RE: reportbug config
From: Hans-J. Ullrich [mailto:[EMAIL PROTECTED] can I change the editor of reportbug from joe to vi ? If yes, where ? In /etc/reportbug.conf. -- Bhaskar
RE: ATI driver slow: which way to look ?
First place to look would be /var/log/Xorg.0.log (use the latest Xorg log file after the X server is started). Perhaps it isn't able to allocate enough memory (for shared memory graphics)? -- Bhaskar S. Manda From: Hans-J. Ullrich [mailto:[EMAIL PROTECTED] All 3D-acceleration functions are accessible and working, But they work real slow. I do not think, it is the driver itself. ... How can I check the rest of the graphics-environment ? Are there any ways i.e.
Re: which kernel for dual opteron ???
On Wed, May 31, 2006 at 03:17:08PM +0100, Adam Stiles wrote: The best kernel for your system is always one you have compiled yourself. I think that's a load of crap. The majority of systems work perfectly with one of the debian provided kernels. I agree, mostly. If you use a packaged kernel, you end up with the same compiled code and drivers that you would have if you compiled it yourself. However you sometimes need drivers that aren't compiled in or modules included with the packaged ones. Ndiswrapper and mppe are two that come to mind. In this case, you have to get the patches for these, which are in dselect, and compile your own kernel. Get yourself some kernel sources from kernel.org, the .config file for There isn't a need to do this even if you want to compile your own kernel. You should really install kernel-package; this will give you the tool make-kpkg, which should be used to make kernels in Debian. It will work with source packages (linux-source-2.6.x) that are installed through dselect, and will create debs of the kernel image and headers. After all that trouble, you might find that the sources won't compile, or you might end up with no console or even no boot disk, because you didn't select the driver for your SCSI card, etc. It is best to start off with a packaged kernel until you get the hang of building your own for your machine. -- Bhaskar S. Manda
RE: I broke my X
Norval Watson wrote: /var/log/Xorg.0.log shows ... Failed to load module GLcore ... Failed to load module speedo ... Failed to load module nv ... Failed to load module mouse ... Fatal server error no screens found The errors about modules by themselves aren't fatal, except possibly for the mouse module. You can fix that by installing xserver-xorg-modules-input-mouse (or something similar, search for xserver.*mouse) in dselect). Look in /var/log/xorg.log.0 (or something like that; 'ls -lrt' in /var/log after you try to start X and it will be one log files at the end of the listing). This log file will tell you why it found no screens. One way for that to happen is when the vert and horiz sync ranges are too narrow and the resolutions too high. You can expand the ranges (the manual for the monitor will have the safe ranges, could be something like 30-65 for horiz and 60-90 for vert); or you can try a lower resolution in the screen section (for example, add 800x600 so that that is looks like Modes 800x600 1600x1200 The log file will be helpful in diagnosing this. If all else fails, you can always uninstall all xserver-* packaegs and re-install; although this may mean reinstalling Gnome packages, etc. -- Bhaskar S. Manda
RE: SATA RAID MV88SX6041 on Sarge 3.1r0a Problem
From: Mushu [mailto:[EMAIL PROTECTED] I tried to install Sarge 3.1r0a AMD64 on a Supermicro server with Marvell SATA RAID MV88SX6041 controller. Because Sarge Installer CD could not recognize the controller I used a temporary IDE disk to compile and got the driver (ftp://ftp.abit.com.tw/pub/download/drivers/linux/marvell/mvsa ta340.zip) I have a Supermicro motherboard H8DAR-T with this Marvell controller on it. It boots just fine off a SATA drive (a single drive, no array configuration). You need the stock sata_mv driver. It is already in the kernel, and the etch netinst disk. However it isn't in the initramfs, so you have to change that per my earlier posting. -- Bhaskar
RE: Dropping to a shell
-Original Message- the instructions for this please? Thanks you for your help!! You can lsmod in the shell and find out whether or not the driver for your SATA or SCSI card is loaded. If not, then you have to boot again with your netinst CD and mount the root partition. Modify the /etc/modules to include the driver, and re-make the initramfs. I'm going off some instructions that I googled, but here it is the best I can remember. I'd installed from the Etch netinst CD, but after the first reboot, it always dropped to a shell because the installed kernel didn't have the sata_mv module, that is needed for my Marvell SATA controller. I went through the netinst again, all the way upto when it asks you to reboot. At this point I switched to another console (Alt-F2 or Alt-F3 or Alt-F3 till you find one to use, because the log is on one of them). Since you are running netinst, your future root partition is mounted at /target. I edited the /target/etc/mkinitramfs/modules and added sata_mv. You can find the module you need by doing lsmod, and looking in the output for the SATA or SCSI driver for your hardware. Some instruction say at this point to mount /rpoc with `mount proc /target/proc', but I don't remember that I actually did this. Then chroot to /target with `chroot /target`. I think at this point you have to mount your future boot partition, if any, at /boot, for example, if /dev/sda2 is your boot partition, `mount /dev/sda2 /boot`. Then re-make the initramfs, with `mkinitramfs -o /boot/initrd.img-version`. You can TAB, and it'll complete the version or give you a choice. Then `umount /boot`, `exit` (to exit from the chroot), `umount /target`, and reboot.
