afunke

There is a spelling mistake above ("noapci"). Perhaps you wanted to write: noapic (apic stands for "Advanced Programmable Interrupt Controller") OTOH, even if you turn off ACPI in BIOS, the Linux kernel may try to turn it on again if it detects a ACPI capable MB, so you need to provide the following command as well: acpi=off (acpi stands for "Advanced Configuration and Power Interface") A third possibility is turning off local apic: nolapic And you may combine 4 or more options: linux noapic nolapic acpi=off nodma Andreas

That's correct. The eth0 interface is already active by the time the ppp0 interface is activated by kppp, and pppd will not set the default route to the Dialup modem interface if it finds an already existing "default" route pointing somewhere else. Check the route by running the following command as the root user: route You'll be able to see several routes, and one of them is probably similar to: default 192.168.1.1 0.0.0.0 UG 0 0 0 eth0 You may fix the problem (temporarily until next boot) by running the following command, as root, BEFORE you connect to your ISP: route --del default The permanent fix is to create a script that runs the command above just before KPPP dials to the ISP, make that an eXecutable script and either call KPPP from it or "teach" kppp to run it before connecting using kppp's configuration gui/wizard. This can be accomplished by typing the following script in a text editor, saving it as "Internet.sh": #!/bin/sh route --del default Make the script executable, and give the script superuser rights with the following commands, run as root: chmod +x Internet.sh chmod +s Internet.sh Configure Kppp to run this "program" before connecting... Of course you may find another way to accomplish this (route command must be run as root, while Kppp can be run as normal user, unless you wish to bring kppp up as root as well, that's a bit simpler but riskier...). :)

Wouldn't it be "Powerpack" instead of "expert"? And what have you seen (error messages, etc...) that made you conclude the media is faulty? There have been several reports of error messages stating some packages could not be read or installed, during the installation process, that may lead to the assumption of a media failure, when this might not be the case (the errors may be related to changes in ide/dma behavior or to the apic management). I suggest below some tests to check for defective media, fell free to describe the one you used if I didn't include it below. Please provide more information on what kind of installation you are trying to do: 1. What happens when you boot from the installation CD (first one)? 2. Do you choose "Install" or "Upgrade"? 3. The rpmdrake errors are seen in you 9.0 system? I don't think DrakX (the graphical installer used by the installation CDs uses rpmdrake). 4. Have you tried to boot the installation cd, press F1, and enter the command: linux noapic ide=nodma 5. Can you load the CD on you 9.0 system, and browse the RPM directory? Can the cd be mounted on your 9.0 system? 6. Is the error specific to one of the 7 CD's 7. Please provide the details above, otherwise no one will be able to help you. Period. I'm sure you are interested in sorting this out, and gather further evidence that may help you advocate for a replacement media (should a faulty media be the true source of your problem).

Edit the .config file, and change the line CONFIG_APM and the lines with APM that follow it (and perhaps also the line CONFIG_ACPI): CONFIG_APM=y change to # CONFIG_APM is not set Quoted from the kernel-docs help:

Congratulations! :D

Knoppix runs on a read-only filesystem (including /usr/lib and /usr/bin), so installing new software to knoppix without first installing it to a read-write filesystem (HD install) is very difficult (although not impossible, I've compiled and installed the lucent winmodem driver, with some difficulties...) It is possible that Knoppix supports rpm, but it will only work properly in a hd install.

