:--}
  /usr1/public_html   {VirtualHost1 DocumentRoot} {usr1:rwx usr1:rw:
w:--}
 /directory/*.htm{usr1:rwx usr1:rw: w:--}
  /usr1/cgi-bin {usr1:rwx usr1:rw: w:--} 
 /directory  {usr1:rwx usr1:rw: w:--}
/*.cgi   {usr1:rwx usr1:rw: w:--}
 /directory

  /usr2/logs
  /usr2/public_html   {VirtualHost2 DocumentRoot}
  /usr2/cgi-bin

httpd.conf

   Global Server Settings

   DocumentRoot "/www/users"

   
   Options ExecCGI
   AllowOverride All
   Order allow,deny
   Allow from all
   

   No Script Alias or  containers.

VirtualHost containers

NameVirtualHost 000etc


   UseCanonicalName off
   ServerName www.domain1.com
   ServerAlias domain1.com *.domain1.com domain1
   DocumentRoot "/www/users/usr1/public_html"
   ServerAdmin [EMAIL PROTECTED]
   ScriptAlias /cgi-bin/ "/www/users/usr1/cgi-bin/"
   ErrorLog  /www/users/usr1/log/www.domain1.com-error.log
   CustomLog /www/users/usr1/log/www.domain1.com-access.log combined
   
   AllowOverride None
   Options ExecCGI
   Order allow,deny
   Allow from all
   



   UseCanonicalName off
   ServerName www.domain2.com
   ServerAlias domain2.com *.domain2.com domain2
   DocumentRoot "/www/users/usr2/public_html"
   ServerAdmin [EMAIL PROTECTED]
   ScriptAlias /cgi-bin/ "/www/users/usr2/cgi-bin/"
   ErrorLog /www/users/usr2/log/www.domain2.com-error.log
   CustomLog /www/users/usr2/log/www.domain2.com-access.log combined
   
   AllowOverride None
   Options ExecCGI
   Order allow,deny
   Allow from all
   


If necessary, this way I can change one user's cgi-bin to "cgi-local", or
"cgi-1" although this still does not make good sense to me and does not fit
in with what I see or what is given as examples in the tutorials out there
(I would like to know how my host for BuddhaDust has been able to give me
"cgi-local" under my document root with me as owner and user...I will ask
next chance I get).

This nearly puts me back at my original configuration with an additional
layer /users/ above the Main Server Document Root and it places the user's
cgi-bin beside, not under their VirtualHost DocumentRoot. If this does not
look right I would VERY MUCH appreciate anyone who would take the time to
change the setups I have here to actually show me by example what should
work. Theoretical explanations just are not getting through to me.

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Secure CGI Implementation (PARTIAL SOLUTION)

[Michael Olds]

Well you may not have know there was a problem. The dialog was why can't I
get CGI working with two VirtualHosts. Then I went down the long trail of
trying to set up CGI outside of the DocumentRoot. I moved everything down a
notch and then also down a notch within each user's space. Absolutely
nothing I did thereafter brought the slightest hint that CGI would ever work
again. I cannot state with absolute certainty that I tried every single
combination of user, group, and permissions and every variation on
cofiguring the apache httpd.conf, or read every single document describing
apache configuration for VirtualHosts but I believe I tried most of them,
and read many. ZipNada.

Bill Moseley's run through a number of possible layouts proved to have at
least gotten me back to square one (one CGI bin working). He mentioned
SuEXEC...and gave me a link to the Apache doc on the subject. I have SuEXEC
installed, and apparently working fine as it is supposed to do. Except that
it does not, for security and efficiency reasons, endquote, work below the
first level of the Server's DocumentRoot, or the First Level under the
VirtualHost DocumentRoot.

So I moved everything back up a notch.

Tomorrow I see if I can get the second CGI bin to work now that I have a
small understanding the role of SuEXEC.

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Apache-ssl .htaccess files

[Michael Olds]

1. Create ".htaccess" file in the directory to be protected.

AuthUserFile /etc/apache/htusers
AuthName "This Directory Requires Authentication by A Valid User"
AuthType Basic
require valid-user
order deny,allow
allow from all

CHMOD: 755

Note: This file can contain adjustments for options to an Apache Directive
or certain Apache Directives themselves.

2. Create etc/apache/htusers

# cd /etc/apache
# htpasswd -c /etc/apache/htusers/htusers_id username1
Password
Password

CHMOD: 644



3. Create additional users.
After setting the first user:

htpasswd /etc/apache/htusers/htusers NewUsersName
Password
Password

Note: More than one .htaccess file can accesess this file.



4. Create sub-groups of users


5. Delete users
delete the whole line to remove a user.




-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Desktop Performance Issue (FIXED)

[Michael Olds]

How did you determine the appropriate values for all of the
other settings you made besides the DMA setting? (Or, where
did you come across this information for these settings 
(besides the man page obviously))?

On Fri, Dec 27, 2002 at 11:17:31PM -0700, Michael wrote:
> You guys hooked me up. I entered the following hdparm command:
> 
> hdparm -X34 -d1 -u1 -m16 -c3 /dev/hda
> 
> Not only are problems resolved but system in general runs much quicker
> and is more responsive. Hdparm is now running at startup!! Thanks all for
> you help with this one. Wish I'd of asked a long time ago.
> 
> Michael

I don't know if this is what he used, but you could start here:


http://linux.oreillynet.com/pub/a/linux/2000/06/29/hdparm.html


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: kernel recompile

[Michael Olds]

||In the continuing saga of my kernel recompile, I did this:
||
||  make xconfig
||  make-kpkg clean
||  make-kpkg kernel_image
||  dpkg -i kernel-image-2.4.18_10.00.Custom_i386.deb
||
||Everything seemed to go well. There were no error messages, and I found
||myself with a newly compiled kernel named vmlinuz-2.4.18 in my /boot
||directory. But on attempting to boot, the kernel would instantly crash
||and reboot again, and again, etc, ad nauseum. I could only recover by
||editing GRUB to point back to the old kernel.
||
||Can't see what I did wrong, other than possibly that my settings somehow
||conflicted. I actually changed very little - I used the stock kernel,
||and just added support for sound, parport, scsi, and my ethernet card.
||
||Has anyone else had this experience? Can I just use my old kernel and
||somehow get these other things to work by loading modules?

I am guessing from the "I could only recover by editing GRUB" that you did
not rename /lib/modules/version number so that a new modules/version number
could be built by the new kernel. Repeat the process doing this first.

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

SuEXEC and CGI to two VirtualHosts

[Michael Olds]

I believe I have traced my problem in setting up CGI bins on two
VirtualHosts to the fact that SuEXEC is incorrectly configured (it is
enabled and I get "suexec: enabled; valid wrapper /usr/lib/apache/suexec".)
The error I get is: "cannot get docroot information /var/www" (the Debian
Apache default docroot).

I do not know how I got SuEXEC on there in the first place!
Does this come as part of the Debian package?

I could disable suEXEC, but in fact it does look like the tool I need (I
need for users to be able to load, unload, and modify cgi-scripts via sftp).

