Re: [expert] Norton Antivirus and Bloodhound
Civileme wrote: The "local" Alaska Linux Users Group reports that Norton Antivirus and Bloodhound, Norton's newer "heuristic" virus hunter, is claiming LILO is a boot sector virus in newly installed dual-boot systems. I think NAV has been doing this for several years. I don't think that NAV actually thinks LILO is a virus, but instead merely sees the mbr is not as it would be on a system with only MS OSs installed, that is, NAV is merely warning the user of something it doesn't understand and therefore can't properly interpret for the user. With the increasing popularity and use of Linux on dual-boot systems, though, NAV or BH should definitely allow for other boot managers being installed in the mbr. It's good to know that someone's on the ball though - McAfee. mike So, for their next installfest, they will be recommending the McAfee virus scanner for linux that is there to protect windows. Civileme -- Beta-Testing Netscape 6 Mailer
Re: [expert] RPM woes
Bruce E. Harris wrote:
Hi,
I am a convert moving over from Caldera OpenLinux 2.3 (and before). So far I
like Mandrake much more. However I have a very disturbing problem. RPM won't
work anymore run anymore. Attempts to install a RPM will lock my system badly.
Usually causing it to log out of X, many time I must do a hard shut down (kill
power), or the HDD will just chatter and do nothing.
I get an error ending with can't update (some lib6 cant remember the
whole name, and the error msg will not generate anymore) but the last part is
/var/lib/rpm/requiredby.rpm
Any ideas?
Not really, but you might try reinstalling rpm from the Mandrake/RPMS directory
on the install cdrom. Then, if you still have these problems, try rebuilding the
rpm database
rpm --rebuilddb
If you still have problems using rpm, then read the rpm man pages, thoroughly.
There's also a --initdb option, but the documentation or man page doesn't explain
either of these options to any noteworthy extent.
What version of rpm are you using, if you still have problems?
rpm --version
Read the man page on rpm and there you'll also find verify options. I'm not sure
if this would do any good wrt the rpm requiredby database, but it might be
helpful in verifying rpm itself.
You can also do a search for RPMs which require the lib*6 file which seems to be
a problem.
rpm -q --whatrequires {filename}
{filename} may need the entire path.
You can also check what RPM provides this problem file
rpm -q --whatprovides {filename}
In this case, you may or may not need the complete path.
Do a search for the directory where this problem lib*6 file is located. It
should be in either /usr/lib or /lib, but if it's not there, then use the find or
rpm command
find / -name {filename}
Using rpm, you need to know which package or RPM provides this lib*6 file, such
as from the --whatprovides rpm query. Once you have this information
rpm -qil {package-name}
{package-name} must only contain the base name of the package, or at least no
more than this part and the version+release. Don't include the .i386.rpm or
.noarch.rpm part, because this part isn't recorded in the rpm database.
If you still have problems, then post a follow-up; although, you might want to
first check what you could do using other rpm options.
mike
--
Best Regards,
Bruce
Re: [expert] Stop using kdm
Jean-Louis Debert wrote: Mike Corbeil wrote: [snipped] Not the same thing. To use the graphical login, you _have to_ be _booting_ in graphical mode, i.e. you have to configure init to start directly in graphical mode. In your current situation, if you get a text login, it's because your init is configured that way, and so you don't use *dm. You probably can find some *getty in ps output. Exactly. There are no *dm instances in the output of ps aux but there are several *getty running. This is fine with me, though. The only condition I'ld see as important to use *dm would be if it provided increased or enhanced security, but I don't know enough about these, yet, to be able to say. Do *dm provide enhanced security, or just a fancier login interface? I should try them just to see what the gui's look like, albeit probably something like on Solaris and gui or character doesn't make any difference; only gui looking prettier. If there's a security difference, though, then a la gui login I'll switch to. mike
Re: [expert] Re: Dependency help
Brian T. Schellenberger wrote:
Mike Corbeil wrote:
[snipped]
Well, if you *only* do --nodeps that's probably true, but if you use
--force you can get in trouble: the un-install could un-install stuff
that was really installed already.
un-install usually means removing stuff that was already installed, as far
as I'm aware.However, guessing, --nodeps could also cause problems.
According to the aforementioned referenced documentation, --nodeps can
install an app and this might work, but it can also cause other apps to be
"broken", and I don't have any difficulty believe that this is definitely
possible, and does happen.
Uninstall only un-installs what the package *thinks* it installed; once
you start messing around with the rpm options you can confuse this
process.
Uninstall can only remove something which has been installed, previously,
regardless of whether the uninstall works correctly or not. You cannot delete
something which doesn't exist.
However, it is important to know that rpm can be caused to screw up, albeit this
should not happen and is indicative of bugginess.
I have found that the best approach is to install/re-install individual
packages from hither yon as infrequently as possible; unless you have
some specific reason to do so, stick to the packages that all came
together on your distribution.
There's no reason to install or reinstall unless the package is of use. Over
the past couple of months, I've reinstalled many packages, because I'm upgrading
the distribution and many packages need to be upgraded. This is one example
where installing updates is important.
Of course, I could purchase a new boxed edition, delete the old configuration
and install the new one, but I'm doing the upgrade manually. In the process,
I'm learning considerably, and learning is important, if not essential;
essential for me anyway, because working with computers isn't a mere hobby for
me. This also happens to help find bugs and therefore help to improve tools or
applications.
For entire newbies, I'ld suggest installing from a packaged cdrom.
I, at least, find it easier to just backup my system and re-install
periodically just to "keep up with what's new" and only install
piecemeal when I have a specific bug or problem.
That's the recommended way, but if doing upgrades entirely with downloads, and
upgrading is definitely appropriate, as it is in my case, then this is an
exception. I believe this is covered in the Software Building HOWTO, or is in
some other LDP documentation.
In the process, I'm learning ample, and have already discovered a couple of
bugs, perhaps one only being a near bug, which will be useful in improving the
reliability, fool proofness, of the tool these bugs occurred with.
A very important and useful approach to developing sw is to apply the "what if"
strategy or approach. "What if the user does ..., will our sw be able to handle
the situation, and is it worth considering, because developing sw does require
time?". Often, sw is developed with inadequate emphasis on the "what if"
approach, but that's the only way to develop reliable sw the first time,
instead of relying on end users or customers having problems with one's sw.
Zacharie Richard sang "travailler est trop dur et voler n'est pas beau", but
there was no real conclusion ("working is too difficult or hard and stealing
isn't nice").
mike
You might try rpmdrake; it can resolve dependencies and install them all
automatically.
Thanks for that point. Definitely sounds better than the rpm I'm currently
using, 3.0.2. I've already found a couple of "bugs" with this version, one
when erasing or removing 3.0.2 after it's been installed "next" to a prior
version, and the other being that I now, somehow, have three instances of
3.0.2 reported with rpm -qa. Not sure how the latter happened, but
definitely seems wrong.
rpmdrake is a GUI on top of RPM, not a replacement rpm.
In general packages don't install next to other packages, they replace
them.
I know of no reason why it wouldn't work with RH.
On the other hand, the first bug is easy to recover from once one knows
how, especially if there's another Linux configuration files and
directories can be copied from, and the second bug doesn't prevent rpm from
working; just that there are more instances reported by rpm -qa than
necessary. Guessing, I believe that this may be more of a rpm database
tracking error, but harmless.
I'll take a look into rpmdrake, and believe it works for both Mandrake and
RH.
Screwed up my test/build system tonight doing configure, make and make
install, for glibc2 2.x.x .tar.gz archive. Can't boot into that
configuration any longer and now need to figure out how to undo or
repair. That teaches me one lesson, to not install to /usr, but instead as
the IN
Re: [expert] Stop using kdm
John Aldrich wrote: KDM/GDM/XDM only work when you're booting into GUI mode. They do NOT work when you boot to console mode. John I'll need to read up on the advantages of one mode over the other. I don't mind console mode, but always run startx when I want to work in a gui desktop environment, and have no problems using a terminal window to do much of my work. I don't mind command line mode and definitely don't see any real advantage with MS Windows in this respect. I spend much time in terminal windows, and the programming editor I normally use is vim (not even bothering with gvim). However, if there's an enhanced security advantage using kdm or gdm or xdm boot mode, then I would definitely switch. Otherwise, am in no hurry to do so. Is there such an advantage booting with these as the boot mode? mike
Re: [expert] Booting problem-----more
J D wrote:
i got into linuxconf. when i got to access local volumes, all it has is:
/dev/zip for source, /mnt/zip for mount point, vfat for FsType.
it says it three times. there is nothing else. could this be the source of
my problems? i think i know what's going on, but since i'm new to this, i
don't want to get ahead of myself.
Understandable.
Add the dos drive or partitions, if that's what you want to do; however, you
need to know what the /dev/{filename} is. You can find that out with fdisk.
If your dos partition is on the first hdd, then run fdisk without any
arguments; if on the second hdd, then run with /dev/hdb for an argument. The
screen you need to use is Config | Boot, instead of Config | Filesystems.
mike
On the other hand, that's for adding filesystem fstab mounts, whereas I believe
what you're trying to do is to configure lilo for a boot manager.
In this case, you need to provide the following information:
- the contents of your lilo.conf file, and
- what hdd your dos boot partition is on, hda or hdb.
