Re: How to mount at boot
T-SNAKE [EMAIL PROTECTED] wrote: ] How do you mount all the drives/partitions at boot time without having to ] actually type in the mount -t msdos / command? Add the partitions you want mounted to your /etc/fstab. See the fstab(5) man page for details. Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
X in hamm problem solved
When i upgraded from bo to hamm, X broke. The fix was to remove the XF86Config file from /etc/X11 and recreate it using XF86Setup. I specified the amount of video RAM on my card manually, overriding the probe. This may or may not have had anything to do with fixing the problem (SuperProbe correctly detectstthe amount of video RAM). Perhaps there were some leftovers in my old XF86Config file that confused the server? I'm using an Orchid Fahrenheit (S3 924) card. Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: cyg-win32 problems
I wrote: ] dump: cannot execute: One of the library files needed to run this application cannot be found Hamish Moffatt [EMAIL PROTECTED] wrote: ] You may need Cygnus's GNU Win32 DLL, I can't remember the exact name. ] I agree that static should remove this requirement though. ftp.cygnus.com. Yes, that was it! I copied cygwin.dll from /usr/i386-unknown-cygwin32/lib to the directory with my cross-compiled DOS program, and that made it work. Thanks a lot! Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
cyg-win32 problems
Hi, i'm playing around with the Cygnus win32 gcc cross compiler in hamm. I can compile and link programs just fine, and file identifies my binaries correctly: make dump;file dump i386-unknown-cygwin32-gcc -Wall -c dump.c -o dump.o i386-unknown-cygwin32-gcc -Wall -c q4401.c -o q4401.o i386-unknown-cygwin32-gcc -static dump.o q4401.o -o dump dump: MS-DOS executable (EXE), OS/2 or Windows However, when i transfer the program to my DOS box (and rename it to dump.exe), it wont run. It says: dump: cannot execute: One of the library files needed to run this application cannot be found I linked the DOS binary static... What's the problem? Any ideas? Is there a 'Quickstart' document for any of the win32* packages? By the way, if i use gcc instead of i386-unknown-cygwin32-gcc, then i end up with a perfectly usable Linux executable. No extra libraries are needed there. I dont do anything fancy, just basically open a file, read it, and print it to the screen. TIA Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
notes on upgrade from bo to hamm
I just finished upgrading from bo to hamm. It worked almost flawlessly. Here are my notes. After running the upgrade script, any new bashes that i start dont work with set -o vi, even if i run that command manually. The problem appears only on the linux console, if i telnet it with a VT100 terminal emulator then vi keybindings work right. X doesnt work any more... When i try to start my X server, it switches to video mode and hangs. I have to log in remotely and reboot. I had a working X installation under Debian 1.3.1. I did the upgrade to hamm, and tried to run X. It hung. I reconfigured it using XF86Setup. The VGA server ran fine, but when it tried to start the S3 serveer it freaked out as above. The partition with my /usr on it filled up while i was upgrading X. I saw no complaints from the packages being installed, fixed the problem (freed up a bunch of space by moving the kernel sources off) and continued the install. Perhaps one of the X packages does not properly deal with disk-is-full errors? I'll try uninstalling everything and installing it fresh on a roomier disk. A couple of problems with installation and configuration of specific packages: Unpacking libc6-dev (from .../devel/libc6-dev_2.0.6-2.deb) ... dpkg: error processing /usr/local/tmp/debian/dists/unstable/main/binary-i386/devel/libc6-dev_2.0.6-2.deb (--install): trying to overwrite `/usr/include/ndbm.h', which is also in package libgdbmg1-dev dpkg-deb: subprocess paste killed by signal (Broken pipe) I uninstalled libgdbmg1-dev and installed libc6-dev, then it worked. Preparing to replace gv 2.9.4-3 (using .../text/gv_3.5.8-4.deb) ... Unpacking replacement gv ... dpkg: error processing /home/kuzminsk/debian/dists/unstable/main/binary-i386/text/gv_3.5.8-4.deb (--install): trying to overwrite `/usr/X11R6/include/X11/pixmaps/mini-gv.xpm', which is also in package fvwm-common Preparing to replace wu-ftpd 2.4-27 (using .../net/wu-ftpd_2.4-27.1.deb) ... Unpacking replacement wu-ftpd ... dpkg: error processing /home/kuzminsk/debian/dists/unstable/main/binary-i386/net/wu-ftpd_2.4-27.1.deb (--install): trying to overwrite `/usr/man/man8/ftpd.8.gz', which is also in package netstd dpkg: dependency problems prevent configuration of grmonitor: grmonitor depends on glut3g (= 3.5-1); however: Package glut3g is not installed. dpkg: error processing grmonitor (--configure): dependency problems - leaving unconfigured This one's just a typo, the package it REALLY depends on is called glutg3. I dont know which should change, the package name or the dependency. All in all i think that went real smooth. Thanks a lot everybody! Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
