Re: Initial install question
On Mon, 2005-06-20 at 03:32 -0500, Mark Anderson wrote: > Greetings. I just installed Debian using the ISOs from your site. > Everything went fine and then i got the "login" command. At this point > I typed in my username and password and what i got was this: > [EMAIL PROTECTED]: ~S: (the "name" being my username i made during > install). I can't get beyond this to the GUI. I have little knowledge > of command line use and none of Linux in general, which i imagine is > obvious by now :) I would appreciate help in how to get beyond the > point i've descibed above. Thank you. I would guess that you didn't select any desktop option during the initial install. That's usual when setting up a webserver or similar, but of course not so useful for general use! No problem, you just need to: * log on as user "root" (the superuser). You will have been prompted to create a password for this account as part of the install * type "apt-get x-window-system-core" to install the basic graphics system * type "apt-get install gnome-desktop-environment", or "apt-get install kdesktop" depending on which desktop you want. You can even type both to get a choice. I think that's all you need; after that rebooting should give you a graphical logon. I don't know if your graphics will be properly configured, though: you've skipped the hardware detection that's built in to the installer and I don't know how to manually run that. If you selected the Gnome desktop then you might need to go into "applications| Desktop Preferences|Screen Resolution" to select a reasonable resolution. If you chose the KDE desktop option then there will no doubt be a similar option somewhere. You might also want to try running "aptitute" as the root user. This does the same job as the apt-get program, but allows you to browse a tree of available packages (text mode of course!). And once you've got graphics running, there are a number of nice graphical applications to do the same job. Actually, it might be easier to try running the installer again this time making sure you select a desktop package... By the way, I hope you're installing Debian 3.1 (aka "sarge") not the old Debian 3.0 (aka "woody") - the new release is a major improvement! Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Good backup software for Linux
On Mon, 2005-06-20 at 12:36 +0300, kalasala wrote: > http://www.amanda.org/ Well, Amanda is really for backing up multiple servers to a single archiving device isn't it? As the original poster didn't state what he wanted backed up I presume it's a desktop machine. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Good backup software for Linux
On Mon, 2005-06-20 at 15:51 +0530, Siju George wrote: > On 6/20/05, Simon Kitching <[EMAIL PROTECTED]> wrote: > > On Mon, 2005-06-20 at 12:36 +0300, kalasala wrote: > > > http://www.amanda.org/ > > > > Well, Amanda is really for backing up multiple servers to a single > > archiving device isn't it? As the original poster didn't state what he > > wanted backed up I presume it's a desktop machine. > > > > Thanks simon for asking :-) > > I have a Sarge installed that I would want to run as a backup server. > this server will take backups from other Linux/BSD/Solaris/MS Windows > systems. The backup will be stored on hard disk (mirrored) for now but > later it will be taken to DVDs. Ahh - in that case kalasala was right - amanda may well be what you are looking for. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How to configure less to read *.gz files?
On Mon, 2005-06-20 at 11:47 +0100, Adam Funk wrote: > On my home computer `less foo.gz` automatically pages through the gunzipped > text (assuming foo is a text file of course). I'd like to get this > behaviour on another machine but I can't remember how I configured it to do > this (instead of `gunzip -c foo.gz |less`). > > I'd appreciate it if someone could remind me! man less and look for "Input Preprocessor" section... Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: compiling custom kernel/Sarge
On Mon, 2005-06-20 at 20:11 -0400, William wrote: > I need to compile a custom kernel, to add raid and a scsi driver. Is > the kernel that comes with sarge just from kernel.org or does it have > some kind of security patches? > If it is just from kernel.org, is it best to use the latest kernel? > thanks william > > Sarge comes with a 2.4.27 and a 2.6.8 kernel, with security fixes backported from later kernels. "apt-cache search kernel-source" will show what kernel source packages are available. I'm no expert on custom kernels, but installing the latest kernel could have interesting side-effects. Most thngs will work, but there can be breakage if bits like udev have changed etc. and you haven't updated the corresponding apps. It's probably safer/simpler to get the source for the standard Debian kernel and recompile that. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Etch experiences
On Wed, 2005-06-22 at 07:21 +0200, Hans du Plooy wrote: > Thanks for the replies, Peter and Rogério. > > The notebook will be my primary work horse. I haven't used Debian much > as a desktop, except on older PC with more minimalistic window managers, > but having seen a fair bit of Sarge, and being totally hooked on apt-get > install, I have decided to put that on the notebook. It's a P-III 1ghz, > 256MB ram, so I don't want to loaded with a too heavy distro... > > The thing is, I want it to be at least reasonably up to date at all > times, so it's really between etch and sid. I just don't know which > will be the most stable, as I imagine etch would undergo a few big > changes. While it may be considered heresy, you might also want to consider Ubuntu. Being based on debian you still get that apt-get goodness. However they have much more rapid release cycles (which is bad for servers, and probably bad for corporate desktops but good for individual technical users). I think life is likely to be very interesting for etch/sid users for the next 3 months or so as major changes get rolled out (gcc4, xorg, gnome/kde updates). Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Setting up a Java development environment the debian way.
On Wed, 2005-06-22 at 10:17 +0100, Alan Chandler wrote: > Ms Linuz writes: > > > Alan Chandler wrote: > ... > >>So my questions are: > >> > >>1) What tools do I need to develop the application. This includes code > >>editiing, build environment, unit testing, I tried to setup eclipse on my > >>workstation but there are unsatisfied dependencies (java runtime?). > >> > >> > >> > > > > Netbeans : http://www.netbeans.org > > the best java ide i see so far > ... > >>Obviously all this from standard debian packages if that is possible > > Netbeans is not a debian package. It might be interesting to look at it, > but what is the debian way? Personally I think you're buying yourself a world of pain if you try to do servlet/ejb/etc java development using the free java tools. They are definitely getting better and are useable for some tasks but aren't completely there yet. This doc was updated 3 june 2005 and so is presumably up with the latest developments: http://www.nl.debian.org/doc/manuals/debian-java-faq/ (I just found this via google). http://java.debian.net is also useful (I found this in the faq). I do a lot of java development on Debian, and simply download Sun's standard java for Linux and a standard Eclipse build and work from there. They just need to be untarred, and $PATH set appropriately. Of course junit is also very important, and possibly ant. I am aware that Fedora ships with a version of Eclipse compiled via gcj which is cool. But I am not aware of a debian package for eclipse. And anyway Eclipse 3.1 is coming out soon with some very useful new features for jsp/j2ee development. If I was writing a GUI app for linux I might be tempted to try java + the gtk java bindings + gcj to generate a native binary from java source. But not for servlet development. If you need more info I suggest you ask on the debian-java mail list instead of here. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Setting up a Java development environment the debian way.
On Wed, 2005-06-22 at 12:24 +0100, Adam Hardy wrote: > On 22/06/05 10:53 Simon Kitching wrote: > > Personally I think you're buying yourself a world of pain if you try to > > do servlet/ejb/etc java development using the free java tools. They are > > definitely getting better and are useable for some tasks but aren't > > completely there yet. > > I re-read that 3 times to make sure I had read it correctly. I > completely disagree! I use JBoss, Tomcat, Sun JDK, Eclipse, ant (or > maven) and I have no issues. > > Which non-open-source tools out there should I seriously consider > spending money on? I only meant Kaffe/Classpath/gcj and things of that level. They've come a long way in the last year or so and I look forward to the day they can be used as a substitute for Sun's JDK. It may not be very far away. But personally I would develop on Sun/IBM jdks and then port/test on Free platforms rather than develop on free platforms; developing on a platform you can't trust to be generally bug-free and feature-complete isn't nice. Sun'd JDK is only free-as-in-beer. Things like JBoss/Tomcat/Eclipse etc. are excellent tools, and at least as good as their commercial competitors IMO. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Setting up a Java development environment the debian way.