Also, look using either fdisk, or looking at your fstab file, to see which
/dev/hd[ab]{n} device file your dos boot partition is mapped to. If it's at
the beginning of hda, then it's probably /dev/hda1, and in this case, if you
want to place lilo boot manager in the MBR (master boot record), then you need
to make sure that the first line of your lilo.conf file (in the /etc directory)
is
boot = /dev/ hda
You can configure lilo using linuxconf, but you described the wrong screen,
above.
Re: [expert] Stop using kdm
Ken Archer wrote:
Xdm (generic), kdm (KDE) and gdm (gnome) all provide another level of security,
but my original question was about doing away with any of the login managers.
Once you have setup Mandrake to launch X and go directly to a gui, the default
is to bring up Kdm. Now I would rather go directly to a consol screen and
then, using "startx" go directly to my window manager of choice as set up in
~/.xinitrc. What I can't find is the file to edit to take kdm out of the
process.
If you're using KDE, then check through the online HELP available in the
KDE desktop. I remember reading about switching gdm for kdm and then doing the
reverse, when I first installed KDE, but don't recall what the commands were. There
should be documentation through KDE Help, which should be accessible through at least
the KDE K menu.
However, I seem to also recall reading about some gui login manager which allows
choosing which wm to login to. Am not sure where you'ld find the documentation on
this, but if a gui login provides enhanced security, then maybe you'll want to search
for this info, before switching modes.
If you definitely want to switch to console login, then another thing you can try is
looking through the configuration files under /etc. Maybe this would be defined in a
file in the /etc/rc.d directory or one of the subdirectories, perhaps the
/etc/rc.d/rc.sysinit file, or /etc/rc.d/init.d/{some-file}, or some other file under
/etc/rc.d. Might also be in some file in /etc. I don't think it'ld be under
/etc/X11/, but may be.
I've resolved a number of problems and performed tasks by learning this way;
therefore, it's definitely do-able, and sometimes faster than asking in a mailing
list.
You should have some configuration howto for the Mandrake distribution, or Linux in
general. Take a quick look through these HOWTOs. I think that the ones for
configuring linux aren't bad.
mike
Wed, 19 Apr 2000, you wrote:
Jean-Louis Debert wrote:
Mike Corbeil wrote:
[snipped]
Not the same thing. To use the graphical login, you _have to_
be _booting_ in graphical mode, i.e. you have to configure
init to start directly in graphical mode.
In your current situation, if you get a text login, it's
because your init is configured that way, and so you don't
use *dm. You probably can find some *getty in ps output.
Exactly. There are no *dm instances in the output of
ps aux
but there are several *getty running.
This is fine with me, though. The only condition I'ld see as important to use
*dm would be if it provided increased or enhanced security, but I don't know
enough about these, yet, to be able to say.
Do *dm provide enhanced security, or just a fancier login interface?
I should try them just to see what the gui's look like, albeit probably
something like on Solaris and gui or character doesn't make any difference; only
gui looking prettier. If there's a security difference, though, then a la gui
login I'll switch to.
mike
--
Kenneth Archer + San Antonio, Texas
[EMAIL PROTECTED] ICQ #24980801
Powered by Linux ++ Mailed by Kmail
Re: [expert] Booting problem-----more
J D wrote:
okay, i figured out what is wrong. my hard drive has all of a sudden become
read only.
It's normally read-only during a fair portion of the boot process; however,
before this process ends, the filesystems are remounted rw, normally.
i didn't think this was possible, but here i am.
what i don't understand is how it got that way, and how to fix it. any help
would be great.
also, i have looked in my fstab file. for some reason, only stuff for the
zip drive is there. now, considering i don't know very much, i think that
that is a major problem.
Definitely. The various filesystems for your Linux configuration should
definitely be in the fstab file.
If youc can, then provide copies of your /etc/fstab and /etc/lilo.conf files.
You should also include the output of dmesg. Redirect the output of dmesg to a
file which you'ld then copy+paste, through what ever means you wish, into the
body of the email.
dmesg dmesg.out 21
This should or must be run as root (don't know if it can be run as any other
user).
Don't send as attachments, because I, for one anyway, certainly won't open
them. Very rare is the attachment that I'll open. Do a copy+paste, or insert
the texts by other means.
I'ld definitely suggest that people not use attachments in these mailing lists;
just a general rule of thumb.
If you don't know of a way to do this and you're, for example, using netscape,
then open an additional navigator window, open the text file in that window, do
a Select All, Copy, and then paste into the body of the email message. This is
one way which definitely works, with netscape 4.05 anyway.
mike
From: Mike Corbeil [EMAIL PROTECTED]
Reply-To: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: Re: [expert] Booting problem-more
Date: Wed, 19 Apr 2000 20:22:47 -0400
J D wrote:
i got into linuxconf. when i got to access local volumes, all it has
is:
/dev/zip for source, /mnt/zip for mount point, vfat for FsType.
it says it three times. there is nothing else. could this be the
source of
my problems? i think i know what's going on, but since i'm new to this,
i
don't want to get ahead of myself.
Understandable.
Add the dos drive or partitions, if that's what you want to do; however,
you
need to know what the /dev/{filename} is. You can find that out with
fdisk.
If your dos partition is on the first hdd, then run fdisk without any
arguments; if on the second hdd, then run with /dev/hdb for an argument.
The
screen you need to use is Config | Boot, instead of Config | Filesystems.
mike
On the other hand, that's for adding filesystem fstab mounts, whereas I
believe
what you're trying to do is to configure lilo for a boot manager.
In this case, you need to provide the following information:
- the contents of your lilo.conf file, and
- what hdd your dos boot partition is on, hda or hdb.
Also, look using either fdisk, or looking at your fstab file, to see which
/dev/hd[ab]{n} device file your dos boot partition is mapped to. If it's
at
the beginning of hda, then it's probably /dev/hda1, and in this case, if
you
want to place lilo boot manager in the MBR (master boot record), then you
need
to make sure that the first line of your lilo.conf file (in the /etc
directory)
is
boot = /dev/ hda
You can configure lilo using linuxconf, but you described the wrong screen,
above.
__
Get Your Private, Free Email at http://www.hotmail.com
Re: [expert] Re: Dependency help
Brian T. Schellenberger wrote: [snipped] This sounds pretty odd to me. I mean, if you get the latest sources and rebuild everything, then it should be reliable in that you know you built for your system with the same libraries, but RPMs should work well, too, most of the time. The previously referenced documentation did say that RPMs should work well, most of the time, but because of problems, at least in the past, the document emphasized using .tar.gz sources, particularly instead of using binary RPMs, but also in any case. However, rpm has probably evolved since that documentation was written. I've been using rpm all along, so far, but do run into dependency problems, some which are a real pain in the neck. And RPMS can contain before and after scripts to fix things up in case there's additional work to do. Many of them seem to contain pre and post scripts. [snipped] Well, if you *only* do --nodeps that's probably true, but if you use --force you can get in trouble: the un-install could un-install stuff that was really installed already. un-install usually means removing stuff that was already installed, as far as I'm aware.However, guessing, --nodeps could also cause problems. According to the aforementioned referenced documentation, --nodeps can install an app and this might work, but it can also cause other apps to be "broken", and I don't have any difficulty believe that this is definitely possible, and does happen. You might try rpmdrake; it can resolve dependencies and install them all automatically. Thanks for that point. Definitely sounds better than the rpm I'm currently using, 3.0.2. I've already found a couple of "bugs" with this version, one when erasing or removing 3.0.2 after it's been installed "next" to a prior version, and the other being that I now, somehow, have three instances of 3.0.2 reported with rpm -qa. Not sure how the latter happened, but definitely seems wrong. On the other hand, the first bug is easy to recover from once one knows how, especially if there's another Linux configuration files and directories can be copied from, and the second bug doesn't prevent rpm from working; just that there are more instances reported by rpm -qa than necessary. Guessing, I believe that this may be more of a rpm database tracking error, but harmless. I'll take a look into rpmdrake, and believe it works for both Mandrake and RH. Screwed up my test/build system tonight doing configure, make and make install, for glibc2 2.x.x .tar.gz archive. Can't boot into that configuration any longer and now need to figure out how to undo or repair. That teaches me one lesson, to not install to /usr, but instead as the INSTALL file says to try first, to /usr/local. mike
Re: [expert] Stop using kdm
Stephen F. Bosch wrote: Matt Stegman wrote: No, startx does not start kdm. startx will start X, and load kde by default, or another window manager - you can use any of several tools to change this. Qué? In Mandrake, startx gives you kdm. -Stephen- The relationship is indirect. startx will give you what ever it's been instructed to give you. One or the principal file for this is $HOME/.Xclients, or ~/.Xclients (~/ = $HOME/). mike
Re: [expert] Unzipping a file
Brian T. Schellenberger wrote: Mike Corbeil wrote: Brian T. Schellenberger wrote: In the vain hope of forestalling a number of "you moron" followups, I will point out that I saw the recent posting explaining that there's a "y" option that does exactly this. In my defense, it's not in the "man" page. man or documentation page bug. y option for what? I just checked the man pages for tar and bzip2 and also didn't see any mention of a y option. Yes, exactly. tar xvyf foo.tar.bz2 extracts and unzips in a single step. Cool, eh? Undocumented, alas. I found some documentation yesterday which gave such examples and I believe that was in the LDP web site. This means that it's documented, except not locally. However, people can mirror the LDP, or download it, or copy it, to have it accessible locally, instead of always needing to hook up to the internet to get to the LDP. Wasn't sure how much I'ld use the documentation, but copied it to my local /usr/doc directory, organized the documentation by categorizing into directories to speed up searches, and it's been useful. Now, I just bring up kfm, as long as I'm using kde, and browse, locally. The only potential problem with that is needing to check once in a while to see if documentation has been added or changed, but most of the documentation is dated, which helps to make the checking fairly quick to do. I agree that it's quicker to use the extra option with tar, and that it may be more useful the way you described. I just haven't had the need to bother, and also wonder if the option is also available with tar on Unix, like Sun, ... One can always try and if it doesn't work, then read the man page, but what if it works, but does something different? Am not sure and I don't have a Unix system available to check. Just some thoughts; although, I don't think they'ld change the meaning or effect of an option for a Unix command, for this reason. The other aspect of copying the LDP and such documentation locally is (besides making sure not to infringe upon copyrights) creating a script or tool to provide one fairly consistent user interface and which searches man pages, info, and all other local documentation files, to locate documentation. This would create one integrated documentation search and display user interface, albeit the various types of documentation would be presented according to what ever tool or type of tool is otherwise used to view the documentation when it's accessed more directly, or directly. I have most of such a script already completed, but presently have higher priorities and need to create a gui front end, before this tool is finished. Also, maybe there already is such a tool, but I'm not aware of one; therefore, this tool is agenda'd, among other things which are also deferred due to higher or more urgent priorities. Without a tool like this, a user must try man, info, and if these fail, then bring up a browser, or use rgrep, to search through the many HOWTOs, FAQs, and other copied documentation. One integrated u.i. seems like a nicer approach. Then, whether people copy documentation from the internet or not, is up to each to choose. Documentation is definitely nice to have, especially up to date documentation, and free documentation puts icing on the cake. mike
Re: [expert] Stop using kdm
Ken Archer wrote: When I installed 7.0 I set it up to go directly to X using "kdm". On a single user machine, however, I prefer to use "guichooser" to choose my window manager. Question is how do I stop it from using kdm and go to a console? If I set it to go to init3 and then use startx, it will still go to kdm if I am not mistaken. Once you're logged in and in kde, bring up kde help and search for how to disable kde. There's some command like kde on/off but have forgotten exactly what it is. This should replace the .Xclients file with one which doesn't invoke startkde. Then, I'ld suggest reading up on the window managers and the tool you mentioned. There may be a man page or documentation under /usr/doc for this tool. mike
Re: [expert] Dependency Help!!
