sdr problems with name resolution...
I'm having dumb problems with sdr, not sure if it's me or sdr being dumb though... [EMAIL PROTECTED] ~]$ sdr SDR: getaddrinfo failed, couldn't resolve 'spoon'! [EMAIL PROTECTED] ~]$ grep spoon /etc/hosts 128.30.28.19spoon.csail.mit.edu spoon [EMAIL PROTECTED] ~]$ host spoon spoon.csail.mit.edu has address 128.30.28.19 [EMAIL PROTECTED] ~]$ nslookup spoon Server: 127.0.0.1 Address:127.0.0.1#53 Non-authoritative answer: Name: spoon.csail.mit.edu Address: 128.30.28.19 what's up with that? -Jon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: nfs-root booting works for 2.2.* client, but not 2.4.19
To point out the obvious... are you sure you flipped the kernel-autoconfig bit in networking and the allow nfsroot bit (somewhere else I forget, probably network file systems) have the right NIC driver builtin to the boot kernel I presume you're getting a can't find init message or a panic about the root device. One issue I had with 2.4 kernels is that I sometimes made them too big. I'm not sure too big for what as the one that works is bigger than 640k and the ones that failed fit on the floppy with syslinux and config stuff... -Jon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Fwd: Silicon Graphics Indy workstations?
http://www.debian.org/ports/mips/ yes you can run Debian on MIPS based Indys
Re: problem installing debian
On Thu, Jan 10, 2002 at 11:31:41PM -, David S wrote: :Hi, I recently got a copy of debian 2.2r4 through one of the official vendors on the debian website and i'm having trouble accessing the x window system : :this is the error message i get when initx or startx command is used: : :_exec of usr/bin/x11/xf86_NONE failed :_x11transSocketUNIXConnect can't connect: errno= (111) :giving up :xinit:connection refused (errno 111) unable to connect to xserver :no such process (errno 3) server error Hmmm, your *first* problem is that XFree86 isn't installed :I've heard about problems with the geforce2 mx400/linux and since using this card have had trouble using graphics with other distros as well (mandrake, suse) Your *second* problem is that XFree86 version 3.3.x doesn't support your video card and stable doesn't have version 4.x yet. Fear not, you can add the following lines to /etc/apt/sources.list : # XFree86 v4.x support # deb http://people.debian.org/~cpbotha/ xf410_potato/i386/ deb http://people.debian.org/~cpbotha/ xf410_potato/all/ then as root: apt-get update apt-get install xserver-xfree86 xbase-clients xutils xterm xprt I'm not sure you need to specify all of those, some may be pulled in as dependencies but being explicit can't hurt. your *third* problem will be getting the /etc/XF86Config-4 file right, if the tools don't do the magic for you, email me directly and I'll send a base config that should work ok (though probably not to the full capacity of your monitor). I support alot of systems with nVidia cards so I have a generic config lying around your *fourth* problem will be getting hardware 3D acceleration working. In the testing branch there's some wrapper .debs for pulling in the non-redistributable nVidia kernel module and OpenGL libraries, but I have a howto (not the best, but a start) at: http://www.ai.mit.edu/lab/sysadmin/debian/nvidia.html Sweet system, though perhaps a little bleeding edge for Debian stable near the end (I hope :) of a release cycle. HTH, -Jon
Re: scp question (scp and exit shell)
On Thu, Jan 10, 2002 at 04:34:34PM -0800, Mike Egglestone wrote: :Hi all, : :I would like to copy a huge file from one server to another. :but because it takes a long time, I would like to exit :my remote shell that I used to login and run the scp command. : :I have tried with thesymbol after the command but no go. :ex. :scp /home/user/file.txt bigserver:/home/user2/ A theory... start the scp: scp /home/user/file.txt bigserver:/home/user2/ then after authenticating and starting the file transfer CTRL-Z to stop the scp then: [EMAIL PROTECTED] bg this will restart the last job in the background. -Jon
Re: Is anyone using woody in a production environment?
On Fri, Jan 11, 2002 at 12:42:20PM +1000, john wrote: :2) We upgrade to testing. : :Is it safe? image of Marathon Man. Who is running production servers :on testing? what if any issues have arisen? Well, not production servers. I do have a mix of workstations some testing, some stable, and a few I've pulled tricks on. We don't support testing so the fact that I haven't heard complaints doesn't indicate a lack of problems. But in a few cases where someone had a demonstrated need for newer tools than stable provides (wxpython, XFree4) I've switched to testing installed the required packages (which ofcourse pulls the new glibc) and then switched sources.list back to stable. This has worked extremely well. As I said these are user workstations not mission critical. If you can install a test system and then stress test it, then you'll see if this can work for you.
Re: Strange boot problems.
On Sun, Dec 30, 2001 at 06:43:25PM +, Jason Wood wrote: :All I need to do now is to figure out how to mount the root partition as read :only, so that I can run fsck on it manually, back to google I go... mount -o remount,ro / after fsck either reboot or: mount -o remount,rw / HTH, -Jon
Re: Strange boot problems.
On Sun, Dec 30, 2001 at 04:03:00PM +, Jason Wood wrote: :# ln /boot/vmlinuz-2.2.19pre17 vmlinuz.old :ln: creating hard link `vmlinuz.old' to `/boot/vmlinuz-2.2.19pre17': Invalid :cross-device link Not clear on your larger problems, but this bit I can. You're trying to make a hard link which doesn't work across different devices. Use a soft link: ln -s /boot/vmlinuz-2.2.19pre17 vmlinuz.old ^^ man ln for details -Jon
Re: My nVidia folly :( FIXED ?
Seems fixed, of course I did a few things not sure which was the magic bullet of if all were required. * Upgrated the nVidia stuff to 1.0-2313 (from 1.0-1541) * removed 'Load dri' and 'Load GLcore' from XF86Config-4 much happier now, though not as stable as the nv driver. -Jon
Re: My nVidia folly :( FIXED ?
On Mon, Dec 17, 2001 at 07:26:48AM -0600, Lance Simmons wrote: :If you're removing dri and GLcore from XF86Config-4, is there still any :hardware acceleration? yes, the glx extention as provided by nVidia does this. There's a FAQ on their site that doesn't seem to come in the tar ball (atleast the older one) which says to load glx but not dri or GLcore. glxinfo shows direct rendering enabled and glxgears runs at 340 or so fps (dual ppro 200MHz riva tnt card) -Jon
My nVidia folly :(
Hi, I admit it, lured by the promise of a faster screensaver I installed the closed source nVidia GL libraries and modules, and now I'm suffering for it. My GL screen hacks are running 4x faster, but I can't switch virtual terminals and am limited to a single running Xsession (I usually have three or four open for other member of my household). If I CTRL-ALT-Fn all I get is noise, *sometimes* I can get back to X sometimes I can't. If I actually close X my console is locked, so I'm forced to run a display manager to respawn X. I sould probably just reinstal the packages nVidia has stomped on and repent of my proprietary ways, but has anyone else experienced this and know a different fix? I'm running sid with: 2.4.14-686-smp (debian official) xserver-xfree86 4.1.0-9 nVidia Corporation Riva TnT [NV04] (rev 04) (PCI) Dual PPro 200Mhz 256M and the evil: NVIDIA_GLX-1.0-1541 NVIDIA_kernel-1.0-1541 Interesting dmesg stuff: Warning: Remapping obsolete /dev/fb* minor 32 to 1 Warning: Remapping obsolete /dev/fb* minor 64 to 2 Warning: Remapping obsolete /dev/fb* minor 96 to 3 Shamefully yours, -Jon
Re: My nVidia folly :(
On Sat, Dec 15, 2001 at 03:55:57PM +0100, Mario Vukelic wrote: :I have a TNT also and no probs (single CPU though). I use the debs for :both glx and kernel. There's .deb's! now I really feel like a fool... -Jon
Re: My nVidia folly :(
On Sun, Dec 16, 2001 at 07:07:19AM +1000, Peter Good wrote: :Using framebuffer (nVidia Riva support) in your kernel? That's what gave me :all my probs with console lockups. Soon as I took it out of the kernel, :everything worked great. I did, I took it out (prior to posting), and it solved the random hard lockups, I'm guessing it's an SMP issue. I've set this up on many newer machines for people who realy use the GL stuff (their medical imaging -vs- my screensavers), but I don't think any were SMP machines. I reinstalled xserver-xfree86 and the mesa libs (don't know if I needed to) so now I'm working and mourning the loss of 28fps screensavers but not all that much... -Jon
Re: emu10k1 troubles...
To a previous question about /dev, I don't usually use devfs on this machine but to findout what the kernel was seeing: [EMAIL PROTECTED] jon]$ ls /mnt/sound/ dsp dsp1 midi mixer Now to see what of this is really in /dev: [EMAIL PROTECTED] jon]$ ls -l /dev/{dsp,dsp1,midi,mixer} ls: /dev/midi: No such file or directory crw-rw1 root user 14, 3 Nov 30 2000 /dev/dsp crw-rw1 root user 14, 19 Nov 30 2000 /dev/dsp1 crw-rw-rw-1 root user 14, 0 Nov 30 2000 /dev/mixer [EMAIL PROTECTED] jon]$ groups user rt systems aries sysadmin uucp Note root has the same trouble On Wed, Dec 12, 2001 at 04:40:33PM -0800, Brian Nelson wrote: :The emu10k1 driver in 2.4.6 is pretty stale; it was updated somewhere :around 2.4.9 to come somewhat up to date with the creative sources. I :believe there were updated because the older driver didn't work for :Linus. I got the latest from creative and built against the running kernel, buth the stock 2.4.6 and the creative modules exhibit the same problem. : emu10k156272 0 : ac97_codec 9264 0 [emu10k1] : :I assume the sound module is loaded as well, right? Oh, yes cut too much :Does the driver load properly in dmesg? Yup: Creative EMU10K1 PCI Audio Driver, version 0.17, 11:36:20 Dec 11 2001 emu10k1: EMU10K1 rev 7 model 0x8022 found, IO at 0xdce0-0xdcff, IRQ 16 ac97_codec: AC97 Audio codec, id: 0x8384:0x7609 (SigmaTel STAC9721/23) :Does something like: :# cat some_sound.wav /dev/dsp Nope, no complaints and no sound. I just discovered /dev/dsp1 in responding to these messages, wouldn't it be clever if that's all my trouble is. Unfortunately I won't be able to test this untill tomorrow. -Jon
emu10k1 troubles...