On Wed, 2005-06-22 at 18:21 +0200, Aurélien Campéas wrote: > Le mercredi 22 juin 2005 à 16:05 +0100, Alan Chandler a écrit : > > > > For the record, the static typing in Java is acknowledged as one of the > worst imaginable (in short : it is mandatory but you have to go through > its backdoor (read : defeat it) all the time to make working programs). It still provides more static type support than completely dynamically-typed languages like Ruby and Python. And if you use "generics" as provided in Java 1.5+ then this problem is resolved. > > OTOH, Working programs have been written in dynamically typed languages > for ages. Yes, industrial strength programs (whetever that means). > > And I don't know any serious study about this topic. All we have is > anecdotal developper experience/opinions. But you will be able to judge > how important is static typing by yourself, again, when trying > ruby/python/lisp ... This mail list isn't really the proper venue for this discussion so I'll just make one reply then shut up. All the following is of course just my opinion. Ruby/Python are excellent languages where the developers on the project are experienced and talented. However this isn't usually the case with any "industrial scale" development; the normal corporate IT department has people with a wide range of abilities, and in that situation the support of a statically typed language can be very useful (IMO). And anyway it is *extremely* difficult to sell non-technical management on development in Ruby/Python. Java's marketing has, however, raised its awareness level to a point where the senior suits will actually consider it. So from a purely practical point of view, if you are interested in learning tools that will be of use in a commercial environment then Java is much more useful to know than Ruby/Python. > > There is an aspect of Java that is quite misunderstood : it is a good > language / platform not because of some technical merits (it is really a > rehash of 60's technology) but because of the division of labor it > allows in the industry (and some other "social" or "political" > properties) : > * separation from the designers (or so-called software architects) that > draw UML pictures and the coders (so-called code-monkeys that do the > supposedly unintelligent translation from high-level pictures to PL > level) > * mass availability of Java programmers (thus supposed > interchangeability of those) > * good marketing pressure from Sun and other big players to make Java > the new COBOL > > All the hype about Java being an OO-language (a very poor one indeed) > and bringing encapsulation to programming in the large masks the fact > that the big organizations that use Java need to encapsulate, indeed, > the many mediocre (read : insufficiently trained on theoretical and > practical levels, thus cheaper) programmers that make their huge dev. > teams. I pretty much agree with this. Java is a middle-of-the-road compromise language. I don't think it's "a very poor" OO language, it's an average one. The only revolutionary thing about Java is the fact that it compiles to portable bytecode (well, it isn't actually the first to do that either). But, in a twist of fate, in Java's "server" niche that is almost irrelevant. And J2EE was horribly over-complicated, a fine example of design-by-committee. However under pressure from alternative nicely-designed frameworks like Spring and Hibernate, J2EE is becoming more streamlined and easier to use. J2EE 1.5 should be a significant improvement over the past. The servlet spec was always a nicer beast, simple and focussed, and hasn't changed much since its first release. And the Java language underlies all of these. Re Tomcat/JBoss: applications should always be separated into "presentation" and "business logic" parts. This allows experts on the "business logic" to work on parts without being concerned about how exactly the information is presented to the user. And it means that the presentation can be updated/modified without any concerns about breaking the critical business logic. In large industrial systems, "business logic" is implemented as a set of "Enterprise Java Beans" (EJBs). And the presentation is implemented as servlets/jsp-pages. Sun have published separate specifications for these two parts: the servlet spec and the ejb spec. And the two combined is called "j2ee". Tomcat implements only the "servlet" spec. JBoss implements the EJB support stuff itself, and incorporates Tomcat to deal with the servlet part (or Jetty which also is a project that just implements the servlet spec). However where there is little/no critical business logic in an application it is possible to use just the servlet level functionality to implement the application. So Tomcat (or Jetty or other) can be used on its own. And that is probably where you want to start - servlets are not trivial, but EJBs can be very complex to use. Anyway, you should probably start by b
Re: Both apache and apache2 (sarge) ?
On Mon, 2004-06-28 at 14:25, Kenneth Jacker wrote: > I'm trying to install 'phpmyadmin' and 'acidlab' on testing/sarge. > > The default web server installed was 'apache2'. Yet when I try and > install the above, dependencies want to force an install of 'apache' > (the fact that 'apache2' is already there seems to be ignored). > > Doesn't seem right installing *two* versions of 'apache'. > > Any comments, suggestions, and/or help is appreciated! Apache2 comes in several variants depending on which MPM (multi-processing module) has been compiled in. There is an MPM called the "prefork MPM" that essentially behaves like Apache1, in that a pool of *processes* is created. And there is the "worker" MPM in which a pool of *threads* is created within the same apache process. Unfortunately a number of apache extension modules are not thread-safe, and hence will not work with the apache2 "worker" version. PHP is one of them. So PHP and friends tend to just require apache1 instead of apache2 (anyway, that's what I gather from articles I have read). Provided you are using the "prefork" mode for apache2, and can find an appropriate apache2 module for PHP, it *should* work. However you will need to fiddle around a bit with the install as this setup isn't the "default" setup. You can probably find instructions on the PHP website. You can run "apache2 -l" to list the installed modules, one of which will be the MPM module that your installed apache2 instance is using. NB: I'm not a PHP user, so all the above is hearsay Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Unable to start x on fresh Sarge install
Hi, I recently installed SARGE from cd, on a PC without internet connection. I only had the first installer disk available initially, so installed the base system and that went fine. I later went back with the first 4 CDs of sarge (downloaded using JIGDO) to complete the install just using apt-cdrom and apt-get. All seemed to go fine, including "apt-get install gnome-desktop-enviroment". But when I then tried to run "startx", I got this message: /usr/X11R6/lib/X11/xinit/xserverrc: line 2: /usr/bin/X11/X: No such file or directory xinit: No such file or directory (errno 2): unable to connect to X server xinit: No such process (errno 3): Server error I installed "gdm" too, and rebooted. But I just got a message saying that X could not be started. I tried "dpkg-reconfigure xfree86-common", but it completed without prompting me for any input, and the situation remained the same. Any ideas what I need to do to get X running? Thanks, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: syslog
On Wed, 2004-06-30 at 12:18, Ketil Froyn wrote: > On Wed, 30 Jun 2004, Matthew Joyce said: > > > I have been asked by a 3rd party (who managed some > > comms equipment for us) if we have a syslog server. > > > > If so, they say they can direct some logs to it. > > > > We have some Debian boxes, don't they all have syslog > > running ? > > > > How does it work ? > > man syslogd, look for "-r". If set, syslogd receives messages on the > network from whoever sends them, and puts them in the local log. Makes > it easier to log from devices without a disk, and makes it harder for > a hacker to hide what he has done (he'll need to break in to the > syslog server too). I'm no expert, but I see that in /etc/init.d/sysklogd there are some comments about how to enable the "-r" option. Once you restart syslogd with /etc/init.d/sysklogd reload I presume that any remote messages received should then show up in /var/log/messages along with the messages logged by local apps via the syslog daemon (syslogd). Apps like "swatch" can then be used to monitor the logfile and generate email alerts or send messages to pagers. If you need info on these options, I am sure someone on this list can explain the best approach better than I can. By the way, can anyone explain why the "syslogd" daemon is started/managed via an init script called "sysklogd"? Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: fast, tabbed, gnome/wm- compliant terminal
On Thu, 2004-07-01 at 16:27, Zenaan Harkness wrote: > On Thu, 2004-07-01 at 14:20, Zenaan Harkness wrote: > > It gets laggy. Inconsistently. When using vim inside gnome-terminal. > > For example, in the topmost line (again in vim, with gnome-terminal at > 87x98 chars) it is really noticeable, and the CPU hits 100% just holding > the cursor key down. > > It really slows you down and is very frustrating after weeks of putting > up with it. > > It could just be a Debian unstable thing... I'm running a mix of Debian testing and unstable. I've got gnome-2.6 with gnome-terminal 2.6.1 (installed via apt-get of the unstable gnome packages), and everything works fine. No lag at all when using vim inside a terminal. I suspect the problem is specific to you... Regards, Simon > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: program starts with 3 threads