Orlando Lewis wrote: Where can I find the following dependencies and/or what packages do they come with? I am using RH 6.1 libstdc++_libc6.1-2.so.0 libcrypto.so.0 TIA That's a good question as far I'm concerned, because I'm also getting problems due to these as well as others, like libc.so.6. I have a symbolic link libc.so.6 to some libc.so.2.0.7 file, or something like that, but either this file isn't the correct version, although it looks like it should be, because the libc.so.6 messages always say either "(GLIBC_2.0)" or "(GLIBC_2.1)", and 2.0.7 is greater than 2.0. However, I read something yesterday saying that libc* packages should be removed, only leaving glibc* packages, but the documentation I've found on this, so far, doesn't cover this much more than this, and I'm a little concerned about removing the libc.5.4.* package until more thorough documentation is found, or someone can answer the question based on real experience. I'm also getting problems with libm, libl, and others, as well as libstdc++-libc6* and libcrypto*. I've searched considerably for the answer, but haven't found anything more than removing libc* and something very vague about symbolic links being wrong and needing redefinition, or the files being copied to the correct directory, and resinstalling some package(s). Haven't found any clear documentation on this and if I do anything of what I've found so far, then it'ld be by using very vague guidelines, so vague it leaves me wondering if who ever wrote the documentation is sure of what he or she is saying, or maybe it's just that they don't like to write, or can't translate experience or knowledge into writing, or don't want to because they wouldn't be getting paid. What ever the reason, I haven't found any documentation leaving me feeling reasonably confident; only leaving me feeling like I'm going by some person's ideas, while it being really, only trial and error kind of testing. However, one piece of documentation I came across yesterday said that this kind of problem is easier to avoid by avoiding RPMS and instead using .tar.gz downloads, and then running configure, make and install. Actually, this is the same piece of documentation first referred to, above. It said that RPMs are supposedly, considerably unreliable in this way, whereas downloading .tar.gz archives is considerably more reliable. Of course, one could always try these guidelines, or installing RPMs using the --nodeps (skip dependency checking) option, but I wouldn't do this without first backing up the system. On the other hand, it may not be necessary to backup the system, because if the --nodeps allows package installation and then apps don't work, then the package(s) can probably be removed without any problems. I don't know installation of RPMs well enough to be able to say this for sure, though. That's what I've found on this kind of topic, so far. The first referenced piece of documentation is in the Software Building HOWTO, which can be found through the LDP. I found it in the html HOWTOs, there. http://www.linuxdoc.org mike
Re: [expert] Stop using kdm
Brian T. Schellenberger wrote: "Stephen F. Bosch" wrote: Matt Stegman wrote: No, startx does not start kdm. startx will start X, and load kde by default, or another window manager - you can use any of several tools to change this. Qué? In Mandrake, startx gives you kdm. No, it doesn't. I think that what the person means is that Mandrake configured kde as the default window manager and desktop. This may have been through a user choice made during the installation process, or the default the Mandrake distribution which was installed. To an unknowing user, this could cause startx to look like it gives kdm or kde as the default. Only by reading documentation on this does a person realize that startx only gives or brings up what it's instructed to do so, through configuration files, e.g. ~/.Xclients. Without reading the documentation on this process, many newbies to X and the wm's can easily think that startx is to "blame". startx is a "front-end" and depends on configuration files, to know what to do. To learn about startx and configuring which wm's to use or have a choice of, people need to read the documentation on this, or wait until someone repeats it. True, startx doesn't decide what wm is to be used, but it can seem like it does for people who don't know what's actually going on. mike
Re: [expert] LILO problem
Kirk McElhearn wrote: I'm really getting fed up with this problem. So many people said it was so easy to use a big HD and boot more than one distribution, I wish there were some way to figure out why it won't work for me I just tried something - I have a Mandrake boot floppy, and I figured that maybe I could just copy it and change some config in the floppy to get the other distro to boot. But when trying to access the floppy, I get a message Could not list directory contents. Is this floppy some format that cannot be read by Mandrake? I don't think that Mandrake should have any problems listing the contents of the disk. Have you tried listing the contents from a dos boot, like into Win9x or NT? If you can list the contents there, then you can definitely list the contents with Linux, but the mount definition may need to be changed to specify the correct filesystem type. To check the latter, you can try mounting the floppy to ext2 and to vfat, if ext2 fails. If both of these fail, then haven't a clue. mount -vt vfat /dev/fd0H1440 /mnt/floppy mount -vt ext2 /dev/fd0 /mnt/floppy mount -vt vfat /dev/fd0 /mnt/floppy You could use /mnt/a:, instead of /mnt/floppy. Doesn't matter, as long as the /mnt mount point isn't already mounted. The only time this can otherwise matter is if you try to mount without specifying the -t option and argument. If none of these mounts works, then haven't a clue. If any of these mounts works and you can't list the contents of the floppy disk, then haven't a clue. On the other hand, the above is based on the assumption that you're trying to list the contents of a Linux boot floppy, which should be listable. As for large disks, you might find some useful information through the LDP http://www.linuxdoc.org Check through the list of HOWTOs. You should find one on using Large Disks. Don't know if it'll contain the answer to your problems, but you should read the howto anyway. mike
Re: [expert] LILO problem
Kirk McElhearn wrote: On 16/04/00 23:11, Andrew George [EMAIL PROTECTED] is reported to have said: /boot/vmlinuz is a symlink to the real kernel file, I noticed both your linux distributons in LILO.conf pointed to it so I was wondering if the symlink had been removed somehow so lilo didn't know where to find the real Kernel Isn't it checking for this symlink in the HD that is specified for each boot setup? In any case, it still tells me No images have been defined Kirk If the /boot/vmlinuz symlink doesn't work with Lilo, then simply enter the full name of the kernel to boot with, e.g., /boot/vmlinuz-2.2.2-x Besides, even if the plain, general, vmlinuz, symlink worked, it'ld still be better to specify the entire kernel image name, because, then, when you work on reconfiguring the lilo.conf file using some tool, or even directly, you explicitly know what you're dealing with. Hence, regardless of whether a symlink like this did work, it shouldn't be used for this purpose, imo. mike
Re: [expert] Stop using kdm
Stephen F. Bosch wrote: Mike Corbeil wrote: Stephen F. Bosch wrote: Matt Stegman wrote: No, startx does not start kdm. startx will start X, and load kde by default, or another window manager - you can use any of several tools to change this. Qué? In Mandrake, startx gives you kdm. -Stephen- The relationship is indirect. startx will give you what ever it's been instructed to give you. One or the principal file for this is $HOME/.Xclients, or ~/.Xclients (~/ = $HOME/). Maybe people are confused about what kdm is? It's the login manager, not the window manager (that's KDE). -Stephen- I'm not sure that kdm is a login manager. I've seen it compared to gpm, but gpm is not a login manager either, because neither needs to run to login, as far as I'm aware. The only thing I recall about gpm is that it allows the use of the mouse in non X login, but doesn't prevent a person from logging in if neither is present. kdm is what was said to be used if using kde, which is a window manager and desktop. mike
Re: [expert] reconfiguring disk