My problem is not knowing how I can reconfigure SuEXEC. The documentation I
have read says that it needs to be recompiled with apache. But I did not
compile apache from sources to begin with.

Hunt's book says: The suEXEC tool is very easy to set up using the APACI
installation script. Again, this is a tool for use in compiling from source.

Before I uninstall Apache and reinstall the package just to see if it (or
some other configuration tool) is included...does anyone know? Or know a
better way?

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: SuEXEC and CGI to two VirtualHosts

[Michael Olds]

Sorry Bill! Sent off-list by accident again...appipollylogies!

Best Wishes!
Mike Olds www.buddhadust.org


||> I believe I have traced my problem in setting up CGI bins on two
||> VirtualHosts to the fact that SuEXEC is incorrectly configured (it is
||> enabled and I get "suexec: enabled; valid wrapper
||/usr/lib/apache/suexec".)
||> The error I get is: "cannot get docroot information /var/www"
||(the Debian
||> Apache default docroot).
||
||That's typically either a permsissions problem or can chdir to a
||directory.

I do not know what "or can chdir to a directory" means. ?It cannot change
directories?

|| It might be that you have group write perms on a directory.
No. Double checked.

||Look at the suexec docs at what it goes through to before it will run a
||user's script.  That might give you ideas where to look.
||
||http://httpd.apache.org/docs/suexec.html  Look at "suEXEC Security Model".
||
||
||Another very useful trick is to run httpd -X (single process mode) and
||then use strace -- although I'm not sure if strace will follow the forks.
||You look for a failed system call and that's noramlly the problem.

The above is beyond my understanding.

|| But it might be easier to try and follow the docs carefully.

This I am trying to do.

I have read the suggested documentation a couple few times now, including
the monstrous routine it goes through to get a connection going...! And I
have it open in front of me now and am running down the list.

1. Am I correct in understanding that I can have one account set up (to test
it) that should work under suEXEC while another is not and it should work on
the test account?

if so:

2. How deep does the "no group write thing need to go?" (in other words,
does it apply above the cgi-bin (Well I tried changing it to Group apache
(the apache group) r -- enter, above /public_html with no change)

This is the Virtual Host setup:


   ServerName name.of.host
   DocumentRoot "/www/user2/public_html"
   User name2 (a test name belonging to no other group than name2 Group,
with no special privelages at all)
   Group name2
   ScriptAlias /cgi-bin-2/ "/www/user2/public_html/cgi-bin-2"
   


This is the Directory Structure:

Server DocumentRoot /www

/www/user2/public_html/cgi-bin-2/directory/*.cgi

for *.cgi (simple test script that works in the working cgi dir.)
   User name2 r r x
   Group name2 r - x
   World --

For directory
   User name2 r r x
   Group name2 r - x
   World --

For cgi-bin-2
   User name2 r r x
   Group name2 r - x
   World --

Calling the script: I get "Forbidden"
and the suEXEC error is as above:
emerg: cannot get docroot information (/var/www)
info: (target actual) uid: (user2/user2) gid (user2/user2): cmd *.cgi

The VirtualHost works fine just serving html docs.

This looks to me like suEXEC is looking in /var/www which is the apache
default DocumentRoot, but is not my DocumentRoot.


Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: SuEXEC and CGI to two VirtualHosts

[Michael Olds]

Bill,

Thanks again for your assistance and ideas on this. The more I get into this
the more confusing it becomes!

If I understand you correctly, to use SuEXEC I am going to have to either 1.
completely change the way I wanted to set up my web directory (ies), or 2.
figure out how to configure SuEXEC prior to compiling Apache and compile it
from source myself. How can Debian leave us in such a fix?...where it is
clear that we can set up the Apache document root any way we want and it is
absolutely silent on the effects changes will have on using SuEXEC? (I have
looked through the docs installed with Apache and there is virtually (ahum)
nothing on SuEXEC.) If they thought it worth while to include in the build,
then it should have been worth some documentation...and I mean more than
"set it up according to the defaults and rules and it might work." (Even
that, put in terms of a warning, would have saved me a few weeks of
aggrivation for no good whatsoever.)

For the record, in order to demonstrate to myself that I did not myself
insert some hairbrained configuration data in the SuEXEC setup, I
uninstalled and re-installed Apache. There is no indication at all during
setup that SuEXEC is being installed (and it is)...the curious thing is that
the re-install picked up much (not all) of my old configuration (before
deleting it...thank goodness I had a backup)(but not the new document
root -- why doesn't the installation include the ability to specify the
document root in the case where SuEXEC is being installed automatically,
since SuEXEC seems to be so tightly bound to the installation
configuration)? The error message I am now getting from SuEXEC log is: error
(a different level of error) command not in docroot.

Time for some more thought.

Sidebar: I contrast this to a PHP/MySQL setup I also have where there is no
grief like this at all!

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: SuEXEC and CGI to two VirtualHosts

[Michael Olds]

Bill,

Again thanks.

>Mike, I think you have to admit your troubles are a result of your
>inexperience with this setup.

Absolutely, and I see your point about the number of other modules (but do
any of them have this particular sort of inability to work if the default
setup isn't used? SuEXEC seems unique in this regard). And in addition to
inexperience, I have an inborn stubborn trait -- I want my directories where
I want them! haha. (This is a selling point of UNIX systems, that is that
the directory structure is "logical"...meaning you can order it in a way
that makes sense to you.)