On Fri, 2004-07-02 at 16:09, jono wrote: > testing/unstable 2.4.25 i686 > > When I run a daemon (written for a telephony card) or the expect > application which uses the daemon, they both start with 3 processes. > I've noticed the same behavior with S20xprint and nautilus. > > The same daemon and application on redhat9 (2.4.20-8 i686) starts with > only the one process. Its causing errors in our application so I was > wondering if there was any reason for this and if it is debian specific? > > TIA, please cc direct. Regards, Jono. > Hi Jono, My memory on this is a little rusty now. However I believe the story goes as follows: With normal versions of the Linux 2.4.x kernel, the "ps" command shows each *thread* as a separate process. The new Linux 2.6.x kernel has a completely rewritten threading implementation and "ps" shows only 1 entry for a multi-threaded process. RedHat has a tradition of distributing fairly heavily patched kernels; I believe that they included the new threading model into the 2.4.x kernel that was included with RH9. I suspect the app you are concerned about is starting *1* process with *3* threads. If that is the case, then what you are seeing on Debian is normal for distributions running 2.4.x kernels. RH9, despite running a 2.4-based kernel behaves more like 2.6.x due to special patches in its kernel version. I remember this because RH9's early adoption of the new threading implementation broke many many applications (eg IBM MQSeries), and I cursed them for it vigorously. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Re: program starts with 3 threads
On Fri, 2004-07-02 at 16:50, jono wrote: > Thanks Simon, that makes a lot of sense. I've just read the ps man page > but cant find out if I can show only processes not threads. Any idea if > this is possible or any another tool that could do it? I'm not personally aware of any easy way to tell threads and processes apart on 2.4.x kernels. All of the "thread" entries reported by "ps" should be traceable back to the real process PID via their parent PID, but I don't know how to tell the difference between this and a tree of true child processes. I believe that in 2.4 the threads are essentially separate processes - they just happen to share the same virtual->physical memory map. Because the kernel keeps separate entries in the process table for them, ps sees lots of processes. And "ps" doesn't bother to look into whether they have the same virtual memory mappings as their parent process (ie are threads) or not (are traditional child processes). Maybe there is a way to tell by looking in /proc... I'm sure others have experienced this issue. Google could well be your friend. I would try "NPTL" (Native Posix Thread Library); I think this is the code that ended up in 2.6.x, and is probably the code that RH included in 2.4.x as a vendor patch. > > Other than that I suppose a kernel patch or upgrade would do it. Yes, I expect that running kernel 2.6.x on debian would do it. I don't know if that's an option for you. If you are running a "production" server then maybe not... I certainly wouldn't recommend trying to patch 2.4 to include an alternate threading model like Red Hat did :-). Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: problem with invisible/disappearing files - but inodes there
On Mon, 2004-07-05 at 16:52, [EMAIL PROTECTED] wrote: > Hi, > > We have an unusual problem with 'invisible' and disappearing files: some files are > not > visible in a directory using 'find' or 'ls' on the directory, yet they > are visible using tools such as 'debugfs' or if you know the actual name > of the file, you can use 'ls' to list it explicitly. > > For instance, in the directory: > > /home/stampy/queues/sendsyphonemailqstore/q_dispatch > > you can ls -la it, and find no files: > > > stampy:~# ls -la /home/stampy/queues/sendsyphonemailqstore/q_dispatch > total 8 > drwxr-sr-x2 louisb staff4096 Jun 25 11:45 . > drwxr-sr-x7 louisb staff4096 Jan 4 14:05 .. > > Yet when you 'ls' the files directly (which we find from the log of a > daemon that places them there), they are there: > > stampy:~# ls -la > /home/stampy/queues/sendsyphonemailqstore/q_dispatch/1088004896841 > -rw-r--r--1 louisb staff 1 Jun 24 01:34 > /home/stampy/queues/sendsyphonemailqstore/q_dispatch/1088004896841 > stampy:~# > Very bizarre. I see that when you do an ls, the following line has: total 8 So it appears there are actually 8 files in that directory. Perhaps there is a file with a very strange character in it, which is causing ls to terminate? Why not try: find q_dispatch -print >/tmp/files.txt then use "od -cx" to see what is really in that dir listing? Look in particular for an "empty line", which might indicate a problem. The "od" program is used to handling any sort of input, so you will definitely see everything generated by find. You might also want to try rm -i * to "interactively" select files for deleting. I have found this useful in the past when a file with a bizarre name needs to be deleted. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: what's the diff between 'aptitude upgrade' and 'aptitude dist-upgrade'?
On Mon, 2004-07-05 at 21:02, Enrico Zini wrote: > On Mon, Jul 05, 2004 at 02:22:40AM -0500, dircha wrote: > > > An equally adequate reference is available as: > > /usr/share/doc/aptitude/README > > ...which is not too hard to paste: > > aptitude dist-upgrade > > This command will also attempt to upgrade packages, but it is more > aggressive about solving dependency problems: it will install and > remove packages until all dependencies are satisfied. Because of the > nature of this command, it is possible that it will do undesirable > things, and so you should be careful when using it. I saw this when I first started to use debian and it scared me. It took me quite a while to realise that when running testing or unstable, you *normally* use dist-upgrade. So I think this bit of info is misleading, and should be qualified in the README to point out that this advice is excellent for production servers and other systems that absolutely must remain functional, but that lots of people happily use dist-upgrade regularly for their home systems. Option "upgrade" is for when you want to apply security patches and minor bugfixes, without new features, right? It never installs new packages or removes old ones, so it is limited both in the stuff it can add and the damage it can do. And dist-upgrade is for when you want to get all the new funky features. Which many of us do :-). It only breaks things if the packages uploaded by the maintainers are broken. Ok, that occasionally happens, which is why you don't use "dist-upgrade" on your production servers except with great care, like backups and rollback plans. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: installing Debian and installing programs
On Wed, 2004-07-07 at 08:28, Jim Knott wrote: > I bought a distro from someone who sells a lot of Debian and when the install > failed, they even sent me a second set in case the first one was no good. The > installs went good up to the 2nd and 3rd disc's respectively, but wouldn't > install all the way. I know I don't understand all of the technical jargon in > the install, so maybe I did something wrong. I tried the install directions > on Debian.org, but that didn't help either, too cornfuzin! Is there anyone > that can tell me how to install, using plain english, step-by-step? Also, > when I do get it installed, I want to be able to use the command line to > download from tucows or others and install programs on my Linux computer, can > you provide step-by-step instructions for that also? I really want to learn > Linux, but the books I have bought are not for new people, they are for > someone who understands all of the tech jargon. Thanks, Jim > [EMAIL PROTECTED] > Hi Jim, Are you using the new "debian-installer" program (which installs the "testing" version called Sarge, or the "unstable" version called Sid)? Or are you installing the (rather old) stable version of Debian called "Woody" (aka Debian 3.0)? [I presume you've noticed the Toy Story naming scheme :-] If you are only trying to install Debian for "learning" and home use, then using the new (and much nicer) "debian-installer" program is a better choice than the old installer that comes with the Woody (3.0) distribution. It's still officially in "beta", but it works great. I have tried debian several times over the years, and only when the new installer arrived did I ever manage to get it properly installed. You will find info on the new debian-installer here: http://www.debian.org/devel/debian-installer/ If you don't have a good network connection, then a reseller should still be able to provide CDs with a recent "Sarge" snapshot, even though it's not yet an official release. Of course if you're intending to set up a production-quality server, you should stick with the official and stable version 3.0. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Cannot login or su
On Wed, 2004-07-07 at 19:11, CN wrote: > Hi! > > I have installed woody in 3 boxes. Suddently I can't login to one of > them. > I am not sure if this problem has to do with the upgrade of wine package > (version 20040615). > > Now I have only 2 ways to access this problematic box: > (1) using boot floppy disks > (2) starting X by sudo in startup script and using xterm from within > window managers > > However, I can't su from non-root accounts in xterm, either. su yields > error message: > > su: Module not found > Sorry. > > sudo seems to be working fine. > When I try "telnet localhost" from xterm in the same box, the system > prompts me "login:", but after any account followed by enter key (line > feed) are entered, the system does not prompt me password. It instead > directly responds with login failure. > > If I telnet from other boxes and give the problematic box the login > account, the problematic box does not prompt password, either. Then it > says: > > System bootup in progres - please wait > > Files passwd and shadow look good. > > How to fix this? Please! Hi, I wonder if "pam"(the "pluggable authentication modules") has somehow got stuffed up? That might explain the "su: Module not found" message. In directory /etc/pam.d I have a number of files. The "login" one is critical to allowing user logins. There's lots of comments in these files, but the critical files/lines are: file login: auth required pam_env.so @include common-auth @include common-account @include common-session @include common-password file common-auth: authrequiredpam_unix.so nullok_secure file common-account: account requiredpam_unix.so file common-session: session requiredpam_unix.so file common-password: password required pam_unix.so nullok obscure min=4 max=8 md5 And of course the pam_unix.so, pam_env.so, etc files need to be available. On my system they are in: /lib/security There's no guaruntee that pam is the problem, just my guess.. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: help; Is C soon to be the programming lang. of the past?
