scoonma

Hi Chiara, which version of Mandriva are you actually using? With 2007 Spring, you could check menu entries by opening/editing them easily. Are you using GNOME or KDE?

Sorry if that was not so clear. With "system monitor" I meant a CPU/system performance tool, not a heat measuring tool. With gnome, you can put a small representation of gnome-system-monitor into the panel. I'm sure something similar does exist for KDE, too. The small app shows how much load the system is actually taking; the output is updated every second or so. When the system monitor output freezed, I knew it was time to reboot (no keyboard or mouse input possible, too). So what I did was the following: After rebooting I opened the BIOS menu and set the FSB multiplicator lower than normally used for this CPU. With a 200Mhz front side bus and a multiplicator of 11, you could tell your system to run at 200Mhz x 11 = 2200 Mhz. The BIOS on this machine allows to set the multiplicator by 0.5 steps. After lowering the factor by 1.5, I could go back to normal operation with no freezing. In the example above, you would get 9.5 x 200Mhz = 1900 Mhz. This is no optimum, but until total breakdown occurs, I can go on working. BIOS settings and options tend to vary at quite an amount between different boards/manufacturers, so you may or may not have this possibility. Older mainboards offered setting FSB multiplicators by manually setting jumpers on the board, newer ones may have this option accessible through BIOS menu. Mine is in the Miscellaneous section. If you have the same possibility to slow down your system, you can try degrading the actual value (probably default now) by minimal steps, then save settings, reboot and observe any possible improvement. But as mentioned before, this is not a cure for any possible error causing the strange behaviour, just a workaround. And it's not guaranteed to help in your situation. Before trying that, I'd make sure you have a "hard" system lockup (indicated by system monitor and no reset sequence or MagicSysRq key sequence possible). I'd try CTRL-ALT-Backspace, CTRL-ALT-Delete, ALT-PrtScreen-b, in that sequence (i.e. try the latter only if the former fails). HTH, scoonma

After copying the live cd to your system, you'd probably have to add boot menu entries manually. Go to MCC (Mandriva Control Center) / System start / Boot options (Boot loader menu) and make appropriate changes. Normally you can skip to the second page and check wether entries for Windows and Mandriva exist. They should be created automatically. After GRUB was installed to the MBR (or your first harddrive/partition), you can exit MCC, reboot and check wether it worked. Good luck!

If your sound works correctly once, you should try running alsamixer (maybe as root, too) from terminal mode after running alsaconf. From my experience, sound level settings are not stored correctly in some cases.

Actually I had the very same symptom (system freezing irregularly) on this machine. After a lot of searching I found it was CPU related. This is a dual boot system too, but I had BSODs pointing in that direction with Win$ rarely. Syslogs were normal the same way. Some hints came from attempts to boot newer kernels, with error messages (kernel panic) directing me to the CPU. A temporary workaround was to decrease CPU speed using BIOS settings (yet the processor wasn't overheated or overclocked before). I could only assume the processor was getting too old to run at the highest freqs (2.5 years)?!? I'm absolutely unsure wether your case has the same origin, but you may give it a try. I would be interested in the results. A good sign of indicating a total system freeze in X windows mode was a system monitor running in the panel. Thus I could rule out input device problems. HTH, scoonma

Does your keyboard show the same behaviour if you toggle terminal mode (Ctrl-Alt-F1 - F6)? Can you log in normally here?

As you chose to upgrade "everything" (option --auto-select) the amount of packages to update is quite normal. However, kernel packages are also included to that with 2007.1, as mentioned before. Doing major upgrades, I've had an amount of even 2GB data to install twice. So you should be fine with "Y". :-)

Yes, a couple of mirror servers (especially in Germany) did not update well in the past. But as long as you can choose repository servers freely... It was annoying, but no big deal.

When looking up a specific topic, it's often useful to google for "linux HOWTO xyz".

If you're not willing to upgrade to the newest Mandriva version (2007.1 will be out soon), you may take a look at the backports directory for your version within the software repositories. Good luck!

Hi Bovy, welcome onboard! You can easily install the package ntfs-utils-client (and dependencies) taking the following steps: 1. Make sure you have all basic mandriva software repositiories installed correctly. You may follow the steps at the top link of these pages, directing to Easy-Urpmi. (You'll need them anyway, for upgrading actual and installing other software) 2. Open a terminal window and enter "su" to become root. 3. Type "urpmi ntfs-utils-client" HTH, scoonma

This is just a rough layout, but I think you'd have to achieve at least the following: 1. Trap a key sequence similar to Ctrl-Alt-Del for system shutdown cancelling. This could be done in /etc/inittab. But you'd have to figure out how (special) keys are addressed. 2. Write a script that is executed when your own key sequence is pressed. That script should go through the list of active processes (i.e. from "ps -A x") and grep out the process number of the shutdown command. It then should terminate the shutdown process number ("kill -9"). Depending on how far the shutdown process has got, you'd also have to go back to runlevel 3 or 5 (init 3/init 5) again.