Best Wishes!
Mike Olds www.buddhadust.org



>-Original Message-
>From: Bill Moseley [mailto:[EMAIL PROTECTED]]
>Sent: Saturday, January 04, 2003 9:05 AM
>To: Michael Olds
>Cc: Debian-User
>Subject: RE: SuEXEC and CGI to two VirtualHosts
>
>
>On Sat, 4 Jan 2003, Michael Olds wrote:
>
>> If I understand you correctly, to use SuEXEC I am going to have
>to either 1.
>> completely change the way I wanted to set up my web directory
>(ies), or 2.
>> figure out how to configure SuEXEC prior to compiling Apache and
>compile it
>> from source myself. How can Debian leave us in such a fix?
>
>How can the debian maintainer know where you want install things?
>
>The SuEXEC docs say:
>
> suEXEC Points Of Interest
> Hierarchy limitations
>
> For security and efficiency reasons, all suexec requests must remain
> within either a top-level document root for virtual host requests, or
> one top-level personal document root for userdir requests. For example,
> if you have four VirtualHosts configured, you would need to structure all
> of your VHosts' document roots off of one main Apache document hierarchy
> to take advantage of suEXEC for VirtualHosts. (Example forthcoming.)
>
>It's not the Debian package that's causing that restriction.
>
>> ...where it is
>> clear that we can set up the Apache document root any way we
>want and it is
>> absolutely silent on the effects changes will have on using
>SuEXEC? (I have
>> looked through the docs installed with Apache and there is
>virtually (ahum)
>> nothing on SuEXEC.)
>
>That's true, but look at how many module options there are in the default
>httpd.conf -- there's no way to list all the details.  Think about
>mod_perl's potential complex issues.  SuEXEC is typically used by ISPs and
>other experienced users -- and is also typically used for ~user type
>directories which is how the debian package has it.
>
>You have to keep in mind that the debian package is basically the Apache
>defaults, plus setup in a standard way to work for a wide group of users.
>Most people just use the default config and it works great.  Once you
>start changing things without have a good knowledge of Apache then you are
>in for a frustrating time.  It's not easy until you do spend those
>frustrating weeks.
>
>> For the record, in order to demonstrate to myself that I did not myself
>> insert some hairbrained configuration data in the SuEXEC setup, I
>> uninstalled and re-installed Apache. There is no indication at all during
>> setup that SuEXEC is being installed (and it is)
>
>It's just another Apache module.  Did it tell you that mod_cgi was being
>installed?  It's a package, and contains a lot of modules.
>
>>  why doesn't the installation include the ability to specify the
>> document root in the case where SuEXEC is being installed automatically,
>> since SuEXEC seems to be so tightly bound to the installation
>> configuration)?
>
>Because it's not a configuration parameter -- it's part of the suexec
>wrapper program and it's compiled in.  When you install a Debian package
>you are getting a binary verison.  It's a package deal.
>
>
>> The error message I am now getting from SuEXEC log is: error
>> (a different level of error) command not in docroot.
>
>Google.  That means you are trying to run a suexec script in a place
>that's not in docroot specified at suexec compile time.
>
>--suexec-docroot=DIR
>Define as the DocumentRoot set for Apache. This will be the only hierarchy
>(aside from UserDirs) that can be used for suEXEC behavior. The default
>directory is the --datadir value with the suffix "/htdocs", e.g. if you
>configure with "--datadir=/home/apache" the directory
>"/home/apache/htdocs" is used as document root for the suEXEC wrapper.
>
>I tried to explain that in my last message.
>
>> Time for some more thought.
>
>I posted a working httpd.conf.  Why not just move your directories?
>
>
>Or you can build Apache from source.  That

RE: Secure CGI Implementation/SuEXEC

[Michael Olds]

Bill, et al

Just a note to say thank you and that I appreciate that you went to extra
lengths to help me out in this. I have considered the options and have
decided that I am going to go with a single CGI bin with one administrator
(me); suEXEC disabled.

I did attempt setting up suEXEC in both /home/user/public_html and /var/www/
but it was late and it was giving me grief and I decided to take a hint. I
may give this a shot again when I am more comfortable with Linux...meanwhile
this is an option where I do not need to worry about user's passwords
falling into the wrong hands.

Best Wishes!
Mike Olds www.buddhadust.org



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Setting-up CDRW

[Michael Olds]

Oops! I see I resonded only to Bob:

Hello, and thank you all for these clues. This is just to let you know that
I can't get to this until later this afternoon. I will report back with
results.

-Original Message-
From: Bob Underwood [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, September 17, 2002 5:54 AM
To: [EMAIL PROTECTED]
Subject: Re: Setting-up CDRW


I just did this again last evening on a new machine and it was loads
easier than i remembered it being.

I'm using kernel 2.4.19-k7 with both an ide cdrom and an ide-scsi
cdrw.

i added this line to /etc/fstab:

/dev/scd0   /cdrw   iso9660 defaults,ro,user,noauto 0   0

my /etc/modules includes this:

options ide-cd ignore=hdd
ide-scsi

hth

bob



On Tuesday 17 September 2002 07:50, Burkhard Ritter wrote:
> hello.
>
> i suppose that your scsi/emulation stuff is not set up correctly.
> have you really loaded all neccessary modules (have a look again at
> the cdwriting howto section 2.1.3; ->
> ide-cd,ide-scsi,scsi_mod,sr_mod,sg)? if that is the case cdrecord
> -scanbus should definitely work (it has nothing to do with your
> fstab stuff).
>
> On Mon, 16 Sep 2002, Michael Olds wrote:
> > Requesting Help Setting-up CDRW
> > ---
> > I had this same hardware setup running under RH7.3
> >
> > Intel 845 P4 1.6G; 768MbRAM
> > Debian: 2.4.18-bf2.4
> > the SCSI emulation module was installed during the initial setup,
> > and lsmod returned: ide-scsi used 0  (Does the "used 0" mean
> > anything?) ---
> > /etc/lilo.conf
> > edited to insert: append="hdc=ide-scsi"  on line 104 above
> > #Boot up Linux by default.
> > #
> > default=Linux
> > ran lilo -v with no errors reported
> > ---
> >
> > /etc/fstab
> > edited to insert:
> > /dev/sc0/mnt/cdrw   iso9660  ro,user,noauto0 0
> > (also tried /dev/scd0 and for mount point: /cdrw having created a
> > directory there)
> >
> > Question:
> > /dev/cdrom  /cdrom  iso9660  ro,user,noauto0 0
> > remains in /etc/fstab. Is this correct? Or should this be removed
> > or linked to
> > /dev/sco...and if the latter, is this the right way:
> > dev/cdrom   /cdrw
> >
> > ---
> > When I run cdrecord I get the following message:
> > "Unable ... could this be the IDE device where you in fact use
> > ide-scsi so that sr0 or sda or so is needed?"
> >
> > My answer is yes, but it does not inform the novice user as to
> > how to get there.
> >
> > ---
> >
> > When I run xcdroast I get: "Failed to scan the SCSI-bus...you
> > have to install the
> > SCSI emulation first, Please see the CD -Writing HOW TO how to do
> > that. Well I did that and I thought that was what I had done
> > above.
> >
> > ---
> > When I run cdrecord -scanbus I get:
> > No such file or directory.
> > (Thinking that this might be something like setting up a new
> > partition, I set up a
> > directory under / named  "cdrw"
> > but that did not solve the problem.
> >
> > ---
> > Variations on the theme are rebooting after every change
> > --
> >
> > I have spent several days on this and have read the relevant
> > sections in O'R. Running Linux; Hunger, D.G.L. Bible, and the
> > CDRW HOW TO, as well as having searched Google Linux and read
> > half a dozen forums and list archives.
> >
> > There is a great deal of contradictory information out there,
> > even cocerning the setup for Debian. Most of the information is
> > also very old. How is it possible that this is not something that
> > is set up to at least be an option at initial setup on Linux?
> > What individual today setting up Linux does not have a CDRW?
> >
> >
> > Best Wishes!
> > Mike Olds www.buddhadust.org
> >
> >
> > --
> > To UNSUBSCRIBE, email to [EMAIL PROTECTED]
> > with a subject of "unsubscribe". Trouble? Contact
> > [EMAIL PROTECTED]


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: latest msttcorefonts broken?

[Michael Olds]

I just checked MS yesterday on this and they no longer offer this download.
Doing the upgrade may have deleted or renamed your old font package.
In any case this needs to be fixed in the packages that depend on it.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of DvB
Sent: Tuesday, September 17, 2002 3:24 PM
To: Debian-User
Subject: latest msttcorefonts broken?