Mage Grimau wrote: Which version of partition magic? 4 doesn't work right for 98 OR linux. 4 should work, because 3.0 did. Actually, I don't know if it worked for win98, but it did work for win95 and linux (ext2), as well as many other operating systems. However, I spoke with a computer store clerk last december who said that althought 98 didn't exist when pm 3.0 came out, 3.0 worked for his 98 system. 3.0 also worked for nt. I don't think you should have any problems using 4 for 98 and I'ld even use 3.0 for 98. Fire it up and read the documentation for 4. Are you considering purchasing pm 4 or borrowing it, because 4 is not the current release, which is at least 5.0; unless that's been backed out. You can also check PM's web site. The company is or was named PowerQuest, but if it's changed, then do a search on "partition magic". Should be easy to find. mike --- lorne schachter [EMAIL PROTECTED] wrote: I've got an interesting problem. 1 have a 2-disk system, 12G and 4G split as follows: cdrive - 8G Windows98 and 4G Linux ddrive - 3G Windows98 and 1G Linux. I boot Windows directly and Linux from a floppy (because I'm past the 1024 sector boundary?). What I want to do is to shrink the Windows partition (I'm only using about 2G of it) and use the rest as more Linux partition. I'm trying to run Partiton Magic to do this, but it keeps telling me that my partition table is bad. It might be, but that's the only place where there's a problem since the two OS's run fine on their own. The partition table is as follows (for the first drive) partition 1 0 - 0 - 0 thru 1023-255-63 partition 2 1023 - 0 - 0 thru 1023-255-63 Clearly, this information is bad, but it won't let me change it to more reasonable values. Any one have any ideas - my next step is to get ahold of the Power Quest folks for their help. Thanks, Lorne -- Lorne Schachter (732) 819-0460, (732)819-0460 (FAX) http://www.intact.com/~lorne = Mage Grimau, Strange Unwashed Somewhat Slightly Dazed VoiceMail/Fax: 1-651-328-1145 __ Do You Yahoo!? Send online invitations with Yahoo! Invites. http://invites.yahoo.com
Re: [expert] reconfiguring disk
Pj wrote: I use Windows version Partition 5.0 to re-partition on the fly. Unless Windows partition is full, you should be able to reduce the size of FAT or FAT32 and increase the size of Ext2. If you want to contact me privately I can try to walk thru it step by step. Do you mean Partition Magic 5.0? PM is an easy tool to use and although it was recommended to always make backups before resizing partitions, I used 3.0 many times for creating partitions, and resizing live partitions, without making any backups, and never had a problem. 3.0 was only really advantageous for resizing live partitions, though. However, I seem to recall hearing a couple or few months ago that PQ ported PM to Linux; therefore, before purcashing, check out this aspect, for anyone interested in investigating this tool. mike Pj [EMAIL PROTECTED] lorne schachter wrote: I've got an interesting problem. 1 have a 2-disk system, 12G and 4G split as follows: cdrive - 8G Windows98 and 4G Linux ddrive - 3G Windows98 and 1G Linux. I boot Windows directly and Linux from a floppy (because I'm past the 1024 sector boundary?). What I want to do is to shrink the Windows partition (I'm only using about 2G of it) and use the rest as more Linux partition. I'm trying to run Partiton Magic to do this, but it keeps telling me that my partition table is bad. It might be, but that's the only place where there's a problem since the two OS's run fine on their own. The partition table is as follows (for the first drive) partition 1 0 - 0 - 0 thru 1023-255-63 partition 2 1023 - 0 - 0 thru 1023-255-63 Clearly, this information is bad, but it won't let me change it to more reasonable values. Any one have any ideas - my next step is to get ahold of the Power Quest folks for their help. Thanks, Lorne -- Lorne Schachter (732) 819-0460, (732)819-0460 (FAX) http://www.intact.com/~lorne
Re: [expert] LILO problem
Marcos Dione wrote: On Fri, 14 Apr 2000, Mike Corbeil wrote: Lastly, this is truly a newbie question and should be asked in the newbie mailing list. don't mean to start a flamewar here... but don't try to discriminate... I mean, I don't think he's a newbie if he's trying to boot another kernel. I think a newbie would not know what the hell is a kernel... anyway, he made the question, so, don't complain about he MUST ask it somewhere else. The crucial underlying point is that many newbies are using the newbie list and in asking there, the answer will be made there, which means that many more newbies will benefit. I like thinking in global manners. I've subscribed to both lists and if a newbie question is more than basic, then I suggest they post to the expert list and therefore to subscribe to both. The question was newbie, but t'was not a crime committed. I'll refrain from proving the newbie part of the point and retract my harsh attitude. mike -- "No tire sus colillas en el mingitorio, las humedece y las hace dificil de encender" "Do not dump butts in the wc. They dampen and it makes them difficult to light." --Tom Sharpe, "Wilt on high"
Re: [expert] Follow-up - Multiple Linux systems on one hard disk?
Andrew George wrote: [snipped] May have already been answered, but in case it hasn't ... Heres a related question The following is part of my partition table Device Mount Point Distribution /hdb5 /boot Debian /hdb6 / Debian /hdc6 /boot Mandrake /hdc7 / Mandrake [snipped] For example...If I'm in Mandrake with dev/hdb5 mounted as /mnt/debboot should I specify the kernal location for Debian as /mnt/debboot/vmlinuz-2.0.38 and Lilo will set the MBR to go to hdb5 (even though it will mount hdb5 as /boot when the kernal reads Debian's fstab)? No. When you'ld boot into Deb, lilo would look for the kernel image in /mnt/debboot/ in Deb's / filesystem, because that's how you'ld have defined or spec'd it in your lilo configuration. If the kernel image for Deb is in /boot of that configuration, then you need to specify the kernel image and location as this for lilo. This means that either you need to copy your Deb kernel image to your Mandrake /boot directory, or create a symlink to it in your Mdk /boot directory. This sym link would be /boot/vmlinuz-2.0.38 - /mnt/hdb5/boot/vmlinuz-2.0.38 or /boot/vm... - /mnt/hdb5/vm... I think it'ld need to be /mnt/hdb5/boot/vm... The boot part does not actually exist in your /mnt path, except once /dev/hdb5 has been mounted. /mnt/hdb5 would be mounted to /dev/hdb5, however I'm not sure if /dev/hdb5 actually needs to be mounted. This would be easy and quick to test, but if you don't want to bother with trial and error, then simply mount /dev/hdb5 to /mnt/hdb5 (or vice versa - in wording) before modifying your lilo conf. If you don't use /mnt/hdb5 for the name of this mount point, then I'ld suggest not placing an entry for this mount in your fstab file. For example, if you name it instead /mnt/hdd, and you define the fstab entry as /dev/hdb5, then to mount /mnt/hdd to any other /dev filesystem, you'ld need to either type out the entire mount command, for mount to not use fstab, or you'ld need to modify fstab, when ever you'ld want to mount a different filesystem to this mount point. If you name it /mnt/hdb5, then you can safely define the entry in fstab, for /dev/hdb5. That's why I used this for the mount point name. Directories don't take up any space worthy of mention, especially when it's only the hardlink of the directory; therefore, you can create as many different mount points in /mnt as you want and then define these in fstab, probably with the noauto option, to avoid needing to type out the complete mount command each time, letting you mount this filesystem for example by simply running mount /mnt/hdb5 which is shorter than mount -vt ext2 /dev/hdb5 /mnt/hdb5 When using the former case, mount automatically looks in fstab for the definition of the mount point. F.e., I have two Linux configurations with different filesystems for /boot, /, /usr, /usr/local, /usr/src, /home, /var, and /tmp. One of these configs is the primary Linux config and the other is test/build, relative to the primary config; therefore, I created /mnt points named boot, root, usr, local, src, home, var and tmp (besides also a:, c:, floppy, and cdrom). Then I created fstab file entries for each of these, because when ever I say, for example, mount /mnt/boot I always want it mounted to the same partition. If I ever want to access my other single and small filesystem configuration, then I merely run mount typing out the full mount command, mapping it to /mnt/root, albeit could also mount it /mnt/boot or any other /mnt point which is not in use. e.g. mount -vt ext2 /dev/hdb2 /mnt/root If I want to access /boot of that config, then it's through /mnt/root/boot. If I want to access the /var directory, then it's /mnt/root/var, or for /root, it's /mnt/root/root. Based on this, you'ld need to create the symlink in your Mdk /boot directory, accessing the Deb /boot through /mnt/debroot/boot/... Add the fstab entry and you can mount by simply running mount /mnt/hdb5 and ls /mnt/hdb5/boot/vm* would show all of the vm* files in the /boot of Deb. This extra rap may be helpful in illustrating the use and flexibility of mount. Hopefully that's not too long winded. mike Trying to burn my boot floppies (and play with another distribution while I've got a spare Hard Drive) Andrew
Re: [expert] How to make hd.img installation see second hdd?