Hi, I'm having weird problems with sound using the emu10k1 module. The system bell is coming out through the speakers, but other sound (xmms, gcd, etc..) seems to play (ie no complaints about devices, visualization shows levels) but no sound comes out. Adjusting mixer levels affects the system bell, but not the others. I'm running 2.4.6, tried the modules that built with it and the latest from creative. Both have the same behaviour. Get the same as root and two different regular users. emu10k156272 0 ac97_codec 9264 0 [emu10k1] I'm puzzled. -Jon
MATLAB testing?
Hi, Some of my testing machines won't run MATLAB6 unless java is disabled. I've made sure the broken machines are running the same versions of X, gcc, and libc as the ones that work, but no change... Now I'm not fool enough to officially support the testing branch, but does anyone know why this is and if I can fix it? TIA, -Jon
Re: server on a 486/100 with 408mb drive
On Sat, Sep 01, 2001 at 04:50:55PM +1000, Sam Varghese wrote: :On Fri, Aug 31, 2001 at 11:24:30PM -0700, Osamu Aoki wrote: :i've already got a server running potato with 2.2.17. I figured :woody was stable enough to be used on a server by now. You can still get bit hard by a stray typo, but I've been running woody/sid on my home server for some time, so shouldn't talk :) :just a question of been there, done that, with potato. thought :i could get a shot at configuring iptables, that's why i :thought of 2.4. you can get kernel 2.4 support in potato (I use this on production workstations), but adding this to /etc/apt/sources.list: #kernel 2.4 stuff deb http://people.debian.org/~bunk/debian potato main deb-src http://people.debian.org/~bunk/debian potato main -Jon
Re: Looking for a wireless ethernet solution...
On Thu, Aug 30, 2001 at 02:20:56PM +1000, Andrew Pollock wrote: :Can you mix and match 802.11(b) gear and expect it to cooperate? Yes, that's the point. Expecting and it actually happenning I've seen a mix of cards talk to Lucent/Oronoco (sp) card, so I'd be optimistic about that. There are (or were) media converters to go from a wired NIC to 802.11 (bring your own card though), I got my hands on one and it didn't quite do what I wanted. My hope was that I could plug this frob into a port on my hub and it would pass through a signal from my laptop. No dice, apparently there's a difference in the standard between an access point and a client node. After much mucking about I did get it attached to my Linux router (parameters need to be passed at module load to set Ad Hoc mode or something), It did finally work, but I had to return the hardware. Not sure where to hunt one of these down or how much they cost, the one I had was from Lucent. a SOHO class wireless hub would be cleaner, and they usually (always) have RJ45 jacks. -Jon
Re: HELP with NFS lossage...
On Wed, Aug 29, 2001 at 11:33:44PM -0500, Michael Heldebrant wrote: :Quick question: : :Did you update the /etc/exports file to allow the hosts to mount and :then run exportfs? yup, other machines mount it fine. The IP block it's exported to covers my whole /16 , I've mounted from machines in the same and in different /24's The problem client can mount NFS volumes from other machines (NetApps and Solaris servers) -Jon
Re: RedHat vs Debian?
Hi, I wish I had numbers for you, but I don't. I do remember we had to build our own patched ssh binaries for the last voulnerability 'cus RH didn't have 'em for a day or two. Debian had packages out withing hours of the patch being posted. Never been able to upgrade a RH box on my network (too many NFS dependencies or something), Debian has handled this. shameless type=self-aggrandizement The MIT Artificial Intelligence Lab has stopped supporting RH and moved to Debian. I managed this by setting up my workstation with Debian and showing howmuch more quickly security patches are out and how much easier it is to admin (OK and writing my own custom installer so noone had to see how tedeous it can be to install with all the config questions that are going to be the same on every system we install). /shameless The joys of apt are reason enough! Of course if Management is as technically inept as the term makes them sound, you may be SOL. I haven't seen any statistics like what you'd need (if you find them send us a link). On the other hand if they're that brain dead how would they know your new webserver was running Debian? I wouldn't recommend replacing the OS on a production server, unless you're building a replacement anyway though. -Jon
HELP with NFS lossage...
Hi, I'm baffled. I have and NFS server (ishmael) which serves /fs/deb as a local apt-repository amoung other things. It also serves /fs/deb/nfsroot as a seperate export used as part of my local autoinstaller. After making some changes to the installer (don't think I touched anything related, same kernel new XFree86 and a few config file changes), I can still install a system using the /fs/deb/nfsroot, but after the install I can't mount /fs/deb the error on the server is: Aug 29 12:20:36 sol4 rpc.mountd: authenticated mount request from 128.52.39.139:801 for /fs/deb (/fs/deb) Aug 29 12:20:36 sol4 rpc.mountd: getfh failed: Operation not permitted The server is a Sparcstation20 running Debian 2.2, the installer installs Debian 2.2 with a hacked apt/sources.list to get kernel2.4 and XFree86-4 support (tthe XFree support was changed in my latest hacking the kernel was not) and is for x86 boxen. Quick help would be GREATLY appriciated as I have 15 machines to install tomorrow am and they need the XFree-4 support. BTW, this will mark the start of official Debian GNU/Linux support at the MIT AI Lab (and the end of support for new RH boxen) -Jon
Re: Can MSN work on Debian?
:I don't see any reason why he can't. MSN is just another ISP like :Earthlink or Mindspring. Exim will get your email from mail.msn.com and :Konqueror accesses the web just fine. All you need is the MSN DNS IPs :during your setup of your Debian network. Because it's M$ and they break things on purpose (though stupidity comes into play at times too), see the previous response... :Heck, if Debian works with AOL it can work with anything :) Does it now??? Is that a good thing??? well I suppose for kids dual booting their parents computer it's a $DIETY send.
Re: Running remote X apps
On Fri, Aug 24, 2001 at 07:27:33PM -0700, Duncan Watson wrote: :X is not listening. So even though I changed /etc/X11/xinit/xserverrc and :removed the -nolisten tcp stuff X is still not open a TCP port to listen :on. : :I have installed xfree86-common 4.0.2-1, could it be that it is a compile :time option? Not compile time as the only available switch is '-nolisten proto' there is no equivelent -listen switch. If you're using a graphical login (xdm, wdm, gdm, ...), the config for the Display manager is used not /etc/X11/xinit/xserverrc, which is used when you startx from the commandline. Using wdm the gonfig you want is in /etc/X11/wdm/Xservers. Since all of these display managers are essentially hacked versions of xdm, I suspect the file conventions are the same. -Jon
Stupid apt tricks?
Hi, I want to have apt-get to fetch from a single repository for a scripted update. in pseudo commands: apt-get -o=Apt:repository:file:///com/debian/dists/stable update apt-get -o=Apt:repository:file:///com/debian/dists/stable upgrade apt-get update # to reset things to a normal state I though I had this working at one point, but can't remember how. TIA, -Jon
Re: adminning mult boxen
On Mon, Aug 27, 2001 at 03:33:02PM -0700, der.hans wrote: : Once you've figured out what exactly you want to do, you can try to : find something in the usual places (like sourceforge), or write a bunch : of cron + ssh + perl + CVS scripts to do what you want. : :I'm gonna as on the SAGE mailing list, then start that way if they don't :have any suggestions. I find it hard to believe cfengine is the only package :that does this type of thing. That is what FAI is using, so maybe I should :give in and use cfengine. There's a package (possibly only in unstable) called replicator, that's basicly a perl wrapper around rsync for installing and syncronizing systems. Haven't used it yet, and probably won't completly replace cfengine, but it sounds interesting and could probably be trivially back ported to stable (since it seems to be all or mostly perl) : Having an all-debian network helps, at least you can use apt's get- and : set-selections to keep packages in sync. SMILEY type=envious/ If you're going to be keeping these as absolutely identical installs, you could cronify a set-selections from a canonical source and dselect-update, but upgrading an installing deb's is a fairly interactive task, stutting it up completely isn't so easily done. I have high hopes for replicator as it seems to be the right thing. see http://www.infrastructures.org for a more generic view of what that right thing might be -Jon
Re: Stupid apt tricks?
On Mon, Aug 27, 2001 at 03:36:02PM -0700, der.hans wrote: :Look at the setup I used to use: : :http://home.pages.de/~lufthans/unix/ : :See the link about security, testing and unstable updater for debian. :They're just calling apt-get with options and alternative dirs in /var. Mmmm, that hits the spot, I was using too few : and too many \ Thanks, -Jon
Re: Stupid apt tricks?