On Thu, 2004-07-08 at 14:41, Paul Tsai wrote: > C# is similar to Java than it is to C, and some people may argue that in > the future, C# may replace Java. Since C# is .NET, it requires a CLR > (or CLI, etc) to run its programs much in the same way Java requires a > virtual machine. However, .NET has a Just in Time compilator that is > able to covert code from CLR code (aka MSIL code, I will use MS names > from now on) to the machines native machine code, so it's slower than > Java the first time the program loads, but after the program is loaded, > performance is not really an issue, compared to Java being interpreted > thoughout the lifetime of the program, hence slower. There is much more > to C# and .NET if you are truely interested. Do a google or browse > MSDN, or the Mono Project's documentation. Pretty much true, although I think that Java/C# are more likely to co-exist than for either to completely eradicate the other. C# has Microsoft's marketing muscle behind it, but Java has a large installed base, a large group of experienced developers, and many supporting products and libraries from commercial and open-source suppliers. But Java has had "just-in-time" compilation for many years now. In fact, it's had this feature for longer than C# has existed. And both can also be pre-compiled, similar to c/c++/etc. [NB: compiled java or C# is still not quite the same as compiled c/c++/etc, as even when precompiled they need some "runtime" support libs]. As an experienced Java developer who has also studied C#, I can say that learning either is a good idea. Once you know one, learning the other is a fairly small step. But as a first language, I would recommend Ruby or Python over anything else, Java or C# as a second choice, and "c" as a distant third or even later, unless you seriously intend to work on certain open-source projects that you know are written in "c". Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: grub problem
On Thu, 2004-07-08 at 16:27, Rodney Richison wrote: > I attempted a repair floppy last night. It didn't work out real well. > Ahhh, for a command like sys a::) Novel idea... > > Anyway here's what I did for future google searches. Maybe it'll help > someone someday. > > How to Repair Grub boot loader after debian ghost restore > > I booted with mepis (Any live-cd would probably do) > Open a terminal window > Mounted the drive (don't use the mepis icon as it mounts read only) > mount -rw /dev/hda1 /mnt/hda1 > chroot /dev/hda1 > grub-install /dev/hda > > Worked like a charm.. Am a bit confused what the drive is hda1 till I > get to the install part, but I'm guessing I'm mounting partition one on > hda, but the grub wants the "drive" to install to. Not the partition. Hi Rodney, The standard procedure for booting on the PC architecture is that the BIOS is configured with a list of boot devices and the order to scan them. It checks the MBR (master boot record, aka "boot sector") of each device in turn until it finds some executable code. Note: the BIOS does not care at all about the partitions on the drive, it only looks at the "boot sector" for the device. What grub-install does is write a small amount of code (512 bytes) into the MBR on a device, called the "stage1" code. This code is just smart enough to be able to read a sequence of hard-wired disk blocks (the "stage1.5" or "stage2" executable) into memory and jump to the loaded code. This data resides as a standard file in some kind of filesystem. Note that no filesystem knowledge is required - but the file being loaded had better not move on disk (eg defragmentation is a really bad idea) as the physical location of the blocks is wired into the MBR. That is why grub-install takes a "device" (hence /dev/hda) as a target, not a "partition"; it is installing the stage1 code into the MBR. And it needs to figure out the physical disk blocks which contain the executable which is the next "stage" of grub. So you can either mount the filesystem that will contain these files as the root partition (which you effectively did), or mount it elsewhere and pass the "--root-directory" option to grub-install. Of course here you are dealing with *filesystems* which reside on *partitions* (hence /dev/hda1). I hope this helps. I learnt much of this the hard way after having to fix unbootable systems just like you are doing now :-) The grub docs at http://www.gnu.org/software/grub/manual/grub.html are of course the ultimate source of info. NB: To confuse issues, each partition also has a "boot sector". The BIOS never looks at these, but a bootloader stored in the MBR of the device can be set up to delegate to the boot code in a partition's boot sector. BTW, there may be errors in the above - anyone who knows better is welcome to correct these. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: new gnome file selection
On Thu, 2004-07-08 at 17:51, [EMAIL PROTECTED] wrote: > I just started using sid, and I've run into the new gnome file > selection interface. It's almost unusable: there's no way to enter > an absolute path (or any path for that matter), and no way to go up in > the file system, or to even see any dot files. Is there a way to > change all this? It's horrible! > > I don't know which package is responsible for this mess, so I can't > report the version; and I don't know what to downgrade if that were > the answer. > > I'm sure this must have been discussed to death somewhere already, but > I wasn't able to find anything useful with google. I'd appreciate any > pointers to more information. I can't believe the gnome developers > would leave things like this. To go "up", look at the top right of the dialog. There is a sequence of buttons that indicate the directories in the current path. Clicking on one of them changes directory to that point. To enter an absolute path, type ctrl-l. There's no entry field for it because "basic" users are expected to point-and-click, and "advanced" users are expected to be able to type ctrl-l. As for dot-files, try "ctrl-l . tab". A name-completion dropdown should appear. Again, "basic" users won't be viewing dot-files. All this is gnome-2.6, *not* anything debian-specific. *Please* do not start flaming all things gnome-2.6-ish until you have got some familiarity with the new setup. Any change is a little annoying at first, but software can't be frozen forever because of this. If something still annoys you after you've given it a good try for a week, then is a good time to do some research about the topic, then submit a polite comment to the gnome hackers. [NB: Personally, I would have liked a simple toggle to show a text input box. But then again, that would clutter the interface and *still* most people wouldn't realise it was there.] Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: gnome tool bar animation - turn it off ????
On Fri, 2004-07-09 at 10:01, Zenaan Harkness wrote: > How do I turn off the "sliding" effect of gnome's toolbar (when I click on one side > button, it "slides away off the side of the screen". > > It frustrates the hell out of me: I want an instant on/off thing, not any Windows 98 > style try-hard hope-the-user's still like me today! Bullshit, that's what it is. Then feel free to fix it. The code is open-source. If you won't or can't fix it yourself, then stop bitching about a *free gift* to you from people who actually can and do write code. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: ping server
On Fri, 2004-07-16 at 16:50, John Summerfield wrote: > debian user wrote: > > > i would like to set up a ping server for my office. I however have > > not found a simple way of recieving a responce of "dead" or "alive". > > I am not to concerned about time etc? I just want to ping a few > > servers and then send the results out in a email to a few key > > personal. Right now I am doing "echo ping -c 3 server-name > file" > > for half a dozen boxes then i email the file to sevel people > > periodicly during the day. I would like it to be a little eaiser to > > read i.e. "dead" or "alive" after the server name. > > thank you, > > alex > > > > > Don't reinvent the wheel. Install mon. >From the description above, I think the OP wants *statistics* about the availability of servers, rather than an alerting system for server failure. There are several good free monitoring packages, including "nagios", "big sister" and mon. Nagios: http://sourceforge.net/projects/nagios Big Sister: http://sourceforge.net/projects/bigsister But for nagios at least, gathering statistics is not its main purpose. It can be done, but requires a fair bit of reading of manuals to set up. I'm not familiar with mon or Big Sister, so don't know how good they are at statistics. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help
On Mon, 2004-07-19 at 15:29, Yossi d Esrig wrote: > I wanted to wipe my laptop (ibm notebook) and put linux on it. I can't > figure out which distribution is right for me. I just need it for > personal use. I actually only need it for the C compiler. (As of right > now, I'm having techinical dificulties installing a c compiler on my PC) > I also do not have internet access on the notebook. > Also, once I download linux, how would I go about intstaling it? My > laptop is quite annoying > in the fact that it can hold either a 3 1/2 inch drive OR a cd drive, but > not both > at the same time, without turning it off in between. > Please help. Thank you. Well, as you asked on the Debian Linux list, the answers may be a little biased :-). Note that all that follows is just my personal opinion. In fact, you're likely to get half a dozen contradictory replies to this question. Good luck sorting them out. Options considered: Fedora, Mandrake, Debian, Slackware, Gentoo, Gnoppix/Knoppix. Assumptions: You want some kind of GUI. Because of your lack of network connectivity, you'll need a distribution on a CD. And I presume because you're asking this question that you are fairly new to linux so would prefer something user-friendly. Debian has excellent features for keeping your OS up to date over the internet, but that's obviously not an option for you. And because Debian's stable release is now so old, a copy of debian 3.0 on CD is not really appropriate for a personal (non-server) PC. You can get a copy of Debian's "testing" release on CD, but without network connectivity you won't be able to update anything that's broken. Its not likely that critical stuff