upgrade to hamm == X problems
Yesterday i upgraded from bo to hamm. That went pretty well, but X broke. I ran out of disk space during the X upgrade, and i thought maybe that was the problem. I removed all the X packages and reinstalled on a roomier disk, hoping that would fix things. It didnt, X still doesnt work. It worked fine with bo but when i start the server under hamm, my machine switches to video mode, blanks the screen, and stops talking to the console. The server output looks normal (looks like it's not aware there's a problem). I cant switch virtual consoles, i cant toggle numlock, but i can ping the machine and i can log in remotely. I can also reboot with Ctrl-Alt-Del. I'm using an old Orchid Fahrenheit 1280 card (S3 based). I can run the VGA16 server ok, but not the accellerated S3 server. Any ideas? I guess the transition from XFree86 3.3 to 3.3.1 broke support for my card... I'll go look at the XFree web site. Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: upgrade to hamm == X problems
David Gaudine [EMAIL PROTECTED] wrote: ] Are you sure you can't switch virtual consoles? I have the same problem ] in all other respects. I have to use ctrl-alt-f1 instead of alt-f1 ] to switch virtual consoles since X is running I tried both Ctrl-Alt-F? and the regular Alt-F?. Neither works. I also tried logging in remotely and running 'sudo chvt ?', and that just sits there. I also cant kill the X server when logged in remotely. Something's definately pretty broken here... Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
notes on upgrading to hamm
I started with an up to date bo distribution, and used the 'upgrade' script (version 0.11) that was posted to the list a few days ago. It mostly worked, but there were three problems. After running the upgrade script, any new bashes that i start dont work with vi keybindings, even if i run set -o vi manually. This may become fixed after i install more packages, i dont know. I had a problem with timezones. Before, with bo, you couldnt select the Pacific timezone, tzconfig would complain that it was ambiguous with Pacific-New. So, i had my timezone set to Pacific-New, and the timezones package didnt like that. The upgrade script exited, and it wasnt clear wether that was a normal termination or not. I ran tzconfig by hand, set the timezone to Pacific, and reran the upgrade script, and everything worked well. After running the upgrade script, dselect didnt work... This was a little scary. I configured the package source (using the 'Access' menu option) to use ftp.debian.org (also tried a couple of mirrors, same problem), base directory /debian, distributions unstable, non-free, and contrib. It complained it couldnt find the Packages file for unstable. If i told it to use /debian/dists/unstable for the debian directory and get distributions main, non-free, and contrib, it could get the three Packages files, but would fail when it actually tried to download. It would cd to the debian directory i entered above, but then tack on dists/unstable again. Sounds like the Packages-files-getting logic doesnt know about the new ftp site directory layout. The .deb-files-getting logic does seem to know about the new directory layout. The workaround was to download the Packages files by hand, set the Access options as the first one tried above, and then install the Packages files by hand. I'm downloading the rest of the packages now, we'll see how that goes. All in all i think the upgrade went really smooth. Good job guys! Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
BIND question
I have a problem with BIND. We have a medium-sized network (192.168.0.0) connected to the internet via a dial-on-demand ISDN line. We pay by the minute, so we want to keep connect-time to a minimum. The ISDN router basically masquerades the internal network. I want to run a nameserver on my Linux box on the 192.168.0.0 network. I want to set it up with a couple of features: 1. It should cache 'real' addresses from outside. 2. It should forward unknown queries about 'real' addresses to a nearby nameserver on the other of the ISDN line. 3. It should serve as a primary nameserver for the 192.168.0.0 domain. 4. It should not forward any unknown queries about the 192.168.0.0 net over the ISDN line. I have got it configured to do 1, 2, and 3. However, i dont have a complete database of 192.168.0.0 names, and queries that arent satisfied by my zone database are forwarded. Is there any way around this? Any help is appreciated. Sebastian -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Installing 1.2 on a headless machine