On Mon, Aug 27, 2001 at 04:19:58PM -0700, Ron Steinke wrote: :When I look at the link, I just get binary garbage. Is this supposed :to be an executable, It's a .tgz file, you'd need to download and unpack it. It's an axample, not much in the way of documentation, though if you were me this is all you'd be looking for :) :and if so where is the documentation on how to modify :sources.list? man sources.list other interesting man pages: apt apt.conf apt-get and /usr/doc/apt -Jon
Stable and XFree86 4.x
Hi, I know this has come up before, but my searching skills aren't up to the task of finding it in the archives apparently... I have about 20 workstations on the way for incoming students and They all have GForce2 cards (AFAIK this requires XFree86 4.x), what's the best way tho keep these machines running stable but also grabbing XFree86 4.x? I seem to recall someone building stable .deb's for this a long time ago, but can't seem to find where they are. TIA, -Jon
Re: Stable and XFree86 4.x
On Fri, Aug 24, 2001 at 10:16:33AM -0400, David Z Maze wrote: :When I did this, I went to ftp.xfree86.org and found a tarball :containing just the server binary and things you absolutely needed to :run the X server, and installed this under /usr/local. Worked like a :charm. (Though at this point, the machine is sitting headless as a :server box. *nostalgic sigh* Five years ago, 100 MHz Pentia were My fall back plan is to get the deb source files and roll my own .deb, as this is going to be a more and more frequent requirement round here, and I do have a locally aptable NFS repository. Good to know the sources build OK. -Jon
Re: Stable and XFree86 4.x
On Fri, Aug 24, 2001 at 02:27:58PM +0200, Erdmut Pfeifer wrote: :see here: : :http://people.debian.org/~cpbotha/ : :or, if you're feeling adventurous: : :http://cpbotha.net/building_xfree86_4.1.0_debs_on_potato.HOWTO Excellent, Thanks. -Jon
Re: Running remote X apps
Hi, Sounds like you've checked everything I would, any reason not to use ssh which does automatic forwarding of X (well if it's config'ed to else ssh -X host). Given the private numbers I'm assuming you trust the local net and wan't to avoid the encryption overhead, but seems worth a shot. -Jon
Re: Choosing a Debian Variant
My $0.02 on make-kpkg, for one off kernels on my home machine I prefer the old fasioned/standard way of make bzImage. I have a lot more control over the build process that way. for kernels that will be distributed to many machines make-kpkg r0ckz. -jon
Re: Running remote X apps
On Fri, Aug 24, 2001 at 02:16:42PM -0700, Duncan Watson wrote: :ssh is not an option as the hosts I actually need to use don't have ssh and :I am not an admin on all of them. That sucks :( You can run sshd on a high port as joe_user, if you configure the client to accept that weirdness. But clearly that's the wrong thing for you (or atleast not the solution to the problem at hand) :I need to get tcp X connections working. Hell I even delved down into :xauth man pages. I was really hoping that would help. Scarry stuff, that man page. well, you've done the xhost +host, set the DISPLAY variable on the remote session, and checked that the -nolisten tcp flag is *not* being sent at X startup, hummm. Have you verified that X is really listening? netstat -tl will show all listening tcp ports you shouls see port 6000 open if display :0 is listening (6001 for :1 etc.) If it really is listening, I'd start wondering about packet filtering rules, have you messed around with ipchains or iptables (I'm assuming you've run remote clients from these machines else where and no filtering is going on there)? -Jon
dumb wav-mp3 question
Hi, what's the command to convert .wav files to mp3 format? this is a dumb question 'cus I was doing it a week ago and after 3h with apropos,dpkg -l,apt-cache search,and ls /usr/bin |less, I still can't find it :( TIA, -Jon
Re: dumb wav-mp3 question
On Sun, Jul 15, 2001 at 09:42:31PM +0200, Joost Kooij wrote: :On Sun, Jul 15, 2001 at 03:30:18PM -0400, Jonathan D. Proulx wrote: : what's the command to convert .wav files to mp3 format? : :You want lame for that. Perhaps that's what I want, but I'm 99% certain I apt-got something that does this, and lame isn't apt-able (well not from normal sources) :Better is to use the ogg format instead of mp3. Ogg is free. :It is also said to be technically better. Install vorbis-tools. I did run across that, where technically better == lossless For my current streaming purposes I converted the files to realaudio (even worse than mp3 from a freesoftware perspective), I'll need to review ogg support in main stream clients before comitting... And it hurts my brain that I did this less than two weeks ago and can't find the tool I used. -Jon
exim and unqualified recipients...
Hi, exim doesn't accept unqualified sender|recipients despite the following lines in /etc/exim/exim.conf: qualify_domain = ai.mit.edu qualify_recipient = ai.mit.edu I'm still getting: mail from: jon 501 jon: sender address must contain a domain mail from: [EMAIL PROTECTED] 250 [EMAIL PROTECTED] is syntactically correct rcpt to: jon 501 jon: recipient address must contain a domain What am I missing? -Jon
Re: mount (2.4.4)
On Mon, Jun 25, 2001 at 07:43:20PM +, Robin Gerard wrote: :hello, :I have succeded compile kernel-2.4.4, but :Y have yet a problem : :I want read my boot diskette, I do : :mount /floppy :and I get the msg : :fatfs : bogus cluster size :VFS:can't find a valide MSDOS filesystem on dev 02:00 :( I am rather happy of this ) :mount:you must specify the filesystem type :Can someone gives me advices or tells me wehere is the :doc of this mater. You probably have a raw kernel image on the floppy (no filesystem at all). Try rdev /dev/fd0 you should get back something like Root device /dev/hda1 if you have a kernel image there. You're probably better off using lilo or another boot loader to boot of the HD, boot ing from floppy is Soo tedious :) -Jon
Re: An *idea* that *might* put Debian on top (?)
On Mon, Jun 25, 2001 at 07:04:19PM -0400, Daniel Barclay wrote: : From: Jonathan D. Proulx [EMAIL PROTECTED] : ... : Except for maintainence, you're reading the best tech support I've : ever gotten, bar none. Most stuff gets atleast some response within : an hour, 27x7. I pay through the nose to get 4h support tht's usually : not as good :) : : I really can't say enough about the quality of this mailing list! : :You obviously haven't suffered the complete lack of response, or responses :that completely misunderstood the question, that I've experienced. Guess not. There's been some misses in the past 3 years, but many more direct hits and alot of insight I wouldn't have even thought to ask for. Hope you're luck turns around, -Jon
Re: [users] Re: An *idea* that *might* put Debian on top (?)
On Mon, Jun 25, 2001 at 08:01:07PM -0700, Nick Jennings wrote: : You are missing the point, installing software and it's dependencies is : one hurdle debian has cleared. But there is much more than that. There : is getting your video card working, your monitor sync set right, your : sound and peripherals working correctly. No argument here, most of this pain is install time rather than maintenance related. If the install time detection were better, simpler, whatever that would also cover maintenence (like a new video card). : Things that I know can be very difficult to an average to new user : of linux. I have installed debian : countless times, as well as other distro's and I know for sure that : debian has close to the worst hardware detection around, this does not : phase me because I can manage to get things working and once they are, : debian is a breeze. That's my point, once it's working you don't have to worry about where to get the right libraries, or bit rot in some mammoth machine managed registry file. : From your post you seem to be under the idea that there is nothing : to impove with debian because of apt-get. I must have misrepresented myself, apt-get rocks for software management, but there are other things and there's always room for improvement. -Jon
Re: An *idea* that *might* put Debian on top (?)
Hi, My initial reaction was a shudder, basicly 'cus I teak everything to within an inch of it's life and hate the limitations of GUI interface for system configuration (you can only stick in so many options). But, upon reflection something like this could have a place, but it's a huge undertaking. I'm just finishing up a one floppy installer and even with the assumptions I can make (fast ethernet LAN, 4G HD, and I know the right network configs), it's been a piece of work to make something that will fit from a PPro 200Mhz with 64M of RAM and 4G HD to Dual PIII Xeon 1.7GHz with 2G of RAM and 80G HD, then all the NIC's, and I've basicly left sound to the user. It's a pretty dirty hack, but even that took some doing. To device something that will do this from any OS, on any machine from a 386 to P4 or Itanium (not to mention all the other architectures Debian supports), with a variety of network connections (LAN, dialup. ADSL, Cable), well I hope patience is one of your virtues :) -Jon
Re: [users] Re: An *idea* that *might* put Debian on top (?)
On Sun, Jun 24, 2001 at 06:42:39PM -0700, Jack Pryne wrote: :machine take care of itself, without ever having to worry about missing :DLL's, or corrupt registry files. That's the whole idea behind this proposed :system. Though I take your point, I encourage you to get Debian up and running so you can see for yourself what types of configuration issues there are. There is nothing even analogous to registry corruption in an Un*x like system. And through Debian's wonderful package/dependency management system you're never missing shared libraries (the analog to a DLL). Actually once you have a basic system installed (which can be tricky), managing it can be as braindead as you want. run gnome-apt two clicks to update evrything you have installed to the latest revision, run a search on available packages pick the new shiny whizbang app you want, it'll grab whatever else you need to make it work. -Jon
Re: An *idea* that *might* put Debian on top (?)