will be broken, but with a "testing" release, nothing's guarunteed. Because of the CD-based approach, I think one of the first 2 on the above list will suit best. If your laptop is oldish (<256MBytes ram, lower than P3 1GHz?) then Fedora will probably not suit. The latest releases are pretty heavy users of memory and CPU. Personally, I would therefore recommend that you obtain a copy of the latest Mandrake release. It's user-friendly, not *too* heavy on resources (I believe), reasonably up-to-date. Or possibly you could select Knoppix or Gnoppix. These are Debian-based distributions that can be run directly from CD (not installed onto a hard drive). But they *can* be installed onto harddrive too. This will allow you try out Linux on your laptop first before doing the install. Of course, if you like the philosophy of open-source, then you may wish to select Debian even though it isn't quite the best match for your requirements. Debian is a community as well as a distribution, and running debian will allow you to participate in the development of a linux distribution far more easily than other distributions will. And if you intend to get into some serious linux work later, then starting on Debian may be a good idea. Debian provides a reasonable desktop OS but really comes into its own when you need robust secure servers etc. In the end, just about any Linux distro you can get on CD will do the job for you. They all come with compilers. And anything at all modern will boot directly from CD (provided your laptop supports this) so you won't need to plug in the floppy drive at all. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Nagios vs Netsaint
On Tue, 2004-07-20 at 10:14, Jacob Friis Larsen wrote: > I see there are both Nagios and Netsaint packages. > Which one should I use, or can I use both on the same server? > Is it possible to use Netsaint plugins with Nagios? Nagios is the "new" name of what once was NetSaint. All new releases are done using the "nagios" name. So you should definitely be using "nagios". I am, and it all works great. As far as I am aware, the nagios-plugins package contains all the functionality that was available with the old NetSaint plugins, so there's no need to use any of the old/obsolete NetSaint stuff. See www.nagios.org Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Re: Experiment: Neophyte versus Windows XP
On Tue, 2004-07-20 at 17:46, Mark e Plummer wrote: > Hello, > I have just read your article as best I could. I am confused, but not > by you but by me. > I have been using Firefox as a browser and Thunderbird as mailbox for > about a week and a half now and I love them. > I am thinking about dumping Windows XP because it is a pain, I was > looking at Mandrake and ReHat. Debian Woody I have never heard off > until five minutes ago. > Is it really hard to understand. Or as I am not a programmer should I > even be thinking about using it. > This is a vague letter I know. > What I am asking, I suppose, is would you go for it. Do I have to dump > Windows before I start downloading Woody? > Any help in the form of ideas would be wonderful. Hi Mark, You should definitely give Linux a go - it's got some great features, and gets better by the day. Provided you have a reasonable-size hard drive, you can have multiple operating systems installed on the same PC by "partitioning" your disk drive. Or if you're short of disk space, you can always buy another disk drive to experiment with. It is probably better to do this than simply wipe Windows; like all drugs, going cold-turkey can be hard :-). With multiple operating systems installed (often called "dual boot" or "multi-boot"), when you turn your PC on you get a list of the installed operating systems to choose from, including your existing Windows setup. There are even Linux versions that run direct from CD without needing to be installed at all, but that is probably not what you're looking for. If you've got Windows 95/98/ME right now, then you can simply run a "disk defragmentation" from inside Windows, then boot from a linux install CD and follow the instructions to split off some spare disk space for the use of Linux. If you've got WindowsNT, 2000 or XP, you'll probably need the commercial tool "partition magic" to create a partition on your existing drive. Or buy that new drive I suggested. There may be a way to partition your existing windows drive with free software, but someone else will have to tell you how, because I don't know of a safe way to do this. And best of all, if you keep your existing windows files around you can access them from Linux. Not your old emails, unfortunately (Microsoft makes that very difficult), but certainly your MS-Word documents, your photos, your mp3 music files, etc. About 6 GBytes is a good amount to use for a linux install. Of course if you can spare more, then that is good - you'll need somewhere to put your music and photo collections ;-). Regarding which linux distribution to install, both Mandrake and RedHat are good choices. Debian is a little harder to use initially, but more powerful once you get to know it. As your first entry into Linux, you might be better off with Mandrake or "Fedora" (what RedHat call their free version). But Debian is ok to start with, too. The "woody" (3.0) version you mentioned, though, is pretty old. If you choose Debian, you will probably be better off downloading the "debian-installer" program from http://www.debian.org/devel/debian-installer/. This is "beta" software, ie not officially released, but it's a lot easier to use than the "woody" (aka 3.0) release of Debian. This will install everything over the internet, so you only need to make one CD. However you'd better have a good internet connection... I suggest you use "google" to search for articles about installing and learning linux. You should find plenty of information there. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Is Linux Unix?
On Wed, 2004-07-21 at 13:16, Paul E Condon wrote: > On Wed, Jul 21, 2004 at 08:35:11AM +1200, Steven Jones wrote: > > hehand at the rate Linux is taking over Unix space he like RISC based Unix > > will be going the way of the dodo. > > > > ;] > > > > Considering how many "Unixes" run Open source applications and utilities I find it > laughable that ppl can look down on Linux. Simple answer is Linux is not Unix > because its better > (or going to be). From an operational point of view I have found that there is > little Unix can do > that Linux cannot, and certainly plenty Linux can do that Unix cannot, kernel > compatibility and > cost effective hardware see to that. The LWN report on the first day of the Linux Kernel Summit gives a very interesting perspective on what Unix can do that Linux still can't. If you're an LWN subscriber, the article is available now. For non-subscribers, it will be available for free next week. See http://lwn.net. The projects at www.osdl.org (Linus Torvald's current employer) also give an interesting perspective on the features Linux still needs in order to catch up with commercial unix in some areas. Of course in many areas it kicks commercial linux. I have to work on AIX as well as linux, and there's many times a day I wish this was a pure linux shop. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How to de-Grub?
On Thu, 2004-07-22 at 12:47, Stephen Cradock wrote: > Then, as one does when playing with Linux, I went stark staring bonkers, and > decide to try Gentoo - a friend told me it was much easier to install. > Rather than make yet another partition, I wiped the Sarge partition and > installed Gentoo there, from a CD. That went OK, but also failed to detect > my ethernet card. So it's back to Windows to go online to find out what to > do next - URGGGH - Grub failed, of course - I had wiped out menu.lst when I > deleted Sarge. So I couldn't get into Windows, or online with Gentoo. > > Now I want to remove Grub - it has gone and installed itself in my MBR, and > I don't know how to get it out. Any help? Well, your machine won't boot from the hard-drive unless the MBR holds code that points to *some* bootloader *somewhere*. So you'll need one of Grub, LILO or the Windows bootloader installed. Doesn't the Gentoo install process complete, even if it can't find your network card? As long as it installs a base system, it should install Grub or LILO, and set up the appropriate config to allow you to boot either Gentoo or Windows. Or is the Gentoo install really a network-based install, without even a base OS on the install CD? I presume that your Windows install disks are capable of rewriting the MBR to point to the Windows bootloader on your Windows partition, by selecting a "repair" option. Another option is to reinstall a basic Sarge system. I think it should be possible to use the Sarge disks as a "rescue" CD (or download and burn a real "rescue" CD). You can then boot from the CD to a basic linux OS, reformat a partition as ext2 or similar, and install *just* grub on that partition. At that point, rebooting should give you a grub menu with just one option: windows. But if you're going to do that, you may as well reinstall the Windows bootloader. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How can I make sure that Linux have initialized all CPUs?
On Thu, 2004-07-22 at 13:07, Greg Folkert wrote: > On Wed, 2004-07-21 at 18:55, Jacob Friis Larsen wrote: > > Greg Folkert wrote: > > > On Wed, 2004-07-21 at 18:19, Jacob Friis Larsen wrote: > > > > > >>I still only see 1 CPU. > > >>On Redhat I can see all 4 CPUs. > > >>Would that be possible? > > > > > > Please do a: > > > > > > uname -a > > > > > > Copy and paste the result please. > > > > react1:~# uname -a > > Linux react1 2.4.26-2-686 #1 Mon May 17 22:19:48 EST 2004 i686 unknown > > Nope... you do not have an SMP capable Kernel. > > What distro are you using? What version? If you do "apt-cache search kernel-image" you can see all the prebuilt kernels available for install. The ones with "-smp" are the smp-enabled ones. The command apt-get install kernel-image-2.4.26-1-686-smp should install that *smp-enabled* kernel, and reconfigure grub/lilo to provide an extra menu entry at boot time to boot into that kernel [at least it does with Sarge]. Of course you have to select the appropriate kernel image for your CPU type, etc. NB: I haven't used smp kernels myself. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How to de-Grub?