One of the biggest differences between Mandrake's custom Linux kernel and and other Distributions' kernel is the "supermount" feature. I don't know if they specifically turn on "experimental/dangerous" flags in the kernel config, like fixes to specific/known chipset bugs. Anyway, I think there should be a good explanation to this. Linux is open source, and it should be possible to track down which line of code is causing the freeze. My goodness, this guy has *paid for* powerpack, and this shouldn't be so complicated... I think that he could register his box in MandrakeExpert and make effective use of the support he is entitled to... Andreas

Sorry, I thought that option would turn off quiet mode. Anyway, it is possible to remove the "quiet" option from /etc/lilo.conf (one of the "append=" options) After modifying lilo.conf, lilo must be run to reinstall the bootloader (with options pointing to the correct device, lilo.conf and boot locations, although running the "reinstall bootloader" of the rescue disk will automate this). The point is that if we never know what the error is, Mandrake will not be able to fix the problem, and no bug-fixes for PowerPack will ever be developed to address the problem... I don't think lots of people who bought PowerPack would be keen to (or able to) recompile the kernel just because a possible but yet unidentified bad choice in the .config file prevents them from booting the system, while at least 2 other Linux distributions work OK in the same system... P.S.: Good luck with your exam!!! Andreas

VeeDubb: Have you tried the suggestion of booting with: linux quiet=off Even if the kernel can't write the logs, it will display error messages and other useful info to the monitor, and then we may have some idea of what can be corrected in the .config file... Andreas

Yes, with knoppix, you may open a console window, then type sudo su (just in case, you'll become root), then cd /mnt ls (locate the likely root partition at the fixed disk, whose mount point is probably already provided by knoppix , like /mnt/hda5), then mount it: mount /mnt/hda5 cd /mnt/hda5/var/log/kernel vi warnings (quit with :q) vi errors vi info Andreas

It is difficult to find an explanation to the problem without have a more verbose output of the booting kernel (before boot.log is written). It may be that the kernel can't mount the disk, and so it would hang. But Mandrake installs the kernel with the "quiet" option on, which makes us think that "Init 2.83 booting" is the last thing that the kernel did, when the error might be elsewere. So, if you can, please press ESC at the boot menu, and enter the following command to turn off the quiet (non-verbose boot) option: linux quiet=off Then look for possible error messages during the kernel boot process (before the disk is mounted or any failed mount attempts). Andreas

System had problems booting to KDE or GNOME after installation, that were related to a bad configuration of XFree86-4 (I'm not sure, but it seems it failed to create the /etc/X11/XF86Config-4 file). Others have reported that it is still possible to boot to IceWM or other lightweight WM in similar situations. In KDE it displayed a blue background, a clockwatch in the center, and a fraction of the taskbar in the botton with 2 icons in it. HD was grinding, and grinding, and never recovered from the situation. I fixed the problem by: 1. booting to runlevel 3 2. XFdrake --expert 3. Open the XFree86-4.3 tree (using the spacebar) 4. Selecting generic "sis" driver ("vesa" also WFM) 5. Plug-n-play monitor 6. Resolution (whatever supported by the card) 7. DON'T TEST (if I test, the program quits without saving the new config) 8. Quit and save 9. init 5 (or reboot)

Interesting info here: http://www.linux-usb.org/USB-guide/x332.html and here: http://www.qbik.ch/usb/devices/showdev.php?id=52 This page seems reassuring as well: http://www.busybits.com.au/item/AM5050?bus...84e022473318ab9 And finally this (german, but not hard to understand, especially the commands to run at konsole as root): http://linux.centernet.de/linuxtag2000/tal.../usb.html#ToC12 Andreas

I've just found out how to get the "command line" boot: Just press <ESC> at the initial (OS choice menu) screen! Andreas

Yes, just add an "install=/boot/boot-text.b" or "install=/boot/boot-menu.b" line just above the default=linux line, and don't forget to run /sbin/lilo as root. Lilo.conf manual page reads: install=<boot-loader> (...) (Version 22.3): The selection of the user interface with "install=" is dependent upon the presence of the substrings "text", "menu", or "bmp" in the parameter to the "install=". The boot loaders selected are no longer files, but are contained within the boot-installer binary /sbin/lilo. If "install=" is omitted, the default boot loader is "menu", unless "bitmap=" is specified (see above), in which case the boot loader "bmp" will be selected. Andreas