On Sun, Jun 24, 2001 at 06:48:49PM -0700, Jack Pryne wrote: :Sounds like you have some valuable insight into the task at hand! But :consider the fact that this system I propose would not be trying to cram :infinite permutations on a disk. Instead, it would create a communal :reference library, holding information on the configurations of many systems :all over the world. Well I didn't cram it all onto a disk, the disk boots up and the real working pieces are mounted over NFS... I say your proposal is much more difficult. I'm working in a comparitively well defined environemnt and most of the important (to me that is) stuff doesn't change. How do you determine who's system is a good model and who's is just a bad idea (after you've matched hardware etc..)? :What would be distributed is analagous to a free :membership card which entitles the bearer to free advice, maintainence, :and/or tech support. Except for maintainence, you're reading the best tech support I've ever gotten, bar none. Most stuff gets atleast some response within an hour, 27x7. I pay through the nose to get 4h support tht's usually not as good :) I really can't say enough about the quality of this mailing list! :After all, if you were a confused newbie, wouldn't you :want the shoemaker elves to come and fix your computer while you slept? ;) I'm with the goblin guy :) aside from my do it yourself ways, the security implications of this are scary. Not to say absolutly insolvable, but far from trivial. -Jon
Re: iptables help
On Mon, Jun 25, 2001 at 01:04:30PM +1000, Tom Tsaknakis wrote: :i will give anyone my first born if you can help me with converting this :\'/sbin/ipchains -A input -s 10.96.8.1 -p IGMP -j ACCEPT\' I have all the kids I need :) But I have this working: iptables -A INPUT --proto icmp -s 10.9.1.1/32 -i ppp+ -j ACCEPT this is my ADSL router and the only 10.0.0.0/8 I want to see on the other side of a ppp link. Can't see why --proto IGMP couldn't be substituted... HTH, -Jon
Re: Careers in Linux
On Wed, Jun 20, 2001 at 01:52:02AM +0100, mark wrote: :Hi, :Iam finding it very hard to even get looked at by :empolyeers.Basically i have made a career change 1 1/2 years ago :from being a factory worker to supporting pc's (sadly win9x/win2k :for a uk comapany,name withheld to protect my job :-), at least it :got my foot in the door). Luck did have a big part in getting my current position! I was the guy who knows computers in my last job (Public Works department of a local Municipality), but hat little official responsibility for them. Try to leverage that foot in the door, can you work in a cheap linux box to do firewalling, mail, fileserving, intranet webserver, VPN ? Be creative and see if there's anything you can do with old machines going out of service or a really low end PC, so it won't cost your employer much if anything, and then you'll have official linux admin on your CV :Also I would be inclined as Iam to start playing with something like :Solaris as its more Unix based than say debian/rh/mandrake (if you :know what I mean) All my important servers are Sun machines running Solaris, so yes it's important to know. But I wouldn't call it more Unix based exept that Unix is a trademark Linux doesn't wear. From an admin perspective some of the commands behave a little differently (though I use the GNU tools mostly so they are the same), and there's a few minor differences that can be trouble till you get used to them. From a user view there's even less difference. I'd go so far as to say the difference between RH and Debian is greater than the difference between Debian and Solaris. That said, they're running Solaris because of the Sun hardware, and Sparc Linux isn't quite where we'd need it to be. -Jon
Re: Problem with woody statd? (was Re: RPC failures)
Hi, I've been having NFS locking trouble on my sid system for a while, but never got around to really looking at it. Based no your mail I did a little debuging and discovered statd was dying here too with: [EMAIL PROTECTED] jon]$ sudo /sbin/rpc.statd -F -d 06/20/2001 12:19:25 rpc.statd[6177]: Version 0.3.2 Starting 06/20/2001 12:19:25 rpc.statd[6177]: Flags: No-Daemon Log-STDERR /sbin/rpc.statd: error while loading shared libraries: /sbin/rpc.statd: symbol __rpc_thread_variables, version GLIBC_2.2.3 not defined in file libc.so.6 with link time reference nfs-common 0.3.2-1 libc6 2.2.3-6 upgrading my nfs-common to nfs-common (0.3.2-2) has fixed this. -Jon
Re: Careers in Linux
Hi, I'm a sysadmin at a computer science research lab. I'd say C is probably the least important part of my job, noone in the group is a real C or C++ hacker. You should be able to build programs from packaged source and be able to tweak a Makefile and library path, but for actually building administrative programs, interpreted language like shell and perl are very important. Especially since you'll endup inheriting your predecessor's tangled ball or scripts. So I would say that understanding programing genericly is a big plus, C isn't crucial. My $0.02 -Jon
Re: iPAQ ?
On Mon, Jun 18, 2001 at 12:19:32PM -0400, Noah Meyerhans wrote: :Which model are you getting? I've got a 3650, and am running Familiar :on it (http://familiar.handhelds.org). Feel free to come downstairs and :talk to me about it (NE43-234). There are a whole mess of people in LCS :who have iPaqs running Linux. I actually know surprisingly little about what I'm getting, whatever version Compaq is handing out to the Oxygen project that they're calling Handy 21's, pretty sure it's a 3650. Thanks for the URL. :The Familiar distribution is based on Debian, although they never seem :to come right out and say that. That's good to know. :You really don't want to try installing a full Debian system on an iPaq. :There is nowhere near enough space on the thing for it. Well not full certainly, but according to http://www.debian.org/ports/arm/ you can do Debian, though there's no clear install procedure. :If you haven't already, you'll want to get a jacket that adds a PCMCIA :slot for it. It works very well on the wireless network, and can :actually be useful if networked. Oh yeh, definately getting that. I tried for the integrated video cam, but apparently there's only six demo models in the world just yet, and I don't rank one of those :) The whole point in me getting it is to make it a useful network device. The idea is to hackup some PAM modules (or something) so that anyone from anywhere can pick this up and access they're networked resources with this as a portal type device. So you'd be able to come up, grab this thing and login with: ipaq login: ssh://[EMAIL PROTECTED] -or- vnc://[EMAIL PROTECTED]:1 and without any direct knowlege of you or your account info the ipaq could get you connected. Well anyway I've only been thinking about this for about 72h, and the next step is to see if anyone with a larger clue is already working on this. -Jon
Re: Configuring X Server during Install
On Sun, Jun 17, 2001 at 04:29:32PM +1000, [EMAIL PROTECTED] wrote: :I have been trying to install Debian on my Toshiba 2180CDT laptop. :But I just can't get X server to work. I'm not familiar with your hardware but you might look up info on your laptop at http://www.linux.org/hardware/laptop.html :Also how do I shut down Debian from the Command Line? [EMAIL PROTECTED] halt to reboot use reboot instead of halt -Jon
Re: Trying to Install Debian
On Sat, Jun 16, 2001 at 06:21:34PM +1000, [EMAIL PROTECTED] wrote: :I am trying to install Debian on a Toshiba 2180CDT laptop. Laptops are notably strange, have you tried making the boot floppies? A quick look on the web shows that this Toshiba model doesn't seem to boot from CD, see: http://www.linux-laptop.net/toshiba.html -Jon
Re: [SLOVED, sort of] kernel 2.4 and cs4232 [semi-urgent]
Hi, Machine one: I forgot to copy over the isapnp.conf from the machine I did get working, oops. Perviously I didn't need the isapnp stuff. Machine two: not a cs4232, but an ac97. Still not working, but I haven't RTFM'd yet. -Jon
Re: getting mad!
Hi, I haven't been following this closely, so my appologies if this isn't relevent. I remeber when setting up tftp for our routers, we had a problem because tftp won't create files so we had to touch /tftp/foo before we could download foo. This caused much frustration untill we stumbled across a clue :) -Jon
kernel 2.4 and cs4232 [semi-urgent]
Hi, Sound, my personal blind spot... When I upgraded to kernel 2.4.x some time ago I lost sound, and haven't been able to get it back on two out of three machines. We have a bunch of demos monday and it would be nice to put something flashy on the kiosk system (which is one of the sound less ones). This used to work, with 2.2.18 and still does on one 2.4.2 system: modprobe cs4232 io=0x534 irq=5 dma=1 dma2=0 now I get init_module: No such device, all systems running unstable. sndconfig doesn't seem to deal with 2.4 kernels (complains about sound_core missing, I wonder if I can alias that away...) TIA, -Jon
Re: kernel 2.4 and cs4232 [semi-urgent]
On Sat, Jun 16, 2001 at 12:50:41AM +1000, Steve Kowalik wrote: :Did you compile in Sound Support as a module? Yup, AFAIK sound_core.o is no longer used in 2.4. On the working machine I don't see it. re: where I compiled, one of the soundless systems was where I did the build using make-kpkg. Thanks, -Jon
iPAQ ?
Hi, This is definately premature, but... I'm getting an iPAQ early next week, and was wondering if anyone out there's wedged debian onto one yet? It's meant to come with Linux, which I can only assume means Dead-Rat, or perhaps a Compaq sepecial mini-distro... -Jon
Re: Latex version 2.09 from 7 Dec 89 ??
On Wed, Jun 13, 2001 at 10:24:27AM -0400, David Z Maze wrote: :Right Answer: LaTeX2e is intended to be a mostly-compatible successor :to LaTeX 2.09. Thanks, I'll try that first, this is all complecated by the fact that I know *nothing* about TeX of any flavor... :(I know that LaTeX2e isn't 100% compatible with LaTeX 2.09, but I :don't entirely know what the differences are. LaTeX2e added some :commands not present in 2.09, so there's a possibility of a namespace :conflict. I think the other issues can only come up if the author of :the document was a 4th-level TeXpert or higher.) Hmmm. We'll see :0 I setup the laptop in question and rather blindly through all the TeX stuff I could find at it, don't know what that works out to though... :MIT Answer: Up until a couple of years ago, Athena's default latex :installation was based on LaTeX 2.09. So if you can find a Linux box :around with the SIPB installation of Red Hat 4.2/Athena 8.0 on it, it :should have a normal LaTeX 2.09. (Then upgrade it to something more :modern!) The AI Lab is it's own creature. We don't use Athena and aren't on 18.0.0.0/8, but thanks for the pointer, if I can't get him going with LaTeX2e I'll check with the SIPB folks. -jon
Re: transfer between win2k and debian
On Wed, Jun 13, 2001 at 04:14:56PM -0400, Hall Stevenson wrote: : On Tue, Jun 12, 2001 at 09:09:51PM -, Em Huynh : wrote: : Hi all, : : I got a little network (100baseT w/switch) going with win2k : and debian. When I transfer from my debian box (p3 566 : w/ultra 66 hd) it seem to peak at about 3000k-4000k. I'm assuming those are kilo-bytes ont kilo-bits... I don't think that's that surprising. I'm seeing 6500kB-8500kB between two PII 400Mhz SCSI systems on 100baseT full duples through a Bay Networks switch. (on an 8.4MB gzipped file) If there's any other traffic to or from either machine, this will drop. In more real world situation, I'm connected to a local Debian mirror over 100Mb with 1000Mb backbone and usually see 1200kB or so transfers :Someone will surely correct me if I'm wrong, but I think full :duplex basically means the network card will receive and :transmit at the same time. Half-duplex means it receives :data, then sends back confirmation, receives some more, sends :back, etc, etc. This could apply to either 10 base-t or 100 :base-t speeds. You *may* be running at 10Mb/s full duplex. :Those speeds suggest the same... More or less right, for unidirectional traffic full-duplex is only marginally better than half-duplex. -Jon
Local auto-update script...