John Aldrich wrote: On Fri, 14 Apr 2000, you wrote: I don't think we're using the same hd.img file, because the only hdd partitions available with the hd.img I'm using are those on hda. This might only be a discrepancy created by the people at Planete Linux mag., because the cdrom I have was obtained with their mag, in a cellophane enclosed package (not from a Mandrake boxed set); and, if not by them, then by who ever created the cdrom for them. Mike: You might wanna just break down and get a "Cheapbytes" CD of Mandrake. It almost costs more to SHIP the blasted thing than it does for the CD itself! :-) Joh Extremely tight budget at the moment and for a while to come, otherwise, I'ld just go to a local store and purchase the package. I merely purchased the mag to get the cdrom, to take a look at Mdk (still want to upgrade my RH config, at least to explore or investigate it), but tight budget and have RH 5.1,upgraded to 5.2. Now, I'm working on the upgrade to 6.0, after which 6.1 and possibly 6.2 shouldn't take long to do. Currently blocked at start of 6.0, though, because of some incompatibilities related to libc.so.6(GLIBC_2.0 and GLIBC_2.1), libstdc++-libc6.1-1.so.2, and it doesn't look like I'll have an answer quickly. RedHat doesn't have mailing lists like this. They have a bug reporting tool, but this isn't a bug. I've been hoping for someone to know how to get around this hd.img problem not being able to see the second hdd. I don't want to switch the drives around, to avoid hassles that'll cause, but if I want to check out this version of Mandrake, then it's looking like this is what I'll need to do. I also don't have a backup system; therefore, turf's a little trickier. On the other hand, the maker of that cdrom (it's not from Mandrake from what I can tell, because it's labeled "Planete Linux, CD1, Avril 2000, No 6") could modify the hd.img file to allow access of the second hdd, send an email message saying where to download the hd.img file, and then I'ld grab it there. I don't need another cdrom; even if one way to resolve the problem, it's not necessary. All I need is an hd.img which will allow using the second hdd. This didn't strike me as a major item, if the maker could just make a corrected hd.img available for download. I don't know what goes into creating the hd.img file and don't have the sources for it, as far as I know. I looked at the lilo man pages, hoping to find something that'ld explain "oh ya, got a work around for ya, ...", but didn't find anything in the documentation related to such a work around. I was hoping someone here would know if this could be done by simply configuring a lilo for the boot floppy, to allow visibility of the second hdd for the installation program to see and provide access to. On the other hand, even if this could be done or attempted with lilo, perhaps the installation program portion on the boot floppy is hardcoded to only use hda. Shouldn't be, but could be. Based on that, I don't see how lilo could help with this problem. Anyway, the distribution is a single cdrom, doesn't look like it was created by Mandrake, and this stupid hd.img problem is a pain in the backside. And, I have no idea why someone would want to create a hd.img which refuses to see other local HDDs; unless an hd.img can't be made to allow use of more than one hdd. Input on these types of topics might be helpful here. Even if it doesn't help resolve my particular hd.img problem, the knowledge would be interesting. mike
Re: [expert] reconfiguring disk
Mage Grimau wrote: I have PM 4 that complains that my partition table is corrupt when I've installed either Linux (Mandrake) or Win98. It's quite happy with NT4 and Win95. PM5 has no problems with my Linux/98/NT4 setup, but I find I don't need it any more. Installing 98 then NT then linux made em all happy. Snipped out all the old stuff. You say that you don't need PM anylonger. What about resizing used partitions? On the other hand, am a fellow of very little financial means; therefore, my 9GB system spread over 2GB and 7GB HDDs is more limiting than what many new systems are coming factory shipped with these days. I guess if someone has a lot of HDD space, then configurations for various OSs can be made large enough that resizing active partitions is never necessary. mike
Re: [expert] LILO problem
Kirk McElhearn wrote: Actually, I figured this question would not get many answers on the newbie list... You may be correct, but it would be useful to other newbies, because many or all newbies will need deal with that kind of question. On the other hand, asking in the expert mailing list also isn't a crime. I've only been using these mailing lists since this past week and kind of like the idea, personally, that at least some more seasoned people pay some attention to questions asked in the newbie list. Maybe I have too much of a bleeding heart for newbies. Of course no one is obligated to do so, and it'ld make more emails to delete, but I don't have a problem doing that. If I see a question or thread I'm not interested in or can't answer, then Delete is quickly clicked. Some people, however, get annoyed by overwhelming amounts of email, but if subjects are sufficiently clear, then more email is less of a problem or nuissance. Clear subjects, though, are therefore important. You say LILO and I know what you mean. If you said , then I wouldn't have a clue, but might read just to find out what is about. If, however, you say the name of an application which I have no experience with and no interest in, then Delete, and zap it's gone. This is one reason why I'm following both threads, however won't be for long if I don't get this cdrom installed. Definitely not a rich man, and I'm not prepared to swap the drives. Laziness is part of the reason, but it's mostly because I'm not interested in dealing with the hassles swapping will bring, at least not now. The rest is merely a story I'm not going to bla blab about. I apologize for being too harsh, though, even if you don't seem too bothered. My global perspective "got the better" of me (I think that's the correct expression). mike
Re: [expert] LILO problem
Kirk McElhearn wrote: On 14/04/00 19:45, Stephen F. Bosch [EMAIL PROTECTED] is reported to have said: OK, I have finally managed to get two distributions installed, and to get lilo to work. But when I try to tweek the lilo.conf, I get a message that each of my choices is added, but then it says No images have been defined. It does not update my changes in the mbr. What could I be doing wrong? Show us your /etc/lilo.conf, man. boot = /dev/fd0 If you want your changes to go to the mbr, then the above line must be changed to /dev/hda, which is where the mbr is located. The way you have it, it'ld used /dev/fd0, which is the floppy. timeout = 100 prompt message = /boot/message default = dos vga = normal read-only map=/boot/map install=/boot/boot.b keytable=/boot/pc-dvorak-latin1.klt image = /boot/vmlinuz label = mdk2 root = /dev/hdb11 image = /boot/vmlinuz label = mdk root = /dev/hda6 other = /dev/hda1 label = win table=/dev/hda other = /dev/fd0 label = floppy unsafe Kirk vice versa Translations - French to English, English to French | Technical Writing Traductions francais-anglais, anglais-francais | Redaction technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [EMAIL PROTECTED]http://www.mcelhearn.com Kirk McElhearn | Chemin de la Lauze | 05600 Guillestre | France
Re: [expert] Unzipping a file
Ron Stodden wrote: Brian, tar can interface directly with bzip2 as needed: tar -cvf archive --use-compress-program bzip2 * tar -xvf archive --use-compress-program bzip2 It all works beautifully (except that bzip2, as you would expect, takes a very very very long time to do its job). Does bzip2 take any longer than gzip? Also, that's one way to do what that does, however the following may be a little easier to write tar cvf archive bzip2 archive and bunzip2 archive.bz2 tar xvf archive The - isn't necessary for cvf and xvf, albeit because most programs require at least one -, it's easier for newbies to remember to use the - with tar, than to not use the -. I don't know of any other programs or utilities which don't require the - to specify options, albeit do like the idea. It's kind of like the Perl saying "there are many ways to do it" (it being what ever). Otherwise, what's the comparative report on the differences between bzip2 and gzip? I think that it was Brian who said that gzip produces slightly less compressed sizes, but other than that up not up on the comparative stats for compression tools. What ever happened to cpio? This one seems to have been the least popular, but worked fine for me when I last used it. From what I recall, cpio was the alternative to tar on Unix platforms, and was thought to be sometimes better than tar. mike
Re: [expert] How to make hd.img installation see second hdd?
Kirk McElhearn wrote: On 14/04/00 9:31, [EMAIL PROTECTED] [EMAIL PROTECTED] is reported to have said: Does anyone know if there's a way to make installing Mandrake from the second hdd, hdb, work? I'm using a boot floppy created with or from the cdrom hd.img file, but when booting with this diskette, the installation program only sees hda. I just installed a second version of Mandrake on my second HD. When you get to the DiskDrake section, you choose which HD to use. Kirk I don't think we're using the same hd.img file, because the only hdd partitions available with the hd.img I'm using are those on hda. This might only be a discrepancy created by the people at Planete Linux mag., because the cdrom I have was obtained with their mag, in a cellophane enclosed package (not from a Mandrake boxed set); and, if not by them, then by who ever created the cdrom for them. We probably are using different installation programs and therefore a different cdrom, because as explained in what you snipped out, when doing the install from the cdrom and the installation gets past package selection and everything before, the actual installation bugs up or out with a stupid failure due to trying to install to var, usr, etcetera under /mnt, instead of to /var, /usr, etcetera, and, again, this appears to be caused in the install2.pm script or Perl package module. After examining the perl-install programs, I'm certain the same thing will happen with a working hdd install, but I'll be able to modify these scripts to work (I can't do that with the cdrom). I've been around a while; therefore, I'm not a newbie newbie, not even wrt Linux. mike
Re: [expert] It doesn't work - Multiple Linux systems on one hard disk?
Kirk McElhearn wrote: [snipped] Then I tried Lilo, and that doesn't work either. I have the following: hda1 is my win partition hda6 is my Mandrake 1 partition What's a Mandrake 1, 2, ... partition? Actually, you explained 2 below, kind of, but even that's not clear to me. What's this 1 and 2 represent? These two can boot fine hdb1 is my Mandrake 2 /boot partition hdb11 is my Mandrake 2 / partition Lilo won't accept anything I try for the second disk. I have tried: image=/boot/vmlinuz root=/dev/hdb1 Your root partition is /dev/hdb11, instead of /dev/hdb1, which is your /boot partition. label=md I thought this would tell it to look in the /boot partition and find the right kernel, but no dice. I get a message saying there is no init. The init is in your / partition. Also, if that's all you have for configurations in your lilo.conf file and you plan to make this the information for lilo booting from the hdd, then you should add your other OSs or configurations before running lilo. mike
Re: [expert] How to make hd.img installation see second hdd?