If there were no special functions of iscan you're definitely relying on, I'd try to get along with the sane package and it's frontends. Depending on your scanner model, perhaps you also need to download the corresponding firmware to be integrated into the driver.

I'd first try to select the module snd-hda-intel from mcc. You should be able to select it in the hardware section/soundcards/Start configuration tool/driver. If that does not work, you could add the following to modprobe.conf: install snd-hda-intel /sbin/modprobe Good luck!

Hi sparrish, having no other clue I'd also try the intel8x0 module with nforce chipset based mainboards. You can try inserting modules using the old way manually, i.e. by executing the modprobe command as root. Once you've found the right one, put it in /etc/modprobe.conf for being loaded at boot time. See also "man modprobe" and "man modprobe.conf". HTH, scoonma

Okay, I put this in the Cooker section, but I wonder wether it really belongs here. My AMD 2700+ / nforce2 based machine produces machine check exceptions in a very irregular way (i.e. not reproducable, can't find any pattern until now). The problem got worse within last week, actually I cannot even boot with a "normal" 2.6.17.11 series kernel. Being in X, the machine freezes left to reboot using MagicSysRq key, when in terminal mode I can see at least the error message. Sometimes I'm left to use the reset button. I've made hardware tests thoroughly (Knoppix memtest etc.), but nothing is found. The machine is dual boot, and I can run Win2000 for hours without any problems. So I doubt this is hardware related, however this is suggested in many forums for this type of error. There's no overclocking or heat problem, too: System runs at normal speed and sometimes freezes even after half a day being off. I suspect some aspect of udev or similar, but not really having a clue. The error occurs sometimes very early during boot (before root fs is being mounted), so I presume it's an aspect also going into initramdisk. I've tried several kernels, too, but they reproduce the annoying behaviour with just slightly different flavour/variation. Now I know it's cooker, being unstable as such, but I'm playing around with cooker versions for some time, and there had never been such a strange behaviour that bad. Anyone else with similar experiences, or (hopefully) hints? Regards, scoonma

This is correct. But there is no "optimal" choice between these alternatives. Both are somewhat compromise solutions. In fact, you can have 3d acceleration in the first case, however not directly through the driver, but through Xgl. AIGLX is generally the better solution (and would be fastest with a complete driver supporting all features - but this is until now the "missing 3rd option" for many ati cards).

How far can you go on with the booting process? I could disable onboard sound on this mobo in favor of using a soundcard simply by preventing the referring module being loaded (i.e. by editing /etc/modprobe.conf). Or maybe you could unload that module and load the module for your alternate card later?

Screen position also depends on your gfx board/chip and on the driver you're using. I'm correcting horizontal screen position with monitor settings (physically).

I browsed through some forums and got the same outcome (See also this one on AGP aperture size in BIOS: http://www.rage3d.com/board/showthread.php?t=33882145 - uff!). That means, if you want full 3D experience, it's better to go on with the free radeon driver - so you're able to use AIGLX, too. OTOH, from my experience the radeon driver was much slower than the proprietary one, so I once decided to switch over to nvidia. But there's constant improvement with xorg, so...

It's okay, you can go on with the stripped version. Besides, what is that ati-kernel? If unsure, you can look it up by "urpmi -vi ati-kernel-2.6.17-5mdv" (if you have full hdlist.cz installed, i.e. not the compressed synthesis hdlist).

Let's try to sort this out from the beginning: 1. Which kernel do you use? You can find out by typing "uname -a" at the command line. Do you have kernel sources installed matching exactly your kernel version? 2. Re-install the package "ati" by "urpme ati" and "urpmi ati" from the command line as root. Do you get any errors here or are the packages being installed correctly? dkms-ati relies on the kernel source tree to compile the correct driver for your kernel. If there are any errors, please post the output here. 3. Modify xorg.conf to get your fresh built driver loaded. If you still have trouble, please post your xorg.conf here, too. HTH, scoonma

These are the correct ones. With fglrxinfo, I've made a spelling mistake. Sorry! (...there's no underscore) With a proper /etc/X11/xorg.conf file you should get along. See also this thread for a good example. If you have still trouble, feel free to ask again! :-) The direct rendering module is not being loaded, which is often due to an error in /etc/X11/xorg.conf. Thus X11 is falling back to software rendering, indicated by use of the mesa drivers. For ati cards, your xorg.conf has to include the following entry: Section "DRI" Mode 0666 EndSection Please see the link above to a working xorg.conf file.

This is probably a driver issue. Where did you get your fglrx from? What does "fglrx_info" present as output, when type in a terminal window at the command line? I found it most convenient to install the ati driver from the plf sources (see the Easyurpmi-Link at the top of these pages). Drivers name is ati, you'll also have to install dkms (should go automatically). There's some settings in /etc/xorg.conf to be correct for working 3D desktop. Please take a look at this thread, so you'll get some more info. HTH, scoonma

Hi there, New (proprietary) linux drivers for nvidia gfx cards are out. This time it's an official version dump to 1.0-9755. It's also the last version before they change to a new version format. However, I've not found something like a complete changelog yet. The link is here.