I ran an apt-get upgrade a while back, which upgraded my msttcorefonts
package. Ever since then mozilla, which I had configured to use the
monotype-arial-iso8859-1 fontset from that package keeps printing the
following error message over and over (but doesn't appear to be causing
any other problems), even after I reset the fonts to the default values:

The font
"-*-*-medium-r-*-*-16-*-*-*-*-*-*-*,-*-*-*-r-*-*-16-*-*-*-*-*-*-*,-*-*-*-*-*
-*-16-*-*-*-*-*-*-*"
does not support all the required character sets for the current locale
"C"
  (Missing character set "ISO8859-1")


Any ideas as to how I might fix this?


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Setting-up CDRW

[Michael Olds]

-Original Message-
From: Christoph Claus [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, September 17, 2002 12:16 AM
To: Michael Olds
Cc: [EMAIL PROTECTED]
Subject: Re: Setting-up CDRW

<--X-->
Question:
/dev/cdrom  /cdrom  iso9660  ro,user,noauto0 0
remains in /etc/fstab. Is this correct?
Do you have an extra cd-rom? Then this is correct.

No, I have only the one CD Drive.
<--X-->
When I run cdrecord I get the following message:
"Unable ... could this be the IDE device where you in fact use
ide-scsi so that sr0 or sda or so is needed?"

What did you give cdrecord as dev= argument? It needs to be the
(pseudo)scsi-device with bus,target,lun e.g. dev=0,0,0

Apologies: this was not cdrecord, this was kreateCD, which uses CD record.
<--X-->
When I run cdrecord -scanbus I get:
No such file or directory.
Thinking that this might be something like setting up a new partition, I
set up a
directory under / named  "cdrw"
but that did not solve the problem.
<--X-->
Did you really load all necessary modules? Can you post the ide- and
scsi-section of your /boot/config-2.4.18? Are the appropriate entries
in /dev present? Especially /dev/sg*? The output looks like missing
/dev entries.

See above, I apparently do not have all the modules, but they are not
available from modconf
<--X-->


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Help needed setting up CDRW

[Michael Olds]

use one is now not able to
make changes there directly and I did not understand the directions for
using Modutls. I did not see anywhere where it looked like I should add the
suggested "options" lines:

options ide=cd ignore=hdc
alias scd0 sr_mod
pre-install sg modprobe ide-scsi
pre-install sr_mod modprobe ide-scsi
pre-install ide-scsi modprobe ide-cd
<--X-->
Direct answers below:

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Christoph Claus [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, September 17, 2002 12:16 AM
To: Michael Olds
Cc: [EMAIL PROTECTED]
Subject: Re: Setting-up CDRW

<--X-->
Question:
/dev/cdrom  /cdrom  iso9660  ro,user,noauto0 0
remains in /etc/fstab. Is this correct?
Do you have an extra cd-rom? Then this is correct.

No, I have only the one CD Drive.
<--X-->
When I run cdrecord I get the following message:
"Unable ... could this be the IDE device where you in fact use
ide-scsi so that sr0 or sda or so is needed?"

What did you give cdrecord as dev= argument? It needs to be the
(pseudo)scsi-device with bus,target,lun e.g. dev=0,0,0

Apologies: this was not cdrecord, this was kreateCD, which uses CD record.
<--X-->
When I run cdrecord -scanbus I get:
No such file or directory.
Thinking that this might be something like setting up a new partition, I
set up a
directory under / named  "cdrw"
but that did not solve the problem.
<--X-->
Did you really load all necessary modules? Can you post the ide- and
scsi-section of your /boot/config-2.4.18? Are the appropriate entries
in /dev present? Especially /dev/sg*? The output looks like missing
/dev entries.

See above, I apparently do not have all the modules, but they are not
available from modconf
<--X-->


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Help needed setting up CDRW

[Michael Olds]

Bob,

Thank you for this response. I am still very new at Linux and have hesitated
doing a kernel rebuild, but that is on the list...I just want to get a way
to do a backup so that when something goes wrong I won't have to start from
scratch ... I'm on about the 15th install and it's getting tiresome!

I am using Synaptic as my package management tool, and I think I can handle
a kernel update, but the question is, has been, I am using ext3 as my
filesystem, do the other kernels besides "bf" support this file system type?

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Bob Underwood [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, September 18, 2002 6:12 AM
To: Debian-User@Lists. Debian. Org
Subject: Re: Help needed setting up CDRW


On Wednesday 18 September 2002 08:53, Michael Olds wrote:
> This boils down to two problems: if a module is missing, where can
> I get it? I did a Google search and there are a zillion references
> to the missing modules but no mention of where they can be found.
>
> And Second, what is the procedure for editing /etc/modules.conf
> using the new modutls tool?
>
> Here is the previous message...and, again thanks for many suggested
> avenues of inquiry. However I am at an impass. I will try to update
> the situation completely here while answering the questions of
> those who asked for more information.

IIRC, the "bf" kernels do not include the modules you're missing.  I'm
using kernel-image-2.4.18-k7 (from woody) which does include them.
It was fairly painless for me to install the new kernel using
apt-get.

If you wish, you can also build your own kernel to include the missing
modules.

bob



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Help needed setting up CDRW

[Michael Olds]

Bob et al,

OK I upgraded the kernel to 2.4.18-686 which went smoothly except that this
kernel uses something called initrd, which I opted out of the first go round
for some reason I can't remember. Needed to add a line to /etc/lilo.conf
'initrd=/initrd.img'
and one to /etc/kernel-img.conf which does not exist (didn't cause a
problem, but I apparently will get some kind of warning message about using
initrd every time I upgrade the kernel)

I then used xcdroast to select a directory and copy it to a cd which is what
I wanted the cd for in the first place (backup, and hopefully restore). This
also seemed to go smoothly.
However to read the CDROM I needed to make a change to /dev/cdrom:

cd /dev && rm cdrom && ln -s scd0 cdrom

which changed it to a link to the scsi device.

I deleted the line for /dev/scd0 etc in fstab without any apparent problems;
I left the line in /etc/lilo.conf append="scd=scsi-ide"; I deleted the CDRW
directory under / that I had created as an experiment; all the required mods
now appear to be present and accounted for

cdrecord -scanbus now gives me:

scsibus0
0,0,0, 0)SONY 'CDRW CRX185e1' 'XYS2' Removable CD-Rom
which is what I think it is supposed to show.

The only problem that remains is that trying to make a CD with kreatecd it
just sits there, but if this does not look like some kind of hidden problem,
I think I will just remove this program and use xcdroast.


Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: ide-scsi

[Michael Olds]

Alan, try

rmmod [option] modules


System administration command. Unload a module or list of modules from the
kernel. This command is successful only if the specified modules are not in
use and no other modules are dependent on them.

Can you tell us what message you are getting that makes you think this is
loaded or interfering with what you want to do?

Best Wishes!
Mike Olds www.buddhadust.org


-Original Message-
From: Alan Chandler [mailto:[EMAIL PROTECTED]]
Sent: Friday, September 20, 2002 12:54 PM
To: [EMAIL PROTECTED]
Subject: Re: ide-scsi


On Friday 20 September 2002 7:33 pm, Amir Tal wrote:
> On Friday 20 September 2002 21:26, Alan Chandler wrote:
> > Can someone explain how ide-scsi gets loaded?
> >
> > I have done grep -r ide-scsi * on /etc and its not reported (and
> > therefore in particular /etc/modules or /etc/modutils have no reference
> > to it).
...
> make sure there is no entry for it in /etc/modules.conf.