[EMAIL PROTECTED] wrote: comment: i have installed mandrake 6.0, 6.1 and also 7.0 directly from hdd hdd (it´s no dupe, it means 4. "harddiskdrive" or secondary slave) with no problems. i always used self-burned cdrom from downloaded iso-files. I think that either you snipped out some of what I had written, or I thought to have written it and didn't. The reply to Kirk will explain. mike heiko Mike Corbeil wrote: Does anyone know if there's a way to make installing Mandrake from the second hdd, hdb, work? I'm using a boot floppy created with or from the cdrom hd.img file, but when booting with this diskette, the installation program only sees hda. Comments? mike
Re: [expert] Follow-up - Multiple Linux systems on one hard disk?
Jean-Louis Debert wrote: Mike Corbeil wrote: Actually, I guess that the boot directory doesn't need to be named /boot, as long as the correct path is specified in the boot configuration. This may be incorrect; however, the reason I say this is because linuxconf lilo boot configuration requires the name of the kernel image file as well as the directory path, at least when linuxconf isn't run in or from the /boot directory, or the directory the kernel image(s) are located in. Otherwise, if /boot is strictly required, then I don't know why linuxconf would require for the directory to be specified. You're perfectly right: lilo only needs to find the kernel image, and you could often, in theory, put that anywhere as long as you specify it correctly in lilo.conf What I was wondering is if not only the vmlinuz kernel file was sufficient, or if I'ld need to create symlinks to all of the /boot files associated with a kernel existing only on another configuration's partition, which is invisible if not mounted. For example, assume the following: Linux config #1: /dev/hda5 - /boot (vmlinuz-2.0.36-3, only) Config #2: /dev/hdb2 - /boot (vmlinuz -2.0.34-0.6, 2.0.36-3, only) Config #3: /dev/hdb5 - /boot (vmlinuz -2.0.36-3, and 2.2.x, only) /dev/hdb6 - /root Config #4: /dev/hdb13 - /boot (Mdk 7.0.2, vmlinux 2.2.y) /dev/hdb14 - /root (Mdk 7.0.2) Configs #1-3 are RH. Now, let's say I want to configure and run lilo from on config #4, say using linuxconf. Can I simply mount /boot for configs #1-3 and create symlinks in #4's /boot directory to the vmlinuz kernels in the /boot directories of configs #1-3, and then run linuxconf successfully, without a glitch? That's what I meant, that is, that this would work and that symlinks to the system map files would not also be needed. I think that I was just trying to be safe, before actually doing this. On the other hand, that was probably just a little nervousness, because linuxconf doesn't ask for the other files; only for the kernel image file itself. It may be necessary to also mount the / filesystems, but that's not important, at least not wrt the question I had in mind. Hence, between your reply and this additional thinking, while recognizing that I was just a little nervous about this, I think this is the answer, that is, other than getting rid of some of my Linux configurations. However, until I get an hd.img file for the Mandrake 7.0v2 cdrom I got with the Planete Linux mag, I only have one unnecessary configuration. Once I get this Mandrake cdrom (not made by Mandrake itself as far as I can tell), then, well, I may still only have one unnecessary Linux configuration to get rid of, if I don't get rid of any between now and getting this cdrom installed. I prefer having at least a couple configurations from each Linux distrib installed. However, the /boot directory is part of the File System Standard (FSS) in linux, and this standard has been elaborated for several reasons: .. mainly to aim for better interoperability between different Linux distributions. .. in the case of /boot, this also provides a suitable mount point to mount a different partition. This is mostly used for older BIOSes that can't deal with the 1024-cylinder limit, so that you can make sure that a boot kernel will be physically below the limit. Was aware of most of that. Wasn't thoroughly aware or remembering the first .., though. However, in following the general guideline, I make /boot a separate filesystem or partition, and do this regardless of the filesystem being above or below the 1024 limit. The reason for this generalized approach is merely to make sure that there's more flexibility for future modification to the system, and I don't think it can hurt to make /boot a separate fs, even if the entire configuration fits below cyl 1024. I'll experiment with creating only symlinks to the vmlinuz kernel images and omitting do the same for the /boot system map files. Doing that would probably only be unnecessary extra. Part of the intent of the question, though, is that some programs seem to not accept symlinks, and I don't recall which program I recently encountered this problem with, but this happened within the past couple or few days (am spread over many, varied tasks and the name of that program is currently, well, absent - still haven't gotten any sleep, been up all night working on various aspects of my Linux configurations, so am zonked, deep fried, right now). Thanks anyway, mike
Re: [expert] LILO problem
Kirk McElhearn wrote: OK, I have finally managed to get two distributions installed, and to get lilo to work. But when I try to tweek the lilo.conf, I get a message that each of my choices is added, but then it says No images have been defined. It does not update my changes in the mbr. What could I be doing wrong? Kirk Provide the contents of the lilo.conf file you're trying to use, and the contents of the /boot directories of the Linux configurations that lilo is supposed to be able to boot. For the /boot listing, you only need to include the name of the kernel image file and symlinks, if there are symlinks. These are typically /boot/vmlinuz... Also: - give the or list the /dev/hd??? files mapped to /, and - what tool are you using to tweek the lilo.conf file? There are at least a couple capable tools; therefore, it may be helpful to know which (I only know linuxconf, whereas others apparently use other tools). Lastly, this is truly a newbie question and should be asked in the newbie mailing list. My hd.img question would resemble a newbie question, except that I'm also asking how the hd.img file for the cdrom I have can be modified or made to work with the second hdd, which it's not permitting to use, currently. This goes beyond what newbies can easily or quickly find worthwhile and authoritative documentation for, particularly because I don't have the sources to create the hd.img file and would prefer avoiding this (will probably just trash the cdrom, instead, albeit maybe not). Am "simultaneously" working on upgrading RH 5.1 to 5.2 and beyond, as well as trying to find a way to install this cdrom for Mdk 7.0v2 (probably or perhaps not worthwhile, but will see); and, of course, there are many other things to do; therefore, am seeking what these mailing lists are supposedly for, i.e., assistance. You should read the documentation on lilo, lilo.conf, and the tool you're using to configure or define the lilo.conf file. You could edit it directly and then simply run lilo, but the gui tools available are certainly adequate and easy to use. Plus if there's anything else which needs to be done related to lilo configuration, then these tools will take care of this business, permitting the user to get on with other more useful things to learn. I never had to ask anyone how to configure and install lilo, albeit would be interested in knowing if it would work for this hd.img problem I'm having. Some reading might turn up the answer, but I already went through the lilo man pages and this wasn't fruitful for this particular task. Don't really like the idea that lilo configuration requires visibility of kernel images, though. Always someone to tell you can't do something which shouldn't hurt anyone except novices and experienced people who are asleep when doing their work. Anyway, if you want help, then you need to provide more information. mike
Re: [expert] Unzipping a file
Wayne Petherick wrote:
How do I unpack a file with a .tar.bz2 extension?
Read the documentation. There are man pages for bzip2 and tar.
That's the general recommendation for what to do, [first]. However,
I'll give you a little rap session on tar. I haven't used bzip2 and
bunzip2, yet, because I mostly work with gzip and therefore gunzip;
however, the tar part I've deal with before, and it's easy to learn from
the man page.
Tip on unarchiving tar files or tar balls: The command is (drumroll
please) ...
% tar xvf {filename}.tar
If it's the only tar file in the directory, then the following will also
work (drumroll, again, please) ...
% tar xvf *.tar
If you want to view the contents of the tar file or ball without
unarchiving, then use tvf, instead of xvf. x = extraction and t= toc
(table of contents).
If you unarchive, delete the .tar file, and then decide you need to
re-tar, then copy the files to a separate, empty directory and run the
following
% tar cv * -f {filename}.tar
{filename} of course being what ever you want to name the file.
For bz2 decompression and compression, refer to the man page on bzip2:
% man bzip2
The bzip2 executables or programs should be in one of your bin or sbin
directories, either under / or /usr. If you don't find the bzip2
programs, then you'll need to install them from the cdrom, and if not
there, then download and install.
There are other tools which can be used and kpackage might work for
this. I believe that it works for .gz files (not entirely sure, though,
because I haven't explored kpackage much, yet); however, it does work
for .tar archives and rpm (supposedly for rpm anyway).
If you're using kde and want to investigate this alternative, then bring
up kpackage and try it against your .tar.bz2 file.
That's the fairly comprehensive pov from arrogant moi. Just kidding,
i.e., joking. Don't mind me, I joke once in a while; although, it ain't
no joke that you didn't bother reading the man pages, first. Tsk, tsk.