RE: Dropping to a shell
You can lsmod in the shell and find out whether or not the driver for your SATA or SCSI card is loaded. If not, then you have to boot again with your netinst CD and mount the root partition. Modify the /etc/modules to include the driver, and re-make the initramfs. -- Bhaskar S. Manda Financial Engineer Cooperfund, Inc. 611 Enterprise Dr. Oak Brook, IL 60523-8811 (630) 573-8700 (630) 573-0652 (Fax) -Original Message- From: William Humphrey [mailto:[EMAIL PROTECTED] Sent: Thursday, March 16, 2006 1:05 PM To: Glenn English Cc: debian-x86-64@lists.debian.org Subject: RE: Dropping to a shell Thanks for the reply Glenn! I just tried your suggestion and it still does the same thing. I also forgot to mention that I am using Quad Xeon's 3.33ghz 64-bit. I did an ls- l on my partition /dev/cciss/c0d0p1 and it finds it, but ctrl-d does get me out of the shell, it just continues to say tty job control turned off. Do you maybe have any other suggestions? -Original Message- From: Glenn English [mailto:[EMAIL PROTECTED] Sent: Thursday, March 16, 2006 12:32 PM To: debian-amd64@lists.debian.org Subject: Re: Dropping to a shell On Wednesday 15 March 2006 15:15, William Humphrey wrote: I am using the latest official testing version of debian (kernel 2.6.15) on a DL580 G3 with Quad Procs. I used the netinst image. Once I installed the program, I have the problem to where is keep dropping to a shell and will not boot. It say that it cannot find my hard drive partition which is /dev/cciss/c0d01. And on top of that, it does not create an /etc/fstab for me to mount it. I have tried creating a fstab myself, but once I reboot, it goes away. Has anyone had this problem? If so please help! Waiting 20 seconds and ctl-d'ing out of the shell works here. When mine does that drop into a shell trick, I just let it sit for 10 or 15 seconds. The boot process prints on the screen that it's created the relevant drive in /dev. And a couple seconds after that, 'ls /dev/sd*' shows the partitions. And ctl-d climbs out of the shell, and the boot process finishes. What you describe is happening here on 2 machines: a dual Opteron Sun running smp etch in 32 bit mode and a single P4 homebrew running sid. Both of them have a SCSI boot drive and a SATA to store big stuff on. (The servers running sarge are fine.) The 2.6.15 kernel and/or the current udev and/or something else I don't know about are/is bent pretty badly. SATA drives are not SCSI drives. If the developers want to run them through the SCSI driver, that's fine. But they could at least call them sdA... so things wouldn't get confused. And whoever's doing that reordering should be put up against the wall. Or at least the installer should be told about the reordering algorithm. -- Glenn English [EMAIL PROTECTED] GPG ID: D0D7FF20 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
debian-testing-amd64-netinst.iso fails reboot
When an installation of debian-testing-amd64-netinst.iso (21-Feb-2006 02:59 89M) is rebooted, it fails because it cannot find the root fs. The error is device-mapper: 4.4.0-xxx initialised Unable to find volume group Debian. ALERT! /dev/mapper/Debian-root does not exist. Dropping to a shell. This is with a SATA disk. The GRUB boot line is root [hd0,5] This iso was supposed to fix harmless error messages from udev and also an incorrect GRUB root specification. -- Bhaskar S. Manda