I already said I could find no entry there:-(

> also check that its not preloading in lilo (should look something like :
> append = "hdd=ide-scsi")

Nothing in /etc/lilo.conf either

(thats why I did a grep on all of etc - I figured at least one of the config
files must load it - but nothing.)

--
Alan Chandler
[EMAIL PROTECTED]


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: ide-scsi

[Michael Olds]

Alan,

OK I think I know what will do it. Go to modconf which is the module
installation dialog you saw when installing, go to scsi modules and find the
ide-scsi module which should have a plus beside it. Select it and click ok.
You should be asked if you want to remove it, etc.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Alan Chandler [mailto:[EMAIL PROTECTED]]
Sent: Friday, September 20, 2002 2:43 PM
To: [EMAIL PROTECTED]
Subject: Re: ide-scsi


On Friday 20 September 2002 10:03 pm, martin f krafft wrote:

> Why not remove it from the /lib/modules tree and watch the startup
> sequence carefully for an error message? This isn't the "nice" way,
> but it might be your only way. I doubt it'll render your system
> unusable, really!

Did that and rebooted - no error message I could see

Its not loaded, all works well and mplayer plays by dvds :-)

Its just so annoying I can't find out why

--
Alan Chandler
[EMAIL PROTECTED]


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: debian menu in kde

[Michael Olds]

Not a direct answer, but a help: by using >System>Appfinder you can make the
KDE menu include many programs it does not include in the default state.

I also wonder what is going on in the mind's of the developers: if you're
going to have an application and services menu, why not put everything in
it? Or at least give us a way to do that.

Best Wishes!
Mike Olds www.buddhadust.org



-Original Message-
From: Jason Pepas [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 26, 2002 2:11 PM
To: [EMAIL PROTECTED]
Subject: debian menu in kde


where can I read up on the internal workings of the kde menu system?  I
would
like to figure out how to get it to use the standard debian menu system (kde
2.2.2 does its own thing).

-jason


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Konqueror slowness

[Michael Olds]

If it helps your case, I had a similar problem. My solution was to take out
Konquoror and use Krusader instead using Mozilla for the browser. In fact I
have strip't out most of KDE from KDE (strip't everything out of Mozilla
except the browser as well). I like the customization of what I would call
the Desktop, Kmail, the Control Center and that's about it.


Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Stephan Hachinger [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 26, 2002 3:13 PM
To: Mark Roach; [EMAIL PROTECTED]
Subject: Re: Konqueror slowness


Hi!

I have absolutely the same problem, although my kdebase and -libs are
self-compiled... I'm already talking about this with Dirk Mueller,
one of the kde team. If you don't bother, I'll forward your problem
report to him.

Regards,

Stephan

On Thu, 26 Sep 2002 14:30:50 -0500
Mark Roach <[EMAIL PROTECTED]> wrote:

> Konqueror takes an extremely long time to load pages on my system
> compared to mozilla/galeon. Has anyone had a similar problem?
>
> I went to theregister in konq and then while it was loading, I
> lanched galeon, loaded the site and came back and started typing
> this before konqueror loaded the site (~17 seconds just for the
> main content area showed up). Reload takes roughly the same amount
> of time.
>
> I don't remember that having been the case in the past, and
> couldn't find anyone else reporting this problem (everyone always
> talks about konq being fast). Any thoughts on what I might have
> done to upset it? I am using the 3.0.3 debs from the kde site.
>
>
> -Mark
>


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Setting up Sendmail

[Michael Olds]

Hello,

Woody.

I am having some difficulty getting Sendmail up and running.
I installed the program apparently with no problems. I ran: sendmailconfig
after the install, also apparently no problems, no errors reported. But
nothing works either. So I looked in /etc/mail/ and there is no sendmail.cf
so I tried running sendmailconfig again and it asked if it should configure
using the existing sendmail .conf and I said yes. And it reported there was
no /usr/sbin/update_conf which there isn't as it's in /usr/share/sendmail/

(Question one: could I just move it to where it wants?)

There is no existing sendmail.conf in /etc/mail/ either.

When I tried to create a sendmail.cf using:

cd /etc/mail
m4 /usr/share/sendmail/cf/m4/cf.m4 I first got a permission denied and saw
that the files in /m4/ were all marked non executable. I marked them
executable but I am not sure if that was the right thing to do.

I re-ran m4 etc and got this error message:
NONE:0:m4: ERROR: EOF in string

I tried looking up EOF (?end of file?) on Google where there are many
references and no explanations.

I have uninstalled and reinstalled the program several times and I always
end up at this spot. Does anyone know how I get out of this?

Thanks in advance,


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Setting up Sendmail

[Michael Olds]

Well I'm listening. If I can't get it set up reasonably soon I will probably
go with Exim.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Bob Proulx [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 26, 2002 8:50 PM
To: [EMAIL PROTECTED]
Subject: Re: Setting up Sendmail


Michael Olds <[EMAIL PROTECTED]> [2002-09-26 20:02:54 -0700]:
> I am having some difficulty getting Sendmail up and running.

I know that Sendmail has been the workhorse of the Internet for many
years.  I have configured Sendmail cf files by hand for years and
loved the recent m4 configuration.  I am just saying that so you will
know this recommendation comes from someone who has spent much time
with sendmail and still thinks favorably of it.  But it has been
passed on both sides by other programs.

You will probably get a hundred replies to the effect that everyone
recommends that you run either Exim or Postfix.  Either are high
quality mailers.  I prefer Postfix.  It is table driven and very easy
to configure.

  apt-get install postfix

Or you could go with Exim, another fine program.  You will find a lot
of Exim users on this list and can be certain of good user support for
it here.

  apt-get install exim

Bob


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Setting up Sendmail

[Michael Olds]

Hello again,

First, thanks to all who responded with suggestions. I'm still working on
Sendmail. I have it working half way following Oliver's suggestions ...I can
send, but I am not receiving anything, either locally or over the Internet.
I figure it can be the settings at my DNS server, the router's firewall
(don't think so), a sendmail configuration I'm missing...(?, I have SMTP and
local, so it must be one of the exotic ones set up by Debian), or I just
don't have my Kmail mail client set up right (ashamed to say it, I can
configure Windows Outlook accounts, but I have no idea what I should be
telling the Linux machine about loading mail to local/remote users).

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Oliver Fuchs [mailto:[EMAIL PROTECTED]]
Sent: Thursday, September 26, 2002 10:04 PM
To: [EMAIL PROTECTED]
Subject: Re: Setting up Sendmail


On Thu, 26 Sep 2002, Michael Olds wrote:

> Hello,
>
> Woody.
>
> I am having some difficulty getting Sendmail up and running.
> I installed the program apparently with no problems. I ran: sendmailconfig
> after the install, also apparently no problems, no errors reported. But
> nothing works either. So I looked in /etc/mail/ and there is no
sendmail.cf
> so I tried running sendmailconfig again and it asked if it should
configure
> using the existing sendmail .conf and I said yes. And it reported there
was
> no /usr/sbin/update_conf which there isn't as it's in /usr/share/sendmail/

It is a bug ... you first have to cd in the directory
/usr/share/sendmail and then run sendmailconfig.