Hi, I'm trying to sort out an automated apt-get update/upgrade script. My plan thus far is running an apt-get upate and -qq -y upgrade against an NFS repository from cron so that I can be sure what is being installed. I'm currently looking for a way to add a bit more security to the mix. Does/can apt check pgp|gpg signatures on packages? any thoughts on this appreciated. Thanks, -Jon
Latex version 2.09 from 7 Dec 89 ??
Hi, Don't ask me why 'cus I don't know, but my fearless leader has decided he needs an ancient version of LaTex on his spiffy new Debian laptop. Since this is from Dec 1989, I'm going to go out on a limb and guess I can't get a .deb for it :) Anyone have a clue where I could find source, and what issues I might run into. I'm just hoping I don't need to pick through backup tapes from 10yrs ago to find this :) TIA, -Jon
Re: star office debian-correct installation
On Sun, Jun 10, 2001 at 10:25:23PM -0700, Mark Wagnon wrote: :I've never used sudo. Whenever I need to do something as root, I use :su. What's the difference? Is one better/more secure than the other? I find that if I use su for an X application I need to meddle with my display security (xhost +localhost or somesuch), where as sudo does the right thing. sudo is more secure in a multi-administrator setting. It logs usage so you can see who issued what command as whom. You can tune permissions so that certain users or groups are allow or dissallowed commands. And it's revocable, you never have to give out the reall root password (users authenticate with their regular password), so if you need to revoke privilege you don't need to change the root password, also people cant leave the root password taped to the monitor, because they don't know it. -Jon
Re: star office debian-correct installation
On Mon, Jun 11, 2001 at 09:10:40AM -0500, Dave Sherohman wrote: :That is a topic of much debate. In general, I fall on the sudo is evil :side of the fence, but the basic arguments are: snip :anti-sudo: It allows you to give limited root access to certain users :without requiring that they know the root password. This allows an :attacker to obtain elevated privileges on the machine by discovering :only a user password instead of requiring that they find both a user :password and the root password. obviously I'm on the other side of this most religious debate :) First, I've seen alot of interesting (and just plain dumb too) ways of breaking into Un*x boxen, but never this one. More importantly, if someone gets a local user there's a very high likelyhood they can force root easily from there. My security policy is based on the presumption that any local account equates to root. If I was cracking a box and had a choice beween a local root exploit and using sudo, I'd take the sploit as sudo does logging which I'd then need to go erase. Like wise, if you don't trust a user with full root access, for the love of $DIETY don't give them sudo. I'm sure even if you restricted the commands to /bin/true someone could find a way to root. In practice the people who have sudo also have root and we use sudo mostly to leave an audit trail. -jon
Re: debian on a 486
Hi, I installed 2.2 from floppy onto a pair of ibm thinkpad 750's (486sx 16M RAM), so the 486 chip shouldn't be an issue. If you're having trouble with a lomem install, I still have some 2.1 CD's I've been dying to donate to a good cause (I belive I wedged debian onto a 386 4M system with the floppy images on these), if yo want 'em email me off the list and we can make arrangements. -Jon
Re: X problem
On Sun, Jun 10, 2001 at 10:34:56PM -0300, Eduardo Gargiulo wrote: :Hi all. : :I have a problem with my X configuration. :When I try to start the server from command line with :startx, the X server don't work; when I try to start it :with xdm, the server starts, I fill the username/password :fields but the system don't say nothing and the login script :starts again. :Both cases, the log says something like Failed SetVCLK. Hmmm. if the xdm screen displays, then your XF86Config file is probably in order. Try starting X as root, there may be some weird permissions issue going on. (verbose log messages, as suggested, would be helpful.) -jon
Re: boot floppy
On Thu, Jun 07, 2001 at 12:22:48AM -0300, xgnu wrote: :Hi all. :2: The kernel version is 2.2.17. I want to compile 2.4.5, but :I have a tar.gz file, not the deb pkg and I had not installed :development tools. Which packages I need to compile the new :kernel? the package kernel-package contains the debian tools for building a kernel, though alot of times I just build by hand, installing this package should pull in the correct dependencies for the build environment. One think I know it usually misses (probably because it's not required only suggested) is libncurses5-dev which you need for make menuconfig, you can still use make config or make xconfig to configure the kernel without this though. -Jon
Re: star office debian-correct installation
On Sun, Jun 10, 2001 at 01:50:16PM -0700, Mark Wagnon wrote: :This started the installation program I put it in :/usr/local/bin/soffice52. After the installation finished, I then :logged in as an unpriviledged user, and ran: : :$ /usr/local/bin/soffice52/program/setup AFAIR, if you run soffice as a regular user it will so the setup thing (and then exit so you must run soffice again) :Then to run StarOffice, you would then run ~/office52/soffice. You :might want to add ~/office52 to your path. On my network install I symlinked /usr/local/Office5.2/program/soffice to /usr/local/bin/soffice, this works fine and is much simple. In fact I didn't realize I had a ~/office52/soffice, untill you pointed that out. :You may need to log in as root, startx, then run the main SO :installation before running the user installation. You don't need to be root, using sudo is fine. If you don't know what sudo is, install it and read the man page then ask here, it's *very* useful. :Hope this helps. If you have any questions, I hope I can answer them. Wow, you really went the extra mile on this one, much respect. -Jon
Re: installing unstable version
On Wed, Jun 06, 2001 at 11:13:27AM +0200, Tobias S. Hofer wrote: :hello people : :i'm interrested in installing an unstable version of debian. :what's the best way to do this. i prefere to download the :packages from ftp servers. :it seems to me that an unstable package tree does not contain :all the required packages. I've installed a number of unstable machines. My prefered method is to use a stable CD or floppy install and when prompted for install method, pick edit sources by hand replacing stable with unstable or testing *except* for the security.debian.org listing. one note the editor you will be in is ae to save use C-x C-w RET, this is hell on us emacs guys, close but not quite... -Jon
Re: Is Woody installable at this time????
On Mon, Jun 04, 2001 at 10:24:53PM -0400, Chris Hoover wrote: :Is woody/unstable currently in a state to do a apt-get dist-upgrade to? If not, is there a eta on that time frame? I've been running it for months and that's how I got there :) But I don't think I'm releasing it on my users yet, what do mean by in a state to do a apt-get dist-upgrade? -Jon
Re: small school: replacements for MS Word and Excel
On Tue, Jun 05, 2001 at 12:29:51PM +1000, Ian Perry wrote: :I have been running Linux 2.0.36 for the past year on 3 sites and have had :ext2 go down on each of them after a power failure. I'd guess the nonbootability was the result of a head crash, and had nothing to do with the filesystem software. Why are you running 2.0.36? I hope you're behind a firewall. We've had two power failures recently, one building wide, one a single floor. AFAIK we had no losses except the drive on an elderly LispMachine (hardware damage to the drive) in the first, and 2 windows machines (filesystem corruption) in the second. This is out of about 400 total machines, of which I'd guess 75 are windows and about 250 are Intel linux. (don't hold me to the numbers, but it's the rough proportions atleast :) Not really frightening, even the M$ machines had less than 2% casualties and one was pretty near collapse before the power failure so say 1% of Windows workstations 0% of everything else (including the one lonely and ignored PowerEdge server running NT) -Jon
Re: small school: replacements for MS Word and Excel
On Tue, Jun 05, 2001 at 02:03:02AM +0100, Colin Watson wrote: :Sean Morgan [EMAIL PROTECTED] wrote: :OK, the two messages previous posts kind of play off eachother so I'm :going to reply to them in one go. First off ext2, it has a really bad :habit of losing files in hard crashes and power outages, this isn't a :problem for someone like you or I as we know how to recover them, for a :student with no root and no knowledge of how to do this, it's called a :couple of hours work down the tubes. : :I have to say, I've never lost a file to an ext2 disk crash, nor even :had to go any further than the odd prompted run fsck manually to :recover it, On extremely flaky hardware, such as Colin goes on to describe, I have lost file on ext2, but only in extreme cases. And I put lost in quotes because I know that they're in the lost-found directory, just with a weird name and some careful use of `file` and `grep` has gotten them back. Though honestly even with crashes under load this doesn't often happen, and even in the rare cases when it does 99% of those are just netscape cache files I don't care about. NTFS and FAT on the other hand I've lost system files that render the machine unbootable (happened on 3 systems in the past 2 months), I haven't lost a file on ext2 ever and the last lost file was about 6 months ago. Linux beats out M$ in my lab atleast 2 to 1. The only Un*x system I've really had a lot of trouble with was a Solaris UFS file system on a very old drive that just went insane while mounted and under heavy NFS use (got that back with fsck too, eventually). So in my experience ext2 is much more robust than M$ offerings. Reiserfs is *probably* a bit better hearsaythough I hear when journaled filesystems do manage to get wedged it's all over /hearsay. RAID mirroring is also an option for the paranoid (not to suggest paranoia is always unwarrented), but this is probably a bit more technicaly complex than you want to get. -Jon
Re: How do I pass init= to the kernel?