If you want to become at all fluent with Linux/Unix, then you need to
use the reference documentation and learn about all there is and how to
access the different forms or formats. There's man, xman, info, as well
as kde Help, and other tools. After learning the basic Linux/Unix
commands, like cd, pwd, cp, rm, mv, cat, echo, id, who, whoami or "who
am i" (for those who forget who they are - joking, kind of), ..., the
next first lesson in learning to use Linux/Unix systems are the
documentation resources.
mike
Re: [expert] You people just don't get it
Alen Salamun wrote: Mike Corbeil wrote: Hello. My name's Mike, aka Frank, as you may discover. Hi Mike! :) You written us a really good "whole look" on the distribution. And yes, the only thing I was upset was, that even when "bugs" are reported some people have xxx excuses not to even investigate them... That may only be because Mandrake is not a HUGE company and therefore doesn't have enough staff to get to everything immediately. This is my hunch and because this is the pov I take, I would simply look for the desired utility and download it. Because Mandrake is highly compatible with Red Hat, if Mandrake doesn't have the utility for download, then I'ld check Red Hat's download directories and give that a try. People need to realize that Linux distributors aren't huge like MS and even that company markets buggy OSs; therefore, a little good will or patience on the part of everyone wouldn't be misplaced or out of place, imo. When you get the utility and it doesn't work, because of some error within the code of the utility, then that's a bug. There's definitely a distinct difference between what constitutes a bug, and merely a missing utility. And a "BUG" in a distribution... Well, if you want to conceptualize bug to that level, then I guess it is a free country. I understand what you mean. There is no such thing as "MISSING UTILITY REPORT FORM" for mandrake OR? :) There is only "BUG REPORT FORM" :)) Good point and this is something Mandrake should add. I understand all of the following, I am using LiNUX for more than 5 years now, and I have contributen many many things to the community. I even offered my help to Mandrake and am translating the distribution into Slovenian language. But there is this thing, that tells me, that if someone will say: "Hey this is not translated right", I will work on the problem, or atleast say "Thank you" I have it it my TODO list... I kind of hunch that Mandrake would place your point about a missing utility on a todo list, too. Based on the list of outstanding bugs, or reported bugs, Mandrake does have todo lists. mike
Re: [expert] Wine Is Not an Emulator
Mike Fieschko wrote: [snip] That is (was) WABI. Caldera were distributing WABI v2.2, and maybe still are. I believe MS Windows 3.11 apps (some) were the last ones supported. Yes, you're right. WABI now rings a historical bell. MS Win3.x apps were all I needed the emulator for, and I'ld still have no other real use for it, unless imposed by the environment on some job(s), which may be the case for some people. It definitely was a useful emulator, though. mike
Re: [expert] You people just don't get it
Alen Salamun wrote: Burkhard Zombronner wrote: Could you please stop shouting on us !!! Hi! Ich bin nicht "schouting" an euch...heheh (I am not shouting on you all) Only on people, that have a kind of funny approach like "...don't use it, don't care...". I have no intetions do shout on anyone, but this really god me down in my LiNUX heart...:)) Really... Yes, people could avoid being overly sensitive, and as far as I'm concerned, your devotion or astuteness is of a worthy sort. However, I also need to admit that your intentions weren't initially understood. Now that they are, you're correct and I have no problems with your initial comments anylonger. mike
Re: [expert] OT: Perfect operating system
On Thu, 13 Apr 2000, Deryk Barker mewed: Thus spake Brian T. Schellenberger ([EMAIL PROTECTED]): AmigaDOS was the perfect operating system, but Unix/Linux is a close second . . . Sorry folks, but Multics was the perfect OS. All others are imitations. From recent readings, some are saying that HURD is currently the best, or best basis for the future best. As for AmigaDOS, I'm not sure. Never worked with it, but have worked with MS and PC DOS, and definitely prefer Mac. Can't really comment otherwise, because I don't have enough knowledge of these other systems. VAX/VMS was considered very good, though. However, this is a Linux mailing list for getting assistance with Linux, in particular Mandrake's distribution. mike
Re: [expert] Follow-up - Multiple Linux systems on one hard disk?
Actually, I guess that the boot directory doesn't need to be named /boot, as long as the correct path is specified in the boot configuration. This may be incorrect; however, the reason I say this is because linuxconf lilo boot configuration requires the name of the kernel image file as well as the directory path, at least when linuxconf isn't run in or from the /boot directory, or the directory the kernel image(s) are located in. Otherwise, if /boot is strictly required, then I don't know why linuxconf would require for the directory to be specified. Is there are slight inconsistency or discrepancy in this sense, or is there a valid technical reason I'm not aware of? Anyone know? mike
[expert] How to make hd.img installation see second hdd?
Does anyone know if there's a way to make installing Mandrake from the second hdd, hdb, work? I'm using a boot floppy created with or from the cdrom hd.img file, but when booting with this diskette, the installation program only sees hda. The cdrom I have came with a planete linux magazine, april/may 2000 issue, and the cdrom can't install, because the install program constantly tries to install or access /mnt/usr, /mnt/var, etcetera, instead of /usr, /var, etcetera. I tried all installation methods (Recommended, Custom, and Expert) from the cdrom, and the "same" /mnt problem occurs each time; different locations, but always after the actual install of the packages is begun. Hence, I created a /Mandrake directory containing the entire cdrom on hdb (not enough space on the 2GB hda) and inspected the perl-install scripts to try to find the culprit and it looks like it's probably install2.pm. In the sub main of this script, if $::testing evaluates to false, then the directory prefix to be used is set to "/mnt". There are a few other possible scripts that could be causing this problem, but install2.pm is the only one I found to actually define the prefix to /mnt. Based on this, I'ld like to do the install from the hdd, modifying the aforementioned script(s) until the problem is resolved, but hd.img doesn't permit access to hdb. I contacted planete linux a few days ago, but haven't received a reply, yet. The only options I see are to either: a) get an installable cdrom from planete linux, which I doubt they'll want to send; or b) have planete linux or who ever created the hd.img file with this lousy restriction modify this .img to allow access to hdb; or c) create the necessary hd.img file myself (don't know how to do this, at least not yet, and don't know if it can easily be done); or d) try something with lilo, creating a lilo boot loader stored on the boot floppy, but am not sure this would permit getting around not being able to use hdb for the /Mandrake directory. I could switch the two HDDs around, to then be able to see the filesystem the /Mandrake directory is located on; however, I foresee some potential problems with this approach, or at least headaches. Plus, I don't see or understand why hd.img shouldn't allow use of hdb. Of course, I also don't understand why the installation wouldn't have been tested as per the real end user's world, prior to shipping the cdrom. This isn't the first time I've encountered this kind of problem, though, and the other time was in 1997, for a Linux distribution I paid $100 for; has never been installable and don't even know where that piece of garbage is, now, and may have thrown it out to the garbage, after the distributor refused to send an installable cdrom. Comments? mike
Re: [expert] Unzipping a file
Brian T. Schellenberger wrote: In the vain hope of forestalling a number of "you moron" followups, I will point out that I saw the recent posting explaining that there's a "y" option that does exactly this. In my defense, it's not in the "man" page. man or documentation page bug. y option for what? I just checked the man pages for tar and bzip2 and also didn't see any mention of a y option. Yeah, I'm more a Unix person than a Linux person, so I checked "man" rather than "info." But I checked "info" just now and it's even worse: It claims that the "I" option does bzip rather than the "y" option. Why not just use bzip2 and tar separately? Or are you talking about some other tool? Sounds like the option -I would be related to tar, and if this is the case, then I'ld just stick with using the tools separately. PS: Does anybody know of a utility that will translate info into man format? Am unaware, of such a tool. Just checked and didn't find any for info to man, but did find pod2man; therefore, if there's a way of getting info converted to man, then it would probably be through an indirect route, perhaps info to dvi or ps, then dvi or ps to pod or man, etcetera, for example. info might also have this information. You might find an answer in the texinfo documentation, which you may or probably have a directory for under /usr/doc. Am not certain if texinfo is related to info, but this documentation should quickly answer this additional question. Actually, I just checked and there isn't much to speak of there; however, the LDP should have documentation. You may find the following documentation of some use, maybe: man mkmanifest (top where it explains - somewhat - how this man page was created) info texinfo Nonetheless, does bzip2 or bunzip2 not work for un-bz2'ing your file? If you're using kde, then have you tried kpackage or karchive, or some other kde utility or tool? "Brian T. Schellenberger" wrote: I find .tar.gzip and .tgz files much more convenient becuase of the integrated support for the gzip: tar xvzf foo.tgz does the whole thing. Yep. The only reason I haven't used this more than once is because I already knew tar and gzip, don't need all tar files zipped, don't do tar'ing and zip'ing an extraordinary amount, and using the tools separately means needing to remember one less option. If I used both functions combined, often, then I'ld use only tar with xvzf. Alas, dude, it's mostly a free world. But the answer the original question, the easist way is bzcat foo.tar.gz2 | tar xvf - Depends on what the person wants to do and it's .bz2. Besides, that's really no easier than the following, or at least not so much easier that it's worthy of note: % bunzip2 foo.tar.gz2 ; tar xvf foo.tar (or running the two commands from separate prompts) Showing newbies the many various ways of running commands as per the Unix ways is good, I guess; however, if people are going to complain about a nano-second of difference in time, then go ahead. At the end of the week, you'll have saved a whole few nano- or micro-seconds. I mostly prefer the practical approach. If you want to sit at work all day trying to figure out how to make your dog do all sorts of tricks, then go right ahead. Until you're accountable to me, I don't care how you spend your time, and I've never worked in such (permissive) environments. Besides the trick I prefer to teach dogs is to stop barking unnecessarily. From one moron to another, mike P.S. Huff puff and the big bad wolf blew the little piggies' house down. Rude wolf, I'll say.