>
> (Question one: could I just move it to where it wants?)
>
> There is no existing sendmail.conf in /etc/mail/ either.
>
> When I tried to create a sendmail.cf using:
>
> cd /etc/mail
> m4 /usr/share/sendmail/cf/m4/cf.m4 I first got a permission denied and saw
> that the files in /m4/ were all marked non executable. I marked them
> executable but I am not sure if that was the right thing to do.

You can copy the example sendmail.mc and submit.mc from
/usr/share/sendmail/cf/debian/ to your /etc/mail directory (and of
course edit them) and then run
m4 /etc/mail/sendmail.mc > /etc/mail/sendmail.cf (the same for the
submit file). Then cd to the /usr/share/sendmail directory and run
sendmailconfig (I think it has to be run as root).

Oliver
--
... don't touch the bang-bang fruit


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: security update

[Michael Olds]

>I think he was asking how to switch the source.list file to go fetch
>from a server rather than his installation CD set. The answer is of
>course easily found by reading documents on the web site in any case.

He must also comment out (# front of each) each line referencing his CDs or
they will be used over the on line links.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Shyamal Prasad [mailto:[EMAIL PROTECTED]]
Sent: Friday, September 27, 2002 3:49 PM
To: [EMAIL PROTECTED]
Subject: Re: security update


"Bob" == Bob Proulx <[EMAIL PROTECTED]> writes:

Bob> Timur V. Irmatov <[EMAIL PROTECTED]> [2002-09-27
Bob> 14:57:46 +0500]:
>> > Check if you have the following in /etc/apt/sources.list:
>> >
>> > deb http://security.debian.org/ stable/updates main contrib
>> non-free
>> >
>> > If not, add it. Then:
>>
>> Something must be added to sources.list, I suppose?

Bob> Hmm...  The message you were responding to contained the
Bob> answer to the question you asked.  Go back, take a breath,
Bob> read it again, then you will be enlightened.

I think he was asking how to switch the source.list file to go fetch
from a server rather than his installation CD set. The answer is of
course easily found by reading documents on the web site in any case.

Try something like this:

deb http://http.us.debian.org/debian/ woody main non-free contrib
deb-src http://http.us.debian.org/debian/ woody main non-free contrib
deb http://non-us.debian.org/debian-non-US woody/non-US main contrib
non-free
deb-src http://non-us.debian.org/debian-non-US woody/non-US main contrib
non-fre
e

deb http://security.debian.org/ woody/updates main contrib non-free

I guess you could s/woody/stable/ there, I guess I never got around to
it after the release ;-)

Cheers!
Shyamal


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Setting up Sendmail

[Michael Olds]

Hello again,

Woody. Can send but not receive mail.

I checked with my DNS server and I believe it is set up properly.

This seems to be the problem:

in var/log/mail/mail.err
No local mailer defined
QueueDirectory (Q) option must be set

and /etc/mail/sendmail.cf.errors
/usr/share/sendmail/cf/m4/proto.m4:232:M4 Cannot open
/usr/share/sendmail/cf/mailer/.m4 No such file or directory

Well, it's no such file. And I did not see it hanging around in any other of
the likely spots either.

I have a little investment in effort here is why I am hanging in and not
switching right off to the much recommended exim or postfix; but all I need
to be told to make the switch is that these are not my problems with
comprehension, but that the Sendmail program is buggy...anyone care to take
a shot?

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Setting up Sendmail

[Michael Olds]

Alvin,

Thank you for this response.
First problem is: I do not have an /etc/mail/sendmail.cw file (I do have a
local-host-names file with no extension); and I do not have an
/etc/mail/access file or an /etc/mail/sendmail.cw file

For these I think the problem is as is stated in the mail just previous to
this: sendmailconfig cannot find /usr/share/sendmail/cf/mailer/.m4 because
it isn't there.

I checked out the site you recommended this morning, and read a few things,
but it seems more directed towards securing a running system than getting a
system running to begin with...or did I miss a link?

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Alvin Oga [mailto:[EMAIL PROTECTED]]
Sent: Friday, September 27, 2002 4:55 PM
To: Michael Olds
Cc: Oliver Fuchs; [EMAIL PROTECTED]
Subject: RE: Setting up Sendmail



hi ya

to receive email
cd /etc/mail
ln -s local-host-names sendmail.cw
- be sure your domain is listed in the file
make
sendmail - stop/start

for popping email stuff...
- make sure you can "telnet localhost 110"
- than you should use secure pop3 instead of port 110
http://www.Linux-Sec.net/Mail/secure_pop3.txt

sendmail howto
http://www.Linux-Sec.net/Mail/sendmail.gwif.html

c ya
alvin

/etc/mail/access
#
# list of domains/ip# this mschine will be relaying emails
#
foo.com
#
# tou local lan
192.168.0.0
#
#

/etc/sendmail.cw  and/or /etc/mail/sendmail.cw
#
# list of domains tyhis machine will be receiving emails
#
foo.com
www.foo.com
#


On Fri, 27 Sep 2002, Michael Olds wrote:

> Hello again,
>
> First, thanks to all who responded with suggestions. I'm still working on
> Sendmail. I have it working half way following Oliver's suggestions ...I
can
> send, but I am not receiving anything, either locally or over the
Internet.
> I figure it can be the settings at my DNS server, the router's firewall
> (don't think so), a sendmail configuration I'm missing...(?, I have SMTP
and
> local, so it must be one of the exotic ones set up by Debian), or I just
> don't have my Kmail mail client set up right (ashamed to say it, I can
> configure Windows Outlook accounts, but I have no idea what I should be
> telling the Linux machine about loading mail to local/remote users).
>


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Can not acces ide cdrom

[Michael Olds]

Hey Bob, a Re: Bob?

I just went through this recently, so I hope I remember this correctly: the
device you want to operate as a CDROM needs to point to the SCSI device.

In /dev/
make /dev/CDROM a symlink to scd0



-Original Message-
From: bob parker [mailto:[EMAIL PROTECTED]]
Sent: Saturday, September 28, 2002 7:32 AM
To: [EMAIL PROTECTED]
Subject: Can not acces ide cdrom


I am using Debian 3.0.0.
I installed it with a 2.2.20 kernel which did not have scsi support.
I later installed the 2.4.18 kernel and got support for my scsi cd burner
but
I have lost ability to read my ide cdrom.