[EMAIL PROTECTED] (Bruce Perens) wrote: ] Change /dev/tty1 on that ] filesystem to the serial device and write it back to the disk. That did the trick! Thanks much. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Installing 1.2 on a headless machine
I am trying to install Debian 1.2 on a computer without a display. This machine has successfully been running Linux for some time, using Miguel van Smoorenburg's excellent serial console patch. I compiled a 2.0.27 kernel as specified in readme.txt on the December 8, 1996 Rescue disk. I also compiled in the serial drivers and serial console driver. I added the appropriate command-line parameter to syslinux.cfg on the Rescue disk, and changed /dev/console in the root.bin ramdisk to point to /dev/ttyS0 (where i have a serial terminal attached). This new Rescue disk boots and mounts the initial ramdisk, and the /etc/rc script runs, but i get no output after that. I am assuming that something runs after /etc/rc, and doesnt use /dev/console for it's user I/O. I'm confused as to the start-up sequence of the Rescue disk... ldlinux loads root.bin and linux, then passes execution to linux, with the specified boot-time parameters. The kernel uncompresses itself and initializes the drivers etc. It then uncompresses root.bin and mounts it as root, and executes /linuxrc. This scripts execs /sbin/init. What happens after this? TIA Sebastian Kuzminsky [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Installing 1.2 on a headless machine
I wrote: ]I'm confused as to the start-up sequence of the Rescue disk... ] ldlinux loads root.bin and linux, then passes execution to linux, ] with the specified boot-time parameters. The kernel uncompresses itself ] and initializes the drivers etc. It then uncompresses root.bin and ] mounts it as root, and executes /linuxrc. This scripts execs ] /sbin/init. What happens after this? Jens B. Jorgensen [EMAIL PROTECTED] wrote: ] After this init reads /etc/inittab runs and starts processes depending ] on the run level. Normally this is the case, but the rootdisk image contained on the Rescue disk has no /etc/inittab... There's an /etc/init_tab which later gets used for the /etc/inittab on the newly-installed system, but i dont think it gets read when booting off the Rescue disk. Does anyone know how /Install/dinstall gets run? Sebastian Kuzminsky [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Serial cable pinout??
David Puryear [EMAIL PROTECTED] wrote: ] Does any one know the pinout of 9 pin serial ] cable? This specific information is available at: http://www.paranoia.com/~filipg/HTML/LINK/PORTS/F_The_Serial_Port2.html#THESERIALPORT_002 The site is run by Fil, and he also has LOTS of other good info available. His site should be required reading for anyone who wants to be a hacker. You can go to his home page at: http://www.paranoia.com/~filipg or straight to the technical info (Fil's FAQ-Link-In Corner): http://www.paranoia.com/~filipg/HTML/LINK/LINK_IN.html Sebastian Kuzminsky [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Mounting a file as a filesystem
The loop device is pretty nifty. I use it to maintain a root-disk image used on a special-purpose diskless machine. To make changes to the root disk, i mount the image, update the FS, unmount it, compress it, and copy the compressed file to a floppy. Then i can bootstrap the system from that new and improved compressed root disk. This works great, except for one annoying detail. After some amount of editing the root fs in the file, the mount count exceeds the routine fsck threshold, and the next time i commit the image to my root floppy and boot off it, the system complains when mounting the file system: maximum mount count exceeded or something like that. So my question is this: how do i fsck the filesystem in a file? TIA Sebastian Kuzminsky [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: Mounting a file as a filesystem
Sebastian Kuzminsky [EMAIL PROTECTED] writes: ]So my question is this: how do i fsck the filesystem in a file? Guy Maor [EMAIL PROTECTED] wrote: ] Use losetup(8) to associate the loop device with a file first, then ] fsck, and then mount. ] ] losetup /dev/loop0 /the/loopback/file ] fsck -t ext2 /dev/loop0 ] mount -t ext2 /dev/loop0 /mnt Oh yeah, that makes perfect sense. Thanks! -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