On Thu, 2004-07-22 at 13:35, Robert William Hutton wrote: > Stephen Cradock wrote: > > > I'd recommend that you install LILO using the first woody CD. LILO, unlike > GRUB, installs itself exclusively in the MBR, so you don't need some other > partition for the stage 2 loader which GRUB requires. Are you sure? The MBR is only 512 bytes. I'm pretty sure that LILO gets the MBR to point to the "real" LILO data residing inside a linux filesystem, just like Grub does. You don't get much code inside 512 bytes... I believe the MBR simply holds a list of block offsets within the disk, and enough code to load the data in those disk blocks into memory and jump to them. The loaded data normally resides as real files in a filesystem (which had better not try to move them, eg via "defragmenting" the disk). The difference (AFAIK) is that Grub has the option of getting the MBR to "raw load" a simple "stage 1.5" app that knows about filesystems, and loads the real bootloader "stage 2" using proper filesystem semantics. Lilo instead essentially puts all its logic in the "stage1.5" file. Quite what benefit Grub gets from the separation of stage1.5 vs stage2 I don't know. And I see that "grub2" is going to do away with the "stage1.5" concept. Potentially, I guess, the LILO bootloader (and grub stage 1.5) can reside in a disk partition that has never been formatted as any kind of filesystem, because the data is fetched using raw disk block addresses. But who ever does that? Maybe it's useful for embedded systems.. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: USB disk drive
On Thu, 2004-07-22 at 14:33, Bill Moseley wrote: > I'm trying out a FireLite USB drive. It's been used on Windows (from > what the owner tells me and has data on it). > > I'm running kernel 2.6.6. > > When I plug it in I see this in syslog: > > Jul 21 17:18:50 bumby kernel: usb 1-2: new full speed USB device using address 2 > Jul 21 17:18:51 bumby kernel: SCSI subsystem initialized > Jul 21 17:18:51 bumby kernel: Initializing USB Mass Storage driver... > Jul 21 17:18:51 bumby kernel: scsi0 : SCSI emulation for USB Mass Storage devices > Jul 21 17:18:52 bumby scsi.agent[1675]: disk at > /devices/pci:00/:00:11.2/usb1/1-2/1-2:2.0/host0/0:0:0:0 > Jul 21 17:18:52 bumby kernel: Vendor: TOSHIBA Model: MK6021GAS Rev: GA02 > Jul 21 17:18:52 bumby kernel: Type: Direct-Access ANSI SCSI > revision: 02 > Jul 21 17:18:52 bumby usb.agent[1642]: usb-storage: loaded successfully > Jul 21 17:18:52 bumby kernel: USB Mass Storage device found at 2 > Jul 21 17:18:52 bumby kernel: usbcore: registered new driver usb-storage > Jul 21 17:18:52 bumby kernel: USB Mass Storage support registered. > > And modules seem to load correctly: > > bumby:~# lsmod > Module Size Used by > usb_storage29696 0 > scsi_mod 81152 1 usb_storage > lp 10564 0 > uhci_hcd 30672 0 > ohci1394 34756 0 > ieee1394 108340 1 ohci1394 > i2c_sensor 2944 0 > > And it shows up as expected: > > bumby:~# cat /proc/scsi/usb-storage/0 >Host scsi0: usb-storage >Vendor: SmartDisk Corp. > Product: FireLite (USB 2.0) > Serial Number: 00010f6f > Protocol: Transparent SCSI > Transport: Bulk >Quirks: > > bumby:~# cat /proc/bus/usb/devices | grep FireLite > S: Product=FireLite (USB 2.0) > > And I've got my device files: > > bumby:~# ls -l /dev/scd0 /dev/scd1 > brw-rw1 root cdrom 11, 0 2002-03-14 13:54 /dev/scd0 > brw-rw1 root cdrom 11, 1 2002-03-14 13:54 /dev/scd1 > > How do I tell what /dev/scd* the device connects to? > I tried installing the sg3-utils: > > bumby:~# sg_scan -i > bumby:~# > > And I can't mount: > > bumby:~# mount -t vfat /dev/sda1 /mnt > mount: /dev/sda1 is not a valid block device > > I tried both of my USB ports. > > > At this point I'm swinging in the dark. > Any ideas? I have an Apacer Handy USB flash device. I have to do this on my Sarge install with kernel 2.4: modprobe usb-storage mount /proc/bus/usb before the "mount -t vfat /dev/sda ..." works. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Debian install breaks on 'Configuring Locales'
On Fri, 2004-07-23 at 09:15, John Hechtman wrote: > Now I've dl'd the Debian CD iso images and burned them to disks. > This is with the 'Woody' 30r2-i386 set of seven CD's, plus the updates > CD. Are you intending this installation to be used for a production server, or for a personal desktop? If you want a production server, you'll have to persist with Debian 3.0 (aka woody). The install procedure is difficult, but the resulting installation is very stable and secure. If you want a desktop, or a more experimental server, then you should download the installer for the Debian Testing release (aka "Sarge"). The packages installed are far more up-to-date, and the install procedure is light-years ahead of the old installer. The only disadvantage is that the occasional package doesn't work (the dangers of using what is effectively a "beta" release). Search for "debian-installer" on the www.debian.org home page for links to the latest releases. Note that the installer is just a single CD, and the rest of the OS gets downloaded over the internet, so you'll need a decent network connection if you want to do this. There are ways to download CDs of "testing"; see documentation for "jigdo" which should be linked somewhere from the main page. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: sunone_asp4.0.1-linux-full
On Mon, 2004-07-26 at 13:10, Support wrote: > Hi! > > Can Debian run asp script in apache ? > > As I know sunone_asp4.0.1-linux-full only run on RedHat 7.3 .. can it run > on Debian ? > FYI, the Mono project provides support for ASP.NET on linux: http://www.mono-project.com/about/technical.html#asp Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Safely Upgrading Packages
On Mon, 2004-07-26 at 18:43, Michael B Allen wrote: > I've been running Debian on the net for a while. I thought it's time to look > at keeping packages up to date. But when I run apt-get update: > > # apt-get upgrade > Reading Package Lists... Done > Building Dependency Tree... Done > The following packages have been kept back > apache apache-common autoconf debconf debianutils e2fsprogs file fileutils > libgd2-noxpm mailman mysql-client > mysql-server php4 php4-mysql php4-pear shellutils textutils > 40 packages upgraded, 0 newly installed, 0 to remove and 17 not upgraded. > Need to get 12.0MB of archives. After unpacking 2192kB will be used. > Do you want to continue? [Y/n] n > Abort. > > Why are packages being "kept back". These are precisely the packages I want > to update. Because the new versions of those packages have new dependencies, and you don't have the new dependencies installed. That's the difference between "upgrade" and "dist-upgrade"; upgrade never installs new packages, for security reasons. But dist-upgrade will install any new stuff you need to satisfy upgrades of existing packages. In this example, the latest version of apache requires libmagic. You don't have a version of libmagic installed, so the apache version has to "be kept back". Try "apt-cache showpkg apache", and check out the dependencies line. > If I try one package: > > # apt-get install apache > Reading Package Lists... Done > Building Dependency Tree... Done > The following extra packages will be installed: > apache-common file libdb4.1 libmagic1 libtool > The following NEW packages will be installed: > libdb4.1 libmagic1 > 4 packages upgraded, 2 newly installed, 0 to remove and 53 not upgraded. > Need to get 2268kB of archives. After unpacking 1954kB will be used. > Do you want to continue? [Y/n] n > Abort. > > Why does it all of the sudden want to install libmagic1 when I don't have > that currently installed at all? Because the package file for the latest version of apache declares that it *requires* libmagic installed in order to work. Apt-get install works effectively like "apt-get dist-upgrade", in that it is happy to install new stuff if the target package needs it. > > Is there a "apt-get update packages just enough so I don't get hacked" > command? :-) It's "apt-get upgrade". For all debian packages, "security fix" upgrades never add dependencies. So they always install fine with "apt-get upgrade". But feature releases can add new dependencies, and if they do then "apt-get upgrade" will report "cannot install; package held back", because adding new packages is not something you want to do automatically on a stable, secure system. This is described in the apt manual. NB: I'm not a debian guru. Any corrections welcome. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Is Linux Unix?