On Mon, Jun 04, 2001 at 10:24:56PM -0400, Margarete Hans wrote: :If you wonder why I don't use a rescue disk to boot: :I get a message that the disk is not bootable. :I tried redownloading the rescue disk files, tried more than 5 :different floppies, same message. :When I tried the compact version, linux booted fin, but stopped at the :kernel panic: VFS: unable to mount root fs on 03:02 :Seems somewhat weird - I don't know what to do anymore. Can't help with loadlin (sorry), but I notice the kernel error you got from the compact floppy shows it's trying to mount /dev/hda2 as the root file system (major number 3, minor number 2). This may be what you eventually want, but at install time (the floppy disk way), you read in the root file system from a floppy disk because there may not even be a partition for root yet, much less all the files that are needed. I missed the model of this laptop, some have secrets and you should look at http://www.linux.org/hardware/laptop.html for a link to your specific model. For example old IBM thinkpads with a 2.88M floppy require you to pass floppy=thinkpad to the kernel, there's also some similar magic to use the sony pcmcia cdrom to install on one of the spiffy new vaios that don't have any removable media on board... -Jon
Re: Help Please
On Tue, Jun 05, 2001 at 11:20:00PM +0100, R. Jorge wrote: :I have install debian GNU/Linux. : :I type login - root then I type the password and after that the next line comes like that: : :debian:~# : :I don't no what I have to do can you please help me? That's the command prompt, you don't *have* to anything, but you can do what ever you want :) The suggested reading is a good start. Also make a regular user account so you can poke agound without breaking anyhting. To do this type adduser you'll be prompted for the necessary info, just use all lowercase letters and no spaces for the user name. Then logout (with logout), and log in with the new user name. Just for fun after that type startx at the command prompt (probably now debian:~$). This is how you start the graphical system with menues and pretty colors. In all likelyhood either nothing will happen or a whole bunch of text garbage will dump to the screen and nothing will happen :) This means XFree86 either isn't installed or (in the later case) is unconfigured. If you get graphic your in business, if not I baet that's what you do next. -Jon
Re: quick question
On Wed, Jun 06, 2001 at 02:04:11PM +1000, Renai LeMay wrote: :just a quick question for those adminning large networks out there - : :what software package do you use for notifications etc? I am a junior admin :in a large network of linux/freebsd machines, and looking at implementing a :system based upon netsaint, which I consider will do all I need (when :combined with mysql), however my senior admin seems to think NOCOL :is a better system. We switched from NOCOL to Netsaint about a year ago (both using qpage to sent text pages in emergencies). I don't much remember why we switched, but Netsaint seem less prone to false positives, able to monitor more details, better able to keep and display history, and the pretty web interface is nice too. Some of this may be better configuration rather than inherent superiority, but Netsaint does the job for me. services status (SMTP, NFS, HTTP etc.) load (number of users, process load, file system capacity) unusual latency machine room temp even printer status and so much more... -Jon
Re: Moving to reiserfs with kernel 2.4.3
On Wed, Jun 06, 2001 at 12:03:40PM +0200, Alex Suzuki wrote: :What are the next steps to move to reiserfs as root filesystem? Is :there a good howto around? : :Can I just cp -a the whole root to another spare filesystem, :reformat / to reiserfs and then copy it back? Hmmm... I haven't done this, but what I'd do is copy the files in / some where (very carefully to preserve permissions and preferably time stamps, I like to use cpio but you can also do it with tar, cp -a might be good enough). Make sure this some where is the top of another partition so you can twiddle you lilo.conf, run lilo and use this new place as root temporarily. Reboot. mkreiserfs on the old root, copy the root fs back, reset lilo, reboot. or it may be simpler to tar up you root, boot from a rescue disk (with reiserfs support), mkreiserfs on the root, untar your magic tarball, and reboot. Though in all likely hood you'll need to rerun lilo in there since the physical location of your kernel image will change. Not sure about that so much. Just some thought, take with generous portions of salt. -Jon
Re: small school: replacements for MS Word and Excel
On Mon, Jun 04, 2001 at 11:42:18PM -0400, Sean Morgan wrote: :The situation I'm reffering to here is that of someone who might see a temporary :interface slowdown or crash(happen quite often in office suites of any kind), :and having no knowledge of how linux works, just hits the reset button(this :could easily get repeated a whole bunch of times throughout the day in a school :environment, and would do all kinds of nasty things to the fs). True, I hadn't considered that. CTRL-ALT-DEL the three fingered salute, does a clean shutdown. I'd stress that and put it in big letters on the machine. Then disconect the front reset and power buttons. This won't keep people from pulling the plug, but it makes it easier to do the right thing than it is to do the wrong thing. :you seem to be coming from is some kind of server farm or similar setup where :the users are all at least competent enough not to do lots of hard resets. You'd be surprised what research scientists in Artificial Inteligence can do to thier machines then shrug and look all innocent :) :Generally speaking, Linux with ext2 is probably the worst OS I've ever worked :with for recovering from mundane stuff like that, I have two preschoolers that love the GIMP and linux, bad things have happened to my machine, but worse things have happened to the Loose95 machine they have for games. That's just my experience. :I don't see how RAID mirroring is an option though, as it would only :serve to guard against mechanical failure, which :is sort of beyond the ability of a filesystem to affect one way or the other. If the power goes during heavy disk usage you risk a head crash, which is physical damage. Though journaled fs (reiser or ext3 and what ever happened to xfs?) would get more bang for your buck. I can't speak much to the current state of these in linux. I've had reiserfs on my home machine for about 6mo with no issues, though I haven't tried to see howmuch abuse it will take either. -Jon
Re: Re. Total Confusion
On Sun, Jun 03, 2001 at 04:36:58AM -0700, Sidney Brooks wrote: :Thanks to Steve Kowalik, who wrote the following, the problem if not the :solution is becoming clear. :I have Windows on partition hda1, Redhat on hda6, and Debian on hda8. :Because of what Steve wrote, I did uname -a for Redhat and got 2.2.14-5.0. :It seems that when I install Debian, it grabs the Redhat kernel instead of :installing its own. The only link is that I am using the same swap partiton :for both. : :Now what? Sharing swap should be fine, I bet your lilo installs are conflicting. I've never run two Linux OS's on the same machine, but that's my guess. What is the output of ls -l /vmlinuz on your Debian root partition? -Jon
Re: differences Debian and Corel?
On Sun, Jun 03, 2001 at 09:03:30AM -0400, Jonathan Freiermuth wrote: :Short answer - yes. :Real answer - yes, if you are willing to learn. : :I've never personally used Corel, but I'm gonna go out on a limb and guess :that Corel did its best to hide the complexities of Linux from the user. : :Debian is a powerful system, and as such requires more knowledge :about how and why things work. In other words, you can probably install :debian off of an official CD, no problem, but you may have to recompile :your kernel to enable reading the NTFS volume. I've never used Corel either so caveat lector, but if it's Debian derived, can't he just edit /etc/apt/sources and do a dist-upgrade? This would leave the kernel in tact and let him keep all his data and whatever frobs he likes from Corel. -Jon
Re: How to use junkbuster in lynx?
Hi, I haven't used junkbuster, but assuming it's a normal proxy server (atleast it's interface the browser deals with) set the environment variable http_proxy=http://proxy.host.name:port; This works for my squid proxy, apt-get will also respect this variable as will wget and probably other command line things that use http. -Jon
Optimizing libc6 et al...
Hi, after reading some recent posts about apt-get -b source and pentium-builder, I'm toying with the idea of optimizing the hell out of my ppro system. Brief testing of a rebuild of the sysutils package got me a 6% decrease in the time to run memtest (I picked this because it takes a non-trivial amount of time and reports this time, also if it broke I'd barely notice). I wouldn't stand too firmly behine the 6% figure as it only compares one run of the stock binary with one run of the newly built sources... Anyway, my question is has anyone replaced system libraries using apt-get -b source? Seems like this would be the most bang for the buck as it were, but also the greatest risk. TIA, Jon
Re: Getting Started!