Re: ethernet card
Karl Ferguson wrote: ] As for PCI versions - they're worth it, they ] speed up transfers accross the network dramatically (that is, if you have ] both cards being PCI). Rick Macdonald [EMAIL PROTECTED] wrote: ] It seems to me that the ETHERNET-HOWTO goes out of it's way to prove ] that ] 16bit ISA ethernet cards can handle 10Mb ethernet just fine, and that ] PCI ] doesn't buy you anything. There are lots of reasons for choosing PCI over ISA... In my experience, the throughput of a good 16-bit ISA card is approximately equivalent to that of a PCI card. However, the CPU overhead of network transfers using a good ISA card is much higher than that of good PCI cards. Your next machine will probably have a PCI bus, but it might not have an ISA bus. The total throughput of a good ISA bus is approximately 5 MB/s, if memory serves, as compared to approximately 50 MB/s for even a mediocre PCI. If you sink 1 MB/s into network transfer (yeah yeah, it's only about 700 or 900 KB/s, but let me make a point ok?), you would consume 20% of your ISA bandwidth, and only have 4 MB/s left over for things like disk and video access. On the other hand, on a PCI bus you can sink 1 MB/s to the net and still have LOTS of room for disk, video, etc. About the only advantage of ISA over PCI that i can see is it's slightly cheaper and more old machines support it. As a side note, i have had very good luck with 3com products. I heartily recommend them to anyone shopping for network stuff. Sebastian Kuzminsky [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED]
inconsistency/confusion in aout-svgalib of Debian 1.1
I'm not sure this is the right place to send this bug report... There's an inconsistency in Debian 1.1. When the aout-svgalib package is installed, it puts the libary files in /usr/i486-linuxaout/lib, but ld.so is not configured to look there for libraries, so the aout svgalibs are unusable. There seem to be two simple ways to fix the problem: * a one-line addition to /etc/ld.so.conf * installing the libraries etc in /usr/lib/i486-linuxaout instead I'm not sure which way is The One True Way. Sebastian
Re: XFree86 3.1.2F plans?
Erik B Andersen [EMAIL PROTECTED] wrote: ] XFree86 3.1.2F is out, according to http://www.xfree86.org. I am ] wondering if a ] Debian package of this beta software will be made available. Anyone ] have any ] plans to make this available??? I would really appreciate this as well. I use a card based on the S3 765 (not the 764), and this chip breaks all the 3.1.2's prior to beta release E. Release E works well, i havent tested F. Sebastian
installation notes
Hello, i'm a new Debian convert. I have been using the ftp method of dselect to install and update Debian 1.1b on a few machines. I have a few questions and comments on the installation/maintenance procedure. In all i am very happy with the quality of the Debian distribution. The installation went very smoothly, and most of the few bugs/hickups i found have since been fixed. The Packages files not always up to date. This causes the dselect tool to think version X of a particular package is available from the ftp server, when in reality the package available is X+epsilon. This causes dselect to try to retrieve a non-existing file, and it fails. This in turn causes some dependencies to fail. For example, the Packages file in unstable/ currently thinks that the available version of xforms is 0.80j-2, but the file present in the unstable/binary/x11 directory is identified as version 0.80j-3. A couple of packages are missing some of the install scripts. For example sudo. Here is the installation error. I guess this should be directed to the maintainer of the sudo package? --- Setting up sudo ... dpkg (subprocess): unable to execute post-installation script: No such file or directory dpkg: error processing sudo (--install): subprocess post-installation script returned error exit status 2 --- The default system init scripts have two shortcomings in my mind. kerneld should be started before ANYTHING else happens, with the possible exception of a root filesystem check. The reason for this is that it may be needed to load binary-format modules, or md-personalities, etc. I had my system lock up during boot because i have swap on a raid0 md device, and i compiled raid0 as a module. Also, mdadd -ar doesnt work. The mdtools man pages are wrong. The workaround is to mdadd and mdrun each md-device individually. All in all i am very happy to have discovered the Debian distribution. Sebastian Kuzminsky [EMAIL PROTECTED]