On Mon, 2004-07-26 at 20:01, John Summerfield wrote: > > > >It's not that no one likes the proposal. It's more of the "herding > >cats" phenomenon. There's no consensus on how the release process > >should work, it's hard to make a thousand scattered volunteers listen > >anyway, and no one other than the current release manager is willing to > >step up and actually try to manage the release... > > > > > > I was thinking about that, thinking back to Mr Worthington, my history > teacher in the early 60s. Teaching me history then was a pretty > fruitless task, but I do recall that democracies don't provide efficient > governance. They do in the long term. Name a system of government that has been more efficient than democracy over a 100 year period. The most significant benefit is that plotting to overthrow the rulers is legal, permitted, and expected. And that the retaliation by the rulers is constrained. With every other kind of government, the paranoia of the leader causes repression of the most able people around them, resulting in the suppression of innovation of any sort. This still happens, of course, in democracies, but the balance of power between the different factions prevents it being taken too far. > > That's why we don't have democratic armies. While a democratic army is > planning the next battle, it would get cut to ribbons by the enemy. Very few democracies that regularly use direct vote on issues (referendums). The swiss one is the only one I know of. Instead, representatives are selected. So the parallel would instead be a system where a set of Generals were elected, and they then took the necessary decisions. An army where the Generals were elected would be an interesting one. A little less mechanically effective, I suspect, but maybe also less prone to the macho posturing that has been particularly *ineffective* in recent times. > > Debian is worse than a democracy, it's a democracy of hackers. No > managers at all, to speak of. That's why people do this in their spare time. Hell, if Debian were full of professional managers, I'd be using Slackware :-) > > Hackers are good at cutting code. They don't like documentation - go > look for debian-installer documentation for example. Hackers also by definition work on projects that aren't complete and are evolving. There's a significant problem in trying to document stuff that isn't complete and is evolving. Still, I partially agree with you on this. > They don't like deadlines. "We'll release it when we're good and ready." > Hackers who show management talent are likely to go get lucrative jobs > managing. Managing the hosts of hackers here must be pretty thankless. > > The ideal government, according to Mr Worthington, is a benevolent > dictatorship. A benevolent dictator's word is law, but he rules to the > benefit of those whom he rules. Taking advice is, of course, allowed and > good. So is making decisions. Yes, except that a truly benevolent, tolerant and trusting dictator won't last long. And one who isn't those things won't be a good ruler. > > That, I think, is about how the kernel works. I would say the kernel development is more like a democracy; developers "vote" by contributing to a particular kernel (BSD, Darwin, Linux, Hurd). And so far Linus has proved to be a fine campaigner :-). Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: restrict ssh login to subset of users?
On Mon, 2004-07-26 at 19:55, Eric Cheney wrote: > Hello folks. I am wonderin gif there an easy way to restrict ssh > login access to a subset of users? The most likely answer is to set things up so that only users who belong to a certain group can log in using ssh. Maybe the ssh server supports this directly. If not, then I believe that SSH executes a set of "pam" modules after authenticating a user (see /etc/pam.d/ssh). Pam (pluggable authentication modules) is very flexible, so I suspect if you read up on pam you'll find an option to reject logins for users who are not in a particular group or similar. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Is Linux Unix?
On Mon, 2004-07-26 at 20:58, John Summerfield wrote: > Simon Kitching wrote: > > > >They do in the long term. Name a system of government that has been more > >efficient than democracy over a 100 year period. The most significant > > > > > > The Roman Empire. >From my memory of history, the Roman Empire was fairly close to a democracy until Julius (you got to vote as long as you had sufficient cash). And went downhill rapidly after that. > > In contrast, Debian makes a decision. Whoops. we didn't mean that. Let's > change it! I think the agreement to exempt Sarge from the new requirement for free documentation & supporting files to be a fine example of when backtracking was useful. > > >>That's why we don't have democratic armies. While a democratic army is > >>planning the next battle, it would get cut to ribbons by the enemy. > >> > >> > > > >Very few democracies that regularly use direct vote on issues > >(referendums). The swiss one is the only one I know of. > > > >Instead, representatives are selected. So the parallel would instead be > >a system where a set of Generals were elected, and they then took the > >necessary decisions. > > > > > > They'd be lucky to get _competant_ generals that way. Witness the Roman > Republic. It's not quite how it worked, but similar. > > Or elected _US judges:-) :-) > > > >>Hackers are good at cutting code. They don't like documentation - go > >>look for debian-installer documentation for example. > >> > >> > > > >Hackers also by definition work on projects that aren't complete and are > >evolving. There's a significant problem in trying to document stuff that > >isn't complete and is evolving. > > > >Still, I partially agree with you on this. > > > > > > Umm. First lesson in project managment. Projects have defined > starting-points and defined end-points. > > I once worked on some IBM projects. Documentation was part of the > project, and it can be written when the specs are done. That's not to > say it will be _complete_ then - it has to be adapted as the product is > changes as design errors are found and some of the fine detail evolves. > Intriguingly, I worked for IBM once in my life. Detailed specs, done by Richard Helm (of the "gang of four"). Documentation up the wazoo. Didn't help in the slightest. The main problem was the design was "tossed over the wall" and the superstar designer then went back home, resulting in an inability to adapt the project design to reality later. Another large IBM project failed around the same time. You can find the details here, in the document titled "Ministerial Inquiry into INCIS" (that's *never* a good title for a report!). I would say that the "waterfall" method of development has had its day. But quite how you document an app being created using Agile/XP/etc is indeed an interesting question. http://www.justice.govt.nz/pubs/reports/2000/incis_rpt/ > > > >Yes, except that a truly benevolent, tolerant and trusting dictator > >won't last long. And one who isn't those things won't be a good ruler. > > > > > If te dictator is truly benevolent, & tolerant then there is no need to > overthrow him. How long since an English monarch was overthrown? About the same amount of time since the power of an English monarch was overtaken by the power of parliament I believe. > Instead, over centuries it's evolved into a democracy arguably better > than the American example. > > Which would you prefer? the US model? or what you have? I don't even believe the US model is a democracy. A system where the third candidate in a presidential election is roundly criticised for running at all, because he can split the vote and allow someone else in is not democracy. Nor is one that requires accepting major donations from corporations in order to have a chance of being elected. [flame proof suit on :-] [And sorry for the off-topic post. I'll shut up now, unless really baited!] Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Metacity window manager
On Tue, 2004-07-27 at 07:13, Otto Wyss wrote: > Marc Wilson <[EMAIL PROTECTED]> wrote: > > > On Mon, Jul 26, 2004 at 03:47:32PM +0200, Andrei Badea wrote: > > > IMHO Metacity isn't meant to be a full-fledged window manager (as is for > > > example WM or Fluxbox), but instead it has to be used with the GNOME > > > environment, which provides the "additional components", such as > > > gnome-panel for the taskbar. > > > > Uh, please define "full-fledged window manager", please. A window manager > > manages windows. Anything else is extra. > > > > Not that I'd ever be caught dead defending Metashitty, but it's doing > > exactly what it's supposed to be doing. > > Maybe, but it doesn't do what I want and it gives no info how to change > this. No matter how good Metacity in it's current state is not usful for > me. I believe Ricky/Mark are correct. The metacity package indeed installs a window manager (and just a window manager). The WM/Fluxbox packages are actually simple "desktop environments", because they bundle panels, etc. I suspect the extra bits you want to complement Metacity can be obtained by installing the gnome-panel package. Or, as Ricky says, any other panel should work with Metacity. That would be odd, but doable. I'm not surprised that the Metacity distribution-independent documentation doesn't tell you what *Debian* package to install. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Locked out of system
On Tue, 2004-07-27 at 10:42, Gary L. Roach wrote: > Well I have been known to do stupid things. This is one of them. > > I changed fstab to mount another drive automatically. I messed up the > entry. Now the system boots to management level ((none) prompt) and makes > the hard drive partitions read only. > > There is no text editor available to edit fstab - assuming that I can write > to the file. How do I get out of this mess. > > System Pentium III dual processors > Linux 2.4.20 > Woody 3.0 > 3 1/2 floppy drive > SCSI CDROM (No drivers loaded until system boots) > > Any help will be appreciated This is where it's nice to have a bootable rescue CD around :-). If you have a copy of Knoppix/Gnoppix, you can boot on that, mount the partition(s) on your hard drive, and edit to your heart's content. Of course there are other (simpler) rescue CDs too. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: DNS
On Tue, 2004-07-27 at 13:07, Paul Johnson wrote: > Marvin Gerardo Aguero Salazar <[EMAIL PROTECTED]> writes: > > > How can I specify the DNS server in debian/linux? > > > > Just to clarify, I don't want to set up a DNS Server (which is what > > the DNS howto talks about). I just want to translate names to IPs. > > man resolv.conf If you just want to be able to refer to remote servers by name, you can add entries to /etc/hosts. The command "man hosts" provides more info. Or have I misunderstood? Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: DNS
On Tue, 2004-07-27 at 13:10, Paul Johnson wrote: > Patrick Albuquerque <[EMAIL PROTECTED]> writes: > > > On Mon, Jul 26, 2004 at 03:54:28PM -0600, Marvin Gerardo Aguero Salazar wrote: > >> > >> Just to clarify, I don't want to set up a DNS Server (which is what the > >> DNS howto talks about). I just want to translate names to IPs. > >> > > You can do this in /etc/hosts > > > > IP hostname > > Most people haven't been doing that since the advent of DNS, because > hosts move and updating /etc/hosts every few hours gets kind of old. > 8:o) Sometimes /etc/hosts is the nicest solution. For example, I need to be able to connect to a couple of servers owned by our customers from a couple of specific hosts only. We have VPNs into their network that allow SSH, but obviously accessing their nameservers is not an option. The hosts file works nicely in this case. And unlike DNS, it can be administered by mortals :-) Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Any chance of Debian adding YaST
On Thu, 2004-07-29 at 10:13, Paul Johnson wrote: > Scott Thompson <[EMAIL PROTECTED]> writes: > > > Does anyone know if there are any official or third party plans to develop a > > version for Debian. > > I think you missed April Fools Day by a few months. > > I can't see anything that YaST can do that $EDITOR and discover doesn't > do better. Besides, we've already tried something similar before. Go > search for experiences with Red Hat's system configurator ported to > Debian around the bo and hamm dynasties. Distro-specific tools just > don't work outside that distro. And when they are, their usefulness is > often distroyed, as was the case with Red Hat's rpm. For those using GNOME, there is the gnome-system-tools package, which is proposed for inclusion in the gnome-2.8 release (previously available as a separate install). I haven't tried it, so don't know if it suffers from the same problems as YAST when editing config files. I don't believe it addresses hardware discovery. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Couple Questions Before I install for first time.