On Sun, Jun 03, 2001 at 10:22:39PM +0200, HawkY wrote: :Hi! : :I'm new to Debian and to Linux too. (I've just (tried to) installed Debian :Potato next to a Win2k.) And I have questions: : :How can I get a list of my installed packages. ( dpkg ???) This is rarely necissary, but if you want to use dpkg --get-selections | more ( the | more send the output of the firts command to the more program that shows it one screen at a time) If you have Xwindows working (BIG if :), gnome-apt is a nice GUI for package management if you're new to it. There's also console-apt, which I've never used. You'd probably need to install these. dselect is the old way, it's console based and to my mind tortureous. :Does this command list the packages I've installed with mc? How do you install packages with mc? :How can I install a bunch of packages with one single command? (I want to :install gnome, but I don't really know the package dependencies.) the basic format is: [EMAIL PROTECTED] apt-get install package-name This will check dependencies and install them too, after giving you a chance to decline (for exmple 300M of dependencies over 56k dial up) :Is there a book I can get from the net that describes Debian? (If it is :free... :o) ) There's lots of good online info, unfortunately I don't have URL's handy. Hopefully someone else does, if not email me off the list and I'll dig around for 'ya -Jon
Re: Re. Total Confusion
On Fri, Jun 01, 2001 at 06:15:02PM -0700, Sidney Brooks wrote: :I know that I am connected because the log says so. Please post the log out put as others have suggested. At one point it will say connected when the modems start talking to each other, later it will give the local and remote IP addresses. If you get so far as having an IP assigned: First ping the local IP that your machine is assigned. Second ping the remote IP If you can ping the remote IP things are good, try route -n the remote IP should be listed as your default gateway, something like this: Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 10.9.1.10.0.0.0 255.255.255.255 UH0 0 0 ppp0 0.0.0.0 10.9.1.10.0.0.0 UG0 0 0 ppp0 I suspect your problem comes before the routing, but it's unclear exactly where it's failing. The dial up ppp sequence of testing goes something like this: * Are the modems connecting (yes) * Is authentication successful (???) * Do you get an IP assigned (???) * Can you ping it (by IP)(???) * Is the default route being set properly (???) * Can you ping it (by IP) * Are the DNS servers IP's being set (no) * Can you ping atleast one (by IP) * Can you ping a remote host like www.debian.org by IP (198.186.203.20) * Can you ping it by name * If you can get this far pretty much everything should work That's my general method (I support about 50 active dialup accounts, though this isn't a big part of my job), comments on the method are welcome. -Jon
Re: X windows dual monitors
On Sat, Jun 02, 2001 at 02:19:58PM -0400, Yea Right wrote: : :How do I setup a dual head Matrox G400 :with 2 17 monitors. You'll need XFree4.0 (but you know this). Then man 5x XF86Config Essentialy you make multiple Device sections for each SVGA port (one card or many, doesn't matter), you get the BusID from scanpci If the monitors are the same, a single Monitor section is ok, or one for each *different type* of monitor. These are just like regular Screen sections, just make sure use use the right Device for each one. Then make multiple screen sections (atleast one per head). These screen sections are assembled in the ServerLayout section. The relevent segnent of my Quad head setup is below: ## Appian JeronimoPro 4port Card ## Section Device Identifier Appian0 # Screen 0 Driver glint BusID PCI:2:1:0 EndSection Section Device Identifier Appian1 # Screen 1 Driver glint BusID PCI:2:5:0 EndSection Section Device Identifier Appian2 # Screen 2 Driver glint BusID PCI:2:9:0 EndSection Section Device Identifier Appian3 # Screen 3 Driver glint BusID PCI:2:13:0 EndSection ## Monitor (works for all 4 OK) ## Section Monitor Identifier Hitachi202 HorizSync 30-100 VertRefresh 50-160 Option DPMS EndSection Screen Sections Section Screen Identifier Screen0 Device Appian0 Monitor Hitachi202 DefaultDepth24 SubSection Display Depth 8 Modes 1600x1200 EndSubSection SubSection Display Depth 16 Modes 1600x1200 EndSubSection SubSection Display Depth 24 Modes 1600x1200 EndSubSection EndSection Section Screen Identifier Screen1 Device Appian1 Monitor Hitachi202 DefaultDepth24 SubSection Display Depth 8 Modes 1600x1200 EndSubSection SubSection Display Depth 16 Modes 1600x1200 EndSubSection SubSection Display Depth 24 Modes 1600x1200 EndSubSection EndSection Section Screen Identifier Screen2 Device Appian2 Monitor Hitachi202 DefaultDepth24 SubSection Display Depth 8 Modes 1600x1200 EndSubSection SubSection Display Depth 16 Modes 1600x1200 EndSubSection SubSection Display Depth 24 Modes 1600x1200 EndSubSection EndSection Section Screen Identifier Screen3 Device Appian3 Monitor Hitachi202 DefaultDepth24 SubSection Display Depth 8 Modes 1600x1200 EndSubSection SubSection Display Depth 16 Modes 1600x1200 EndSubSection SubSection Display Depth 24 Modes 1600x1200 EndSubSection EndSection ServerLayout Section ServerLayout Identifier Default Layout Screen Screen0 Screen Screen1 RightOf Screen0 Screen Screen2 RightOf Screen1 Screen Screen3 RightOf Screen2 InputDevice Generic Keyboard InputDevice Configured Mouse InputDevice Generic Mouse EndSection -Jon
ANSWER: how to add extra chat options in pppconfig
Hi, I accidently deleted the post this answers (oops). Someone needed to add some more chattiness after connectig to their ISP to select PPP rather than SLIP or some other things... After going through the regular set up, it shows what you've selected for connection settings and allows you to change them if you mistyped. It also has an Advanced Options option. Select this Then Post-Login from the menu it presents. It explains itself pretty well. You can add as many extra expect-send pairs as you need here. Hope the right person reads this :) -Jon
Re: Wireless Conection
On Sat, Jun 02, 2001 at 06:55:05PM -0300, Sergio E. Schvezov wrote: :Well so far it fairly works for almost everything except for http :( I had a similar problem when I put in my DSL. Not the same, it only affected boxes behind my Debian box that was runing NAT to share the connection, the machine that was directly attached to the DSL router/gateway worked fine. Any way the fix for me was lowering the MTU on the machines HTTP wasn't passing through for (to 1464 looks like). Not quite the same problem, but it's a place to start looking. HTH, -Jon
Re: xscreensaver
On Sat, Jun 02, 2001 at 06:28:58PM -0500, ktb wrote: :Try - :exec xscreensaver :kent NO DONT DO IT! This will stop processing the file and make the screensaver the last thing that executes (ie no window manager) and if you close the screensaver you Xwindows session will exit. Only exec a windowmanager, unless you really want to do some thing odd. Sorry to shout, but I saw someone exec xsetroot which ofcourse promptly closed the Xsession after changing the background color, and they had a hell of a time figuring out why X was crashing :On Sun, Jun 03, 2001 at 01:27:08AM +0200, Jeroen Valcke wrote: : xscreensaver : However it doesn't work. What's wrong. Adding to the .xinitrc file is : correct isn't? For some reason I've been unable to fathom if you start X from the console with startx it looks at one config file if you're using xdm (or gdm, wdm, what ever) to get a graphic login it looks at a different one. These two files are .xinitrc and .xsession, I've symlinked mine (ln -s .xinitrc .xsession) and forgotten which method uses which file. -Jon
Re: Packet Traffic
On Sat, Jun 02, 2001 at 04:06:55PM -0700, Alvin Oga wrote: :think you'd end up with snmp and/or perf tools : :- a nice link i found ... nothing that helps you ?? I liked the link :) I agree snmp to get the info. There's plenty of tools command line and otherwise to view it. We use snmp with cricket (which takes the snmp data and creates near real time graphs via a CGI interface) to monitor our network equipment. We don't look at the servers directly, just the switch ports they're on, but you can go host by host and get more detailed info if you want. -Jon
Re: xscreensaver
On Sat, Jun 02, 2001 at 07:21:30PM -0500, ktb wrote: :Every .xinitrc I have put together has exec lines for each line in the :file. If I understand it correctly all exec does is execute the :program. The tricky part is when back-grounding programs with there :has to be one, usually the window manager, which *isn't* back-grounded. I :have one .xinitrc for a kiosk which has everything back-grounded except :Netscape. Hmmm... exec executes something in place of the current process. If you are in a bash shell and type exec tcsh when you exit tcsh you will exit you login session (or xterm or what ever). If you had simply typed tcsh when you exit you'll be back in bash. Obviously, this is working for you though. I suspect it's because you're backgrounding the jobs. The example I was refering to wasn't backgrounded. Alot of people don't background setting the root window (not required because it exits quickly and goes on to the next thing, though consistency would be good too). I didn't consider that backgrounding the job would make the exec work. -Jon
Re: I need a windows-linux solution
On Sat, Jun 02, 2001 at 10:06:17PM -0400, Brian Schramm wrote: :I have tried samba but it is aparently blocked at the :cable co. Apparently because it doesn't work or because the cable co. said they're blocking it? You can use nmap to determine if bits on ports 137-9 are filtered between the hosts (or call the cable co and ask, but the support folks probably won't understand...) Just on the winblows side of things: you are maping the drive (or trying to) by hand \\cable.host.ip.num\file-system not relying on network neighborhood which is well um, S*(T. You have NetBIOS over TCP enabled on the Window$ box. The share name in samba.conf must be under some # of characters to work (12 or 14), it took me three weeks to findout why I could use 3 out of 4 exports. That's all my thoughts for now. I feel your pain, -Jon
Video Capture (meteor pro ati all in wonder)
Hi, I'm looking to play around with video conferencing and I have a Matrox Meteor Pro (SAA7116 chip acording to scanpci) and an ATI all-in-wonder 128 to play around with. I'm running a 2.4.4 kernel and trying to figure out what module to load (and with the ATI if it's supported at all). The Meteor was in use in another box running Linux (2.0.36 with patches I think), so I could probably boot that and findout what modules it was using... As an aside anyone have experience with IEEE1394 video cams? support seems sketchy to non existant, but people have got to be working on it. Thanks, -Jon
Re: 12 box network in small school