Here is my fstab
# /etc/fstab: static file system information.
#
#  

/dev/hda8   /   ext2errors=remount-ro   0   1
/dev/hda7   noneswapsw  0   0
proc/proc   procdefaults0   0
/dev/fd0/floppy autouser,noauto 0   0
/dev/cdrom  /cdrom  iso9660 ro,user,noauto  0   0

My cdrom mount points
debian:/home/bob# ls -l /cdrom*/
/cdrom/:
total 0
lrwxrwxrwx1 root root7 Sep 28 23:35 cdrom0 -> /cdrom0
/cdrom0/:
total 0
/cdrom1/:
total 0

my /dev/cdroms
debian:/home/bob# ls -l /dev/cdrom*
lrwxrwxrwx1 root root   11 Sep 28 23:35 /dev/cdrom ->
/dev/cdrom0
lrwxrwxrwx1 root root9 Sep 28 23:35 /dev/cdrom0 ->
/dev/scd0
lrwxrwxrwx1 root root9 Sep 28 23:35 /dev/cdrom1 ->
/dev/scd1

How do I regain access to my ide cdrom?
Thanks
bob


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Can not acces ide cdrom

[Michael Olds]

Bob,

Sorry, I think I misunderstood your situation. You have a CDROM and a CDRW.
Your CDRW is actually a real SCSI device? I was thinking it was scsi
emulation. If I understand this now, the ide CDROM should NOT be being
pointed to the scd0 or anything else but itself. (You may actually have
implimented scsi emulation for the CDROM, which could be another cause of
this problem.) If I have again misunderstood, and you are speaking about two
ide devices (one ide CDROM and one ide CDRW with scsi emulation) you need to
link the CDROM to the scd0...I would think directly not via the CDRW...and
also as just posted, check permissions. Next up you need a more experienced
advisor. If you are speaking about one ide CDRW with scsi emulation then it
needs to be linked to the scd and there should not be a second device in
/dev ... just /dev/CDROM pointing to scd0.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: bob parker [mailto:[EMAIL PROTECTED]]
Sent: Saturday, September 28, 2002 8:33 AM
To: Michael Olds
Cc: [EMAIL PROTECTED]
Subject: Re: Can not acces ide cdrom


On Sun, 29 Sep 2002 00:59, you wrote:
> Hey Bob, a Re: Bob?
>
> I just went through this recently, so I hope I remember this correctly:
the
> device you want to operate as a CDROM needs to point to the SCSI device.
>
> In /dev/
> make /dev/CDROM a symlink to scd0
>
It already is indirectly, ie /dev/cdrom symlinks to /dev/cdrom0 which
symlinks to /dev/scd0. Reading from the scsi device is ok. It is the
ide device that got lost.
What I'm not clear on is whether I need to have the /dev/cdrom1
linked to /dev/sc1 as it now is or /dev/hdb as it is ide.
Cheers
bob

>
>
> -Original Message-
> From: bob parker [mailto:[EMAIL PROTECTED]]
> Sent: Saturday, September 28, 2002 7:32 AM
> To: [EMAIL PROTECTED]
> Subject: Can not acces ide cdrom
>
>
> I am using Debian 3.0.0.
> I installed it with a 2.2.20 kernel which did not have scsi support.
> I later installed the 2.4.18 kernel and got support for my scsi cd burner
> but
> I have lost ability to read my ide cdrom.
>
> Here is my fstab
> # /etc/fstab: static file system information.
> #
> #  
> 
> /dev/hda8   /   ext2errors=remount-ro   0   1
> /dev/hda7   noneswapsw  0   0
> proc/proc   procdefaults0   0
> /dev/fd0/floppy autouser,noauto 0   0
> /dev/cdrom  /cdrom  iso9660 ro,user,noauto  0   0
>
> My cdrom mount points
> debian:/home/bob# ls -l /cdrom*/
> /cdrom/:
> total 0
> lrwxrwxrwx1 root root7 Sep 28 23:35 cdrom0 -> /cdrom0
> /cdrom0/:
> total 0
> /cdrom1/:
> total 0
>
> my /dev/cdroms
> debian:/home/bob# ls -l /dev/cdrom*
> lrwxrwxrwx1 root root   11 Sep 28 23:35 /dev/cdrom ->
> /dev/cdrom0
> lrwxrwxrwx1 root root9 Sep 28 23:35 /dev/cdrom0 ->
> /dev/scd0
> lrwxrwxrwx1 root root9 Sep 28 23:35 /dev/cdrom1 ->
> /dev/scd1
>
> How do I regain access to my ide cdrom?
> Thanks
> bob


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Setting up exim: was Setting up Sendmail

[Michael Olds]

Hello,

Not because I gave up on Sendmail, but because of another problem (for some
reason (I believe it had to do with uninstalling Sendmail which I did in
order to start from scratch and install the package from source) when I shut
down and restarted, Apache was unavailable, all the files are in place, but
it is listed as not installed and doesn't work. apachectl start would get no
sucha commandment...anyway, to fix the Apache problem I reinstalled,
reinstalling installed exim...so I thought I would give it a spin...). But
here I am in the same situation: I can send mail outside, but I cannot
receive mail from outside or internally. The error log reads:
date refused relay (host) to <~internal@~> from <~external@~>
and one from local to local

In other words, from my little understanding: the mail is getting to me, but
it isn't being distributed. In other words, no problem at the DNS level or
at the router, some problem either with exim config or in my Debian/linux
setup. I am not sure if my Sendmail problems were caused in the same way, if
so, the problem is with my Debian/Linux setup. I have read a couple of books
and looked on the exim site with no joy.

I have the following in exim.conf
local_domains = listed:all:variations domain:domain.net:mail.domain.net
I added:
hosts_treat_as_local = all the above listed here as well
relay_domains = all the above listed here as well
NOT UNCOMMENTED: RELAY_DOMAINS_INCLUDE_LOCAL_MX = TRUE
host_lookup = *
host_auth_accept_relay = * (none set up)

I also listed all the domains and variations of the names in
/etc/hosts_allow
ALL: domain
ALL: domain.net
and so forth one per line; did not do anything with hosts_deny

Any pointers? Any and all help appreciated.






-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Setting up Sendmail

[Michael Olds]

Thanks for this Oliver, however I think the setup is too different to apply
to my situation: DSL always on, DNS servers running.

Best Wishes!
Mike Olds www.buddhadust.org

-Original Message-
From: Oliver Fuchs [mailto:[EMAIL PROTECTED]]
Sent: Sunday, September 29, 2002 4:33 AM
To: [EMAIL PROTECTED]
Subject: Re: Setting up Sendmail


Sorry,

it has to mailq.

Oliver
--
... don't touch the bang-bang fruit


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact
[EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Email (was Setting up Exim was Setting up Sendmail)

[Michael Olds]

Hello,

Exim on Woody

Trying again here, I posted on Friday afternoon and by Saturday I knew there
wasn't a hope of my request for help being seen.

At this point I have again run out of possible solutions:

I have Exim set up. I believe it is set up properly as I used the defaults
mostly. I can send mail from my Linux box, but I cannot receive mail, either
internally or from the outside. The log indicates that mail is being
received and is being delivered no errors. I am using the Kmail e-mail
client, which I have uninstalled and re-installed a couple of times just to
check. If there is a better client I am listening. I am sure this is
something silly, like permissions, but I do not know where to look next for
the solution. As far as I can tell the relevant directories and files are
user me and group mail, not sure if it is correct or safe, but also with
read, write, and execute for both.

Anyone have an idea as to what I should dig into next to get this up?

Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Email (was Setting up Exim was Setting up Sendmail)

[Michael Olds]

Carel,

No problem replying off list, and thank you for replying, period, I have two
machines one Windows 2000 Pro and one Linux. I will answer your questions,
below your questions below:

[ sorry for sending this offlist too, but as you have email problems I'm not
  sure list trafic will reach you.  If it does, pray tell and I wil adjust ]

> I have Exim set up. I believe it is set up properly as I used the defaults
> mostly. I can send mail from my Linux box, but I cannot receive mail,
either
> internally or from the outside. The log indicates that mail is being

so it isn't set up properly:)  For starters, we need some more info.
You say you've used the defaults during exim install/setup.  Did you
choose the "Internet site" or "Satellite" or...

So "Internet" or what was the #1 choice.

The next question is, how much do you need/want your machine to do:)
I normally prefer to have an ISP take care of all delivery and receiving
problems, but others really want to run their own full blown mail server.

In general what I am aiming to do is to get to the point where I have a self
sufficient setup running off my DSL line. I have a very low traffic site
plus one test site and one personal site, and will have only a few e-mail
addresses to service. My DNS service is being handled by VeriSign, and I am
only working with the test site and the personal site, BuddhaDust is being
handled by a hosting service. When I get the rest set up and running I will
take DNS on as well...right now I can use one less problem.

> received and is being delivered no errors. I am using the Kmail e-mail
> client, which I have uninstalled and re-installed a couple of times just
to
> check. If there is a better client I am listening. I am sure this is

The email client just hands it over to a email server, and depending
on your configuration of your email client that could well be your
IPS's mail server.  To shortcircuit such debugging fun, you could try
to use "mail" (install mailutils or mailx if you're missing this one)
like this:

   # to check the mail gets delivered open a log window and enter your
passwd
   $ xterm -e su - root "tail -f /var/log/exim/*log" &

   # send a mail from and to your normal non root login
   $ echo just testing | mail -s test $USER

   # look at the log window, this should look like:

==> /var/log/exim/paniclog <==

==> /var/log/exim/rejectlog <==

==> /var/log/exim/mainlog <==
2002-10-02 02:20:10 17wXFK-0005dU-00 <= [EMAIL PROTECTED] U=carel P=local
S=319
2002-10-02 02:20:11 17wXFK-0005dU-00 => [EMAIL PROTECTED]
<[EMAIL PROTECTED]> R=smarthost T=remote_smtp H=schaduw.felnet
[192.168.1.1]
2002-10-02 02:20:11 17wXFK-0005dU-00 Completed

Little confusion here, but it comes down to the fact that the log reads as
does your last entry here.

I am talking about when I send mail from the Windows machine to the Linux
box, the linux box is recording such entries in /var/log/exim/mainlog and
similar entries for mail being sent from root to userme. I am just not able
to get them to show up in my email client. Again, I can send from Linux to
the Windows box no problem.



> something silly, like permissions, but I do not know where to look next
for
> the solution. As far as I can tell the relevant directories and files are
> user me and group mail, not sure if it is correct or safe, but also with
> read, write, and execute for both.

I'm not sure what files you're talking here, but I sure hope your not
refering to e.g. files in /etc?

in /etc/exim
I have permissions for /etc/exim/ user root rwe; group mail re (is it Debian
or Krusader that gives permissions in terms of "show entries" "write
entries" "enter" why not use the numbers, or the usual rwx?)
for /etc/exim/exim.conf user root rw; group root r
for /home/userme/mail user me rwe group mail rwe
for /var/mail/userme user root rwe group mail rwe set Gid others re sticky
(?)

Again, thanks,
Best Wishes!
Mike Olds www.buddhadust.org


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Email (was Setting up Exim was Setting up Sendmail)

[Michael Olds]

Stephen,
Thank you for this reply. Answers to your questions below your questions:

-Original Message-
From: Stephen Gran [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, October 01, 2002 6:16 PM
To: Debian-User@Lists. Debian. Org
Subject: Re: Email (was Setting up Exim was Setting up Sendmail)


This one time, at band camp, Michael Olds said:
> Hello,
>
> Exim on Woody
>
> Trying again here, I posted on Friday afternoon and by Saturday I knew
there
> wasn't a hope of my request for help being seen.
>
> At this point I have again run out of possible solutions:
>
> I have Exim set up. I believe it is set up properly as I used the defaults
> mostly. I can send mail from my Linux box, but I cannot receive mail,
either
> internally or from the outside. The log indicates that mail is being
> received and is being delivered no errors. I am using the Kmail e-mail
> client, which I have uninstalled and re-installed a couple of times just
to
> check. If there is a better client I am listening. I am sure this is
> something silly, like permissions, but I do not know where to look next
for
> the solution. As far as I can tell the relevant directories and files are
> user me and group mail, not sure if it is correct or safe, but also with
> read, write, and execute for both.
>
> Anyone have an idea as to what I should dig into next to get this up?


OK, if you have a default Debian install, exim installs mail into
/var/mail/$USER, and /var/spool/mail is a symlink to /var/mail.

ls -l /var/mail/$USER will tell you if there's any mail there (if it
has a nonzero size, you do.)  If there's mail there, then the problem
is the setup of your mail clients in reading it.  All of the mail
clients (AFAIK, correct me if I'm wrong) can be made to read from the
mail spool.


ls -1 /var/mail/$USER gives me
/var/mail/userme and that is all.
the file itself if 700kb which isn't zero.


<===>

If you can send mail, exim is probably working - it's job is SMTP only.
exim (or postfix/sendmail, etc) do not _retrieve_ mail.  You will
only get mail sent to you by local jobs (cron, etc) or other users on
your box, unless this box is also the MX in the DNS for your domain.  I
don't have any details of your setup, but if this is a standard
workstation install, without a domain name attached to it, this doesn't
seem likely.  Again, correct me if I'm wrong, and fill in other details
as you think necessary.

I can send mail out. Are you saying exim does not process the mail that is
received? ( don't mean retrieve it as with an email client, but put it in
/var/mail/$USER ?

In any case yes this box is the MX in the DNS for this domain, and one
other...fully qualified, registered and, again, apparently working because I
just sent me an email from the Windows box and the /var/log/exim/mainlog
recorded it "completed"

<===>
The thread got drowned in the weekend traffic. I was having this same
problem with Sendmail. When I uninstalled sendmail, it boinked Apache (don't
ask me how), re-installing Apache installed Exim, so I thought I would give
it a try...about brings you up to date.


I missed where this thread started, but you make reference to sendmail.
I don't know sendmail that well (does anybody? (^8 ), but I do get the
gist at least of exim, and I'll help with what I can.


<===>

Steve
--
To see a need and wait to be asked, is to already refuse.

These two are hard to find in the world: one who offers and one who is
grateful. -SGS


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

RE: Email (was Setting up Exim was Setting up Sendmail)

[Michael Olds]

Stephen,

These are the settings I have in Kmail (I have tried multiple variations)

name: userme
location: /var/mail/userme
checked: Procmail locfile /var/mail/userme.lock

I also tried:
location: /home/userme/Mail and Mail/

nada

Best Wishes!
Mike Olds www.buddhadust.org 


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED] 
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]