Re: installation notes
I (Sebastian Kuzminsky) wrote: ]The Packages files not always up to date. This causes the dselect ] tool to think version X of a particular package is available from the ] ftp server, when in reality the package available is X+epsilon. This ] causes dselect to try to retrieve a non-existing file, and it fails. Guy Maor [EMAIL PROTECTED] replied: ] The scripts that update Packages run every day at about noon, PDT. I ] typically install packages in the evenings, so you're right that the ] files are not always up-to-date. ] ] I'll think about running the scripts by hand after I install. The Packages files in both the buzz/binary-i386 and the unstable/binary-i386 directories STILL think the available xforms version is 0.80j-2. In both cases the available version is 0.80j-3 Surely it has been noon in PDT since installation of the updates xforms packages. What's up with those two directories anyway? Is buzz the code name for Debian 1.1? Should i retrieve the buzz stuff rather than unstable? I wrote: ]kerneld should be started before ANYTHING else happens, with the ] possible exception of a root filesystem check. The reason for this ] is that it may be needed to load binary-format modules, or ] md-personalities, etc. I had my system lock up during boot because ] i have swap on a raid0 md device, and i compiled raid0 as a module. Guy Maor replied: ] But swap should be turned on before doing any memory-hungry things, ] like fsck for example. kerneld is already started pretty early. You ] can rename the S files in /etc/rc2.d (changing their numbers) if you ] want to change the order. ] ] My advice is not to compile raid0 as a module. It'll never be removed, ] you realize, so there's really no benefit to compiling it as a module. I administer a couple of machines, some of whom can take advantage of a raid0 driver (because they have many disks) and some who cant. I like to compile as minimal a kernel as possible, and localize the kernels for the different machines using loadable modules for everything that is common to all machine (like raid0 and serial drivers for example). I will read up on init 2.62 and change the rc file names as appropriate to fix this problem on my local network. Also, how about including the Network Administrators Guide and the System Administrators Guide in debian-doc? Sebastian Kuzminsky [EMAIL PROTECTED]
Re: ncr-810 support?
Dale Martin [EMAIL PROTECTED] wrote: ] It would appear that the boot disks for debian 1.1 beta don't have ] generic ncr-810 SCSI support - the .93 disks did. Are there ] alternative boot disks available with this support? I perused the ] installation document and saw no mention of it. I ran in to this same problem. I made a custom boot disk with the NCR driver. It's available at: ftp://ftp.cs.colorado.edu/users/kuzminsk/boot-disk.gz It's not the latest and greatest version kernel, and it doesnt mimick the Debian boot disk exactly, but it's worked well for me. Sebastian Kuzminsky [EMAIL PROTECTED]
Debian 1.1 X problems with S3
Just installed Debian 1.1beta. It was pretty smooth, except for X, and i dont think the X problems are really Debians fault, but what the heck, i'll ask for help here anyway. I have two systems... System 1: motherboard: ASUS SP3G CPU: AMD 486 dx4 100MHz Graphics card: Orchid Fahrenheit 1280 (ISA, S3 924, 1 MB) Linux: Very old, very hacked Slackware (ELF, X11R6, pretty stable) System 2: motherboard: ASUS P55TP4N CPU: Intel Pentium 166MHz Graphics card: Diamond Stealth64 VIDEO 2001 (PCI, S3 765 (Trio64V+), 2 MB) Linux: Debian 1.1beta I've had System 1 for a long time, it's been really good. I'm currently building System 2. My problem is this: When i start X on System 2, the S3 server prints the normal startup messages, switches video mode, and locks solid. The SCSI light is on, but there is no disk activity (I use the NCR 810 SCSI controller on both machines). The machine stops taking keyboard input (NumLock LED doesnt change state, cant switch VTs, etc), and stops responding to pings. When i start X on System 1, everything works fine, no problems. System 2 runs svgalib programs fine, no problems. System 1 with the video card from System 2 also works fine. System 2 with the video card from System 1 works fine too. The S3 X server from System 1, when run on System 2, hangs just like the X server on System 2. It's very confusing. Any idea what's up here? It seems like maybe the PCI bus is misconfigured, or the X server is missing some little flag or something... TIA Sebastian Kuzminsky [EMAIL PROTECTED]