Re: [expert] Need someone to point DNS
This is kind of an off beat question, with a little teasing added; however, am I correct in thinking that when you people say IP, you're actually referring to ISP, instead of IP? As I'm sure most of you already know, but some real newbies may not: IP = Internet Protocol whereas ISP = Internet Service Provider There's a major difference in the meaning and relatively few people ever deal with IP in any way other than (totally, or at least very) transparently, whereas ISPs are businesses and therefore not dealt with or used in a manner as transparent as IP. Little sarcasm added just for a touch of light spice, of an "haachum(e)" pepper variety (but no harm intended). mike Pj wrote: If I understand this subject correctly there are two issues: the host and the account holder. They are not necessarily serviced by the same IP, and do not forget: IP's are not all created equal. If you have a hosted, active, paid DNS account anyone anywhere in the world should be able to type in www.yoursiteneame.com and see your page in all it's glory. However there is a caveat writtn in fine print in most IP-Client contracts: It's called site content. The IP agrees to host your site as long as the the content is not purient, and doesnt promote antisocial behavior on the site. Most IP's reserve the right to cancel the account, and delete the page from the server without warning under the conditions as set forth in the contract. SPAM and XXX sites are types of accounts that usually get removed from the server quickly. If this is a new account or if your are moving your business from one IP to another IP time can be a big factor. One IP didn't set up my client's page for almost 6-months even though the IP received the NIC approval within three days after the name was approved. Conversely, because of extenuating circumstances-and with the help of NIC and the new IP-a friend was able to move his very large and active business account and be active in less than 24-hours. I've had accounts under some incredibly bad IP's who simply knew nothing about web page hosting set-up but were too ego-centered to admit it. In short, the IP that hosts your webpage should know how to set up the browser to point to your page and service page requests properly. If your acccount is new, your host should be able to tell you how long it will take to set up your account on his server once he receives notice from NIC or ICANN. My IP will usually do it within 24-hours. If is usually convenient to host your page with a local IP. However, it is not a requirement. You should be able to see your page from your home, office or from a hotel in another country as long as you can get on the web. If you cannot see your page, then you need to find out why. Pj
Re: [expert] Wine Is Not an Emulator
I thought WINE was what's been used on Solaris for several years, but may be confusing it with another MS Windows emulator for Unix, or Solaris anyway. mike John Aldrich wrote: On Tue, 11 Apr 2000, you wrote: Lucky you about the only thing I have been able to get to run under wine is Forte Free Agent 32bit Haven't tried ityet. :-) That's one reason I keep my Windows machine around. :-) John
Re: [expert] What is the thing not installing GCC AND EVEN MAKE in anything else than developer?
If the server in question has enough disk space, the server does not need to be constantly up, and the the person who posted the question of this thread does not have a second machine to use, then this person could and perhaps should consider installing a second, separate, configuration on the same machine, for doing builds and such work. Then, when builds and such need to be done, shutdown the server, boot into the development configuration, do the coding or code changes, builds, and tests, and copy the files to the appropriate locations on the server partition(s), and reboot into the server. However, if the server must be constantly, or near constantly, up, then this won't be a good solution, because this kind of work can require considerable time. I agree that these tasks should be kept out of a serious server configuration. Even if it would or could be difficult for a hacker to break into the server and do damage using make and or running compilers on the server configuration, the added security of not having these tools accessable at all definitely makes much sense, at least for serious environments, and serious environments can usually afford a second machine for this kind of work, especially when considering PCs. If these tools are available, but there's no source code available, then this might help to decrease risk, especially if people using the server cannot upload to the machine. The least best choice, imo, is to make these tools available on the server, within the server configuration; unless as per the last above paragraph. However, if this is the only choice, then perhaps there's a way to allow only root and perhaps some special user account to have access to these tools. If there's no way to prevent people using the server from having access to these tools, then monitoring would need to be done "microscopically" using yet another daemon. Or, the server administrator sh/could install, do the builds and testing, and then uninstall, these tools, on an as needed basis, instead of leaving these persistently on the server. If this kind of attack on the server is not a concern, then advance at your own risks. If no hacker penetration ever occurs, then great; otherwise, "live and learn". On the other hand, perhaps it is possible to install these development tools and leave them on the server configuration, while assigning these tools [strictly] to a specific group, e.g., "developer", making sure that the server does not belong to the developer group, creating a special user account for doing development, making this user part of the developer group, and while the server is up, login as a developer to do this development work. However, this would probably be better using a separate machine, to login to the server as a developer user, and if the person who posted this thread has the ability to do this, then this person should consider making that second machine the second machine already refered to, which would leave the server strictly a server. Server and developer or development platforms are not necessarily synonymous. A development server is, however file servers and isp servers, for example, are not. In this sense, it might be useful to know exactly what type of server is at the center of the question of this thread. Servers can be used as central to or strictly for development, but this probably isn't what the most typical use is. Servers are usually thought of as for file servers, isp servers, and database data servers, for example; however, servers can also be used for development tools. In this sense, it might be useful to have a little more clarity on the exact nature of the server this thread's about. mike P.S. A little long winded, eh. Alen Salamun wrote: Charles Curley wrote: Yes, they are the most common utils in the Unix world. Which is EXACTLY why you don't want them on a server. If a cracker were to gain access, would you want the cracker to be able to compile for your computer? For proper security, do your own compiling on some other computer and copy the executables over. If you must compile on your server, install the compiler as needed and remove it when you are done. Hi! Ohhh what a smart thing...It is SO HARD for a cracker to compile his hacking/cracking tool on other machine and move it to this one...Expecially on x86 architecture that is so uncommon... I have (almost) never saw a computer without c compiler on it...This is like having a screw-driver in a trunk of a car and saying, it will help a burglar to brake in...He can bring his own with him Bye, Alen -- *---* *E-Mail: Alen Salamun [EMAIL PROTECTED]* * LiNUX - The choice of GNU Generation! * *---*
Re: [expert] You people just don't get it
See below. Alen Salamun wrote: Hi! Hello. My name's Mike, aka Frank, as you may discover. I am really SUPRISED to see such replies like: ...I don't use it, I don't care... ...This should be fixed by configuration alone... ...mrproper should be run... Ya, that stinks for a reply, but it helps to show that although this is supposed to be an expert mailing list, well, not all using it are experts. Eh! Or, maybe the experts don't care to share their knowledge. Hey THERE ARE NOVICE USERS TOO! Not all are experienced in all this stuff and Mandrake should be USER FRIENDLY TOO, expecially for beginners! Again, this is the expert mailing list, albeit even experts are on a learning curve in some respects. Don't expect GOD, here, because you probably won't find Him, here. It's one thing to specialize as a systems administrator, or a developer (in some set of tools and languages, albeit no one knows all), etcetera. However, when installing Linux, doing the administration, setting up servers, and doing development, then there's much more to know than in typical niche jobs in large companies. It's important to remain realistic, always and in all ways. And I AM TOTALLY TOTALLY CONSEARNED BY THE FIRST ONE! I DOESN'T MATTER IF YOU DON'T USE IT, OTHERS ARE! Good point and friendly approach, that is, what you're suggesting or implying, and I like friendly people. I despise stuck up people, who remind me of something that's better left unstated, here. (I'm a people person and therefore believe in "people building".) And bugs like missing utility, that is not missing in any other distribution (even not in .tgz) are NASTY ONES! T'is not a bug. T'is only a missing utility. When you get the utility and it doesn't work, because of some error within the code of the utility, then that's a bug. There's definitely a distinct difference between what constitutes a bug, and merely a missing utility. For example, do you consider MS Windows NT Server 4.0 or Windows 2000 Server buggy after installing it and it lacks MS Office, compilers, etcetera? I'll bet 100:1 (of a non-existing money) that you don't. Take it easy here, this is Linux and Linux definitely doesn't cost the price of Windows NT Server, let alone Windows 2000 Server. I visited a little computer store in Quebec, recently, and Windows 2000 Server costs (if memory serves me accurately) around $5,000Cdn for six users, and $3,000Cdn for five users. Conversely, Mandrake Linux, a version which can be used for development and server, cost around $80Cdn. Hence, get a grip on reality, here, now. That's one of my prescriptions or recommendations, and you can send me $50Cdn for the prescription, which comes without pills. I can understand that you find some things don't work with Linux and or certain Linux distributions, but don't forget that Mandrake is not charging thousands of $$ for it's Linux distribution. You pay less for Mandrake and other Linux distributions than you would for MS Windows 2000 Workstation, and it doesn't come with all of the server capabilities as the many Linux distributions do; therefore, take it easy. Otherwise, create your own distribution, which is something I'ld definitely like to do. There are many good things said about Linux and some web sites like Network Computing, PC Magazine, ZD Net, and CNet, for example, contain articles from people who've examined Linux distributions to various extents. http://www.networkcomputing.com http://www.pcmag.com http://www.zdnet.com http://www.cnet.com There are also other locations and I think Information Week is one, but don't know the url off of the top of my head. I agree that some things don't work with some Linux distributions; however, I like to try to keep a real, clear, objective, sound perspective about such things, especially when the alternative (MS) solutions would cost MUCH more (which I can't afford - wallet's paper thin, like as in one sheet of thin paper, ya know what I mean?). Therefore, another prescription is PATIENCE, mon vieux. Know what you're dealing with and if you're not prepared to deal with it, then leave it, instead going for something else. I certainly don't expect the ride to be entirely smooth, however when comparing costs, well, I'll go for this ride any time. The little extra smoothness MS might provide isn't worth the astronomical difference in cost. Also, remember or learn about the history of Linux and how the distributions came to be. Linux was not developed as some capitalistic means to quick wealth. There was major amounts of work done by many people around the Globe and for free. There's much to be appreciated, but some problems therefore, inherently, need to be expected. If you're entirely intolerant, then you either need to move on to another OS, or take time to tranquilize the nerves, to adjust your expectations to what you're dealing with, realistically. Play it as if you're part of a symphony, and all