Hi, Sorry if I'm coming into this late, I've been away from the list for a while and blew away the 3k messages I hadn't read so only caught the last post of this thread I've been working on an NFS root installer, which starts off essentially diskless so it can repartition the harddrive(s). One gotcha I found is that strange things can happen if /dev isn't writable by root (by default nfs exports are root squashed so local root=nobody on the NFS volume). I worked around this by using a 2.4.x kernel with devfs, so that each client has it's own writable devices without any possibility of conflict with each other. Traditionaly I think a compressed ramdisk is loaded off the boot floppy and mounted on /dev (but I've not done this). The advantage of the devfs system is that you can just dump raw kernels to floppy without mucking with lilo or syslinux, down side is that running 2.4.x is slightly more adventuresome than stock stable (though see http://www.debian.org/News/2001/20010415 for how to run 2.4.x with an otherwise stable system). I'd be happy to go into more detail, but I really missed the question :) HTH, -Jon
Re: Two keyboards in X
Hi, know is too stong a word but, you'll need XFree4.0 then RTFM for XF86Config. I was poking through for my multi head setup and there seems to be away to add multiple input devices which I *think* can be bound to a specific display. Both displays will have the same login (UID). Multi head started fine for me but exititing results in a system lockup (net down and everything), requiring a hard reboot. This may be due to my weird video card (4 port Appian Graphics Jeronimo pro GLINT chip sets), but I advise caution :) Good luck, -Jon On Thu, May 31, 2001 at 08:40:09PM -0300, Christoph Simon wrote: : :Somebody knows how to get two Xservers running, each using its own :physical keyboard (PS/2 + USB or 2 x USB)? : :Thanks,
Re: SSHD-config
On Fri, Jun 01, 2001 at 02:09:40AM +0200, Frans Schreuder wrote: :Hi, : :I'm having trouble setting up a SSH-connection to my Debian box. 1) Are you logging in as root? this is disabled by default in Debian systems for security reasons (look in /etc/ssh/sshd_config) 2) Check /var/log/auth.log to see what the Debian box thinks is going on. 3) Is sshd running on the Debian box? check the output of netstat -tl (t = tcp; l = listening), it should show in part: [EMAIL PROTECTED] jon]$ netstat -tl Active Internet connections (only servers) Proto Recv-Q Send-Q Local Address Foreign AddressState tcp0 0 *:ssh *:* LISTEN -Jon
Re: Two keyboards in X
On Thu, May 31, 2001 at 10:05:20PM -0300, Christoph Simon wrote: : :On Thu, 31 May 2001 20:49:29 -0400 :Jonathan D. Proulx [EMAIL PROTECTED] wrote: : : : Hi, : : know is too stong a word but, you'll need XFree4.0 :Thanks for the reply, but multihead is actually more complicated than :I need. I plan to run two servers, and just have to find out how to :tell X which keyboard to use. I did RTFM (XF86Config(5x), wrong one?), :but couldn't find anything, while expecting some option to specify :something like the mouse device. Are you using XFree 4.0.x? this is required, 3.3.x doesn't have this: ---from XF86Config(5x) for Xree86 4.0.3--- INPUTDEVICE SECTION The config file may have multiple InputDevice sections. There will normally be at least two: one for the core (primary) keyboard, and one of the core pointer. InputDevice sections have the following format: Section InputDevice Identifier name Driver inputdriver options ... EndSection The Identifier entry specifies the unique name for this input device. The Driver entry specifies the name of the able server, the input driver module inputdriver will be vice section is considered active if it is referenced by an active ServerLayout section, or if it is referenced by the -keyboard or -pointer command line options. The most commonly used input drivers are keyboard and mouse. loaded for each active InputDevice section --- I haven't really read the section I just searched for InputDevice, whick is the section that caught my eye about this. -Jon
Re: SSHD-config
On Fri, Jun 01, 2001 at 03:07:11AM +0200, Frans Schreuder wrote: :Thanx for your trouble :I checked /var/log/auth.log on advise of H.H.Gebel :Changed an host entry (casesensitivity)of the FREESCO router. :More importantly, added client in host.allow. So it works now? Hmmm. on my system hosts.allow and hosts.deny are both empty (exept comments), which allows everything from everywhere (nobody get excited, I'm also running netfilter packet filtering :), and this is the default I think as I haven't touched them. These files only apply to services run out of inetd. I've always run sshd as a stand alone deamon (though there may be and inetd option on install I forget) The case sensitivity of host names is news to me, but after alittle testing it looks like /etc/hosts *is* case sensitive while DNS lookups are not, hum, live and learn :) I'm be interested to know if it still works if you remove the hosts.allow entry (assuming you haven't edited hosts.deny) :Thanx again for the fast response. :Sure to use this channel again I think this list is one of the best things about Debian (though apt is a close second). I owe it alot... -Jon
Re: Jumpstart install of debian?
Sorry I missed the start of this... I'm very near a solution for my Lab to do this very thing. It would be working (I think :) except that I broke the custom kernel I want so that one of the important NIC cards 'round here doesn't work. The concept works (aslong as you have a 3com or eepro card and not a tulip based card :) It's a single floppy network install. The boot floppy is syslinux based with the inportant network info in syslinux.cfg (it's a switched multi subnet environment and I only want one sever not six) The floppy kernel mounts root from an NFS server and then the installer script (#!/bin/bash) is run from inittab. The install.sh script does: assign_host_variables # from ifconfig and DNS lookup detect_drives # uses sfdisk partition_drives# uses sfdisk format_drives mount_drives base_install# cd $TARGET tar -xzf $BASE_TGZ mkfstab # with info saved from partition_drives mkhosts set_hostname mkinterfaces# setsup /etc/network/interfaces set_packages# chroot $TARGET dpkg-setpackages $PKGS do_lilo #get_packages #moved to post install ai_config # semi-random stuff mostly in /etc # also setsup self erasin second stage # installer /etc/rc2.d/S99install I broke off the second stage install because, some of the packages had trouble installing in the chrooted environment. It detects the NIC by attempting to modprobe all the network device modules untill one works then does: apt-get -qq -y dselect-upgrade apt-get clean anXious reset apt-get -qq -y dselect-upgrade The -y to apt-get is safe here as the configs are already inplace from ai_config. Gasp! It's convoluted, a pain to set up initially, and unpackagable because of all my site specific hacks and FAI may be in a place where it will do this for me (it wasn't when I started, but alot of developments gone on there and I haven't checked back), but damn it I'm *almost* there :) If anyone is up for some hacking I can go into more detail (off the list I think?) but this is quite long winded enough for now. A few months back there was talk on boot-flopies about this type of functionality, but I decided I couldn't live with RH that long (7.0 had just come out), and I haven't checked the state since, so that's another ray of hope. On Thu, May 31, 2001 at 06:45:44PM -0700, Krzys Majewski wrote: :with any luck, I can eventually brainwash my department to switch from :redhat to debian. (Lots of problems with redhat, for the record. Not :sure how much better debian would be, but at least debian truthfully :advertises their releases as unstable, rather than the ambiguous :7.1 (Seawolf).) Well, you see what lengths I'll go to to banish RH :) -Jon
Re: SSHD-config
On Thu, May 31, 2001 at 09:55:37PM -0400, Harry Henry Gebel wrote: :If sshd has been compiled with libwrap support (Debian's has) it will use :hosts.allow and hosts.deny . I think the default hosts.deny in Debian is :now 'ALL: PARANOID', at least that was the setting when I installed :Debian (in September). That'd 'splain it. In mine ALL: PARANOID is commented out shrug -Jon
Re: X :1 works but startx -- :1 doesn't anymore
On Thu, May 31, 2001 at 11:43:57PM -0300, Sergio E. Schvezov wrote: : :i've jsut fixed it i downgraded from xbase-clients_4.0.3-4_i386.deb 2 :xbase-clients_4.0.3-3_i386.deb did you file a bug report? see http://bugs.debian.org for howto info
NIS broken in unstable
Hi, I just upgraded NIS on my machine (unstable w/2.4.4) and now it won't bind to the domain, so I'm basicly screwed... Anyone know what happened? -Jon
Re: How to copy the binary image of a dos floppy ?
On Fri, May 04, 2001 at 10:17:13PM -0700, Karsten M. Self wrote: : [EMAIL PROTECTED] dd if=/dev/fd0 of=floppy.img bs=72k : :I'm not sure where you're getting your bs= value. Ok, I admit it's just what we use here, I don't know if there's any particular reason for it. To an extent the larger the block size the quicker the dd goes, though floppies are so slow anyway, I don't think it matters much. BTW bs=72k results in a count=20 on a 1.44 floppy That is the inverse of the command I use to make boot floppiesa and it's worked well. This is not to say anything against bs=1440 :) -Jon
Re: Minor Technical Difficulties
On Fri, May 04, 2001 at 09:56:28AM +0200, Alberto Cabello S?nchez wrote: : :For the machine don't start at graphics mode, it suffices to change the line (of /etc/inittab) : id:5:initdefault: :5 is the default runlevel, so you can put in it 3 (multi-user, no graphics mode). Unless I'm mistaken, the runlevel is irrelevant in Debian (unless you as the administrator configure it to be relevent), by default if xdm is installed it starts in all multi-user runlevels (2,3,4,5). RedHat (possibly others?), starts xdm in /etc/inittab rather than from a normal rcN.d script, and uses runlevel 5 for graphic mode, runlevel 3 for console mode. Debian uses runlevel 2 as the default. -Jon
Re: Minor Technical Difficulties
On Fri, May 04, 2001 at 02:52:45PM +0200, Danie Roux wrote: :I always wondered, why are there different ways of using the runlevels? :It took me a while to figure out why telinit 3 won't kill X! You can configure it to do this, by setting xdm (or gdm or wdm or whateverdm), to only start in runlevel 5 and be killed in all others. man update-rc.d for the Debian Way of doing this. You can also tailor the runlevels to do more complex things. For example my laptop uses runlevel 2 by default, this is basic stand alone mode. Runlevel 3 is configured to tie into the lab NIS and automount maps. -Jon
Re: I forgot my root Password
On Fri, May 04, 2001 at 09:12:21AM +0200, Alberto Cabello S?nchez wrote: : : At LILO prompt, type: linux 1 and press enter. This will cause linux to start at runlevel 1 (single-user) so you can run passwd. After that, don't forget your passwd again. Debian asks for a root password before entering single user mode, go with the LILO: init=/bin/sh # mount -o remout,rw / as recommended above. -Jon