On Thu, 2004-07-29 at 13:15, Jason G Skala wrote: > This weekend I am getting ready to switch my server over to Debian > Woody from Red Hat 9.0. And have a few questions before I go and do > this about SMP Kernels. > > I have a Intell LX440GX+ Motherboard with Dual PIII 500's running > software raid currently, I have found some great articles on getting > the Software Raid portion to work with debian so I think I am ok on > that. My real concern is getting the SMP Kernel working with it, and I > have yet to really find any good examples or docs on this. I am not > new to linux but new to debian, I am used to Red Hat were I have a GUI > install and select SMP kernel and that is it. Now is there an easy way > to get a SMP kernel for debian or should I just plan on creating my > own from source? > > The motherboard has a built in Adaptec aic7896/97 Ultra2 SCSI adapter, > is this supported by default without any trouble? > > > Also I have an intel Nic card 82559 that uses the e100 module/driver > and have read that this can be trouble some to get to working any info > that some one can provide to me on that as well. > > Thanks for any help anyone can offer I don't know whether the install provides an option for selecting an smp-enabled kernel, but if it doesn't, you should be able to install one pretty easily. After the install, you can list the set of kernels available with: apt-cache search "kernel-image.*smp" If there is one you like, just: apt-get install kernel-image-xxxyyyzzz The smp-enabled kernels have "-smp" on the end of their names. I presume you'll be wanting a 2.4 kernel (because you're installing Woody). For my Sarge (testing) install, I get: $ apt-cache search "kernel-image-2.4.*smp" kernel-image-2.4.18-1-686-smp - Linux kernel image 2.4.18 on PPro/Celeron/PII/PIII/PIV SMP. kernel-image-2.4-686-smp - Linux kernel image for version 2.4 on PPro/Celeron/PII/PIII/PIV SMP. kernel-image-2.4-k7-smp - Linux kernel image for version 2.4 on AMD K7 SMP. kernel-image-2.4.25-1-686-smp - Linux kernel image for version 2.4.25 on PPro/Celeron/PII/PIII/PIV SMP. kernel-image-2.4.25-1-k7-smp - Linux kernel image for version 2.4.25 on AMD K7 SMP. kernel-image-2.4.26-1-686-smp - Linux kernel image for version 2.4.26 on PPro/Celeron/PII/PIII/PIV SMP. kernel-image-2.4.26-1-k7-smp - Linux kernel image for version 2.4.26 on AMD K7 SMP. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: repeated entries in /var/log/messages...
On Thu, 2004-07-29 at 14:11, Mario Flores wrote: > Hi: > > Since day 1 I installed debian, I see the following in the > /var/log/messages: > > Jul 28 21:01:08 woody -- MARK -- > Jul 28 21:21:08 woody -- MARK -- > Jul 28 21:41:08 woody -- MARK -- > > and they repeat every single hour at the exact same intervals. Does > anyone know what they mean and how can I stop them? They are there so that you can be sure the syslogd daemon is still running. Without these, if syslogd failed you wouldn't be able to tell from the log files. And as syslogd is often part of the security infrastructure of a server, that's not good. I believe syslogd can also log to remote servers, so these mark messages would be extra-useful then. man syslogd for more info, particularly the -m option. syslogd is started by /etc/init.d/sysklogd (note the k). So you can edit this file to pass whatever -m option you want; zero disables these messages. The /var/log/messages file is automatically archived, so I don't see why you would want to disable the mark messages. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: alternatives to NIS and NFS
On Tue, 2004-08-03 at 10:53, John Summerfield wrote: > Paul William wrote: > > > Hi, > > > > I am in charge of a small office network. The server is running Debian > > stable with some testing packages and the desktops are running > > mandrake 10.0. > > > > Currently we are using NIS for authentication and NFS to share the > > home directories. > > > > I have been having some hassles with NIS and would like to upgrade to > > a more modern system. > > > > Are there any alternatives to NFS and NIS? > > > NIS and NFS are differen tissues. What's your problem withNFS? > > For authentication, take a look at LDAP. I mean to. > > > > > As long as it not to complex to setup and is fairly easy to administer > > its fine. X is not on the server so all admin takes place over > > ssh.Security is an issue. > > > > It is safe to assume that there will not be any windows clients on the > > network, ever :) There is one osx ibook being used but it does not > > need to 'login' to the network. > > > OSX should be able to authenticate against LDAP. Watch your IUDs tho. > Authentication in Linux is (usually) done via the PAM library, ie applications wanting to do authentication link against the PAM library and then call PAM apis to do authentication. In particular, openssh, login, etc. all do this. Here's some info on PAM: http://www.kernel.org/pub/linux/libs/pam/Linux-PAM-html/pam.html There are PAM modules for ldap, which enable PAM to authenticate against LDAP servers. Here's one link I found on this by googling: http://www.metaconsultancy.com/whitepapers/ldap-linux.htm Also, try google for "linux pam". Note that I haven't used PAM and LDAP myself (yet). Like John, it's on my to-do list. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Restricting a shell user to his home dir ?
On Thu, 2004-08-05 at 18:16, [EMAIL PROTECTED] wrote: > Hi, > > I am learning a lot from this mailing list :). I have few shell users, i want to > restrict their shell login to their home directories. > > Like they should not be able to move around in the system and see other user's home > directories. > > Any suggestions would be usefull. Hi, When you installed debian, you should have been asked "should home dirs be readable by others". If you answered no, then when new users are created, their home dir will automatically have permissions of rwx--, which means that no other user on the system can see their files. I believe this is even the default setting. [NB: perms of rwxr-x--- are also secure, provided each user has their own personal group setting]. If you answered "yes, home dirs should be readable by others" then when you added users, their home dir will be rwxr-xr-x, which allows others to see their files. This is often the default for "traditional" unix systems, like AIX. In this case, users (or you) can use chmod to make the home dirs unreadable. But generally, users *do* need access to the rest of the system. They need access to /bin, /usr/bin, etc or they won't be able to run even simple tasks like "cp", "more", etc. And unix is designed to allow users to access system dirs without any harm occurring. Note that users will be able to see that dirs /home/fred, /home/sue etc exist but won't be able to see anything in them if the permissions on those dirs are set appropriately. If you're really paranoid (and this can sometimes be healthy) you can look into something called "chroot", which means that when users log in, they get a special restricted view of the local filesystem. But this is a major pain to set up and administer. Only for the brave... Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How popular is Debian (popularity contest)
On Sun, 2004-08-08 at 22:19, Jaap Haitsma wrote: > I came across this website http://popcon.debian.org/ > It shows statistics how many people have installed a certain package on > what kind of system etc. etc. If you want to anonymously report what > packages you are using you have to have the popularity-contest package > installed. > I guess many people using Debian have it installed. That's why I was > amazed that the number of submissions is only around 6000. > Are there only around 1000 debian users on the world (assumption 60% of > them sends reports) > > Also 9 architectures have 20 or less submissions. 3 architectures have > only 1 submission. Seems to me supporting all these architectures for > all the developers is quite some burden just to help out a few users. > > Is there something wrong with my reasoning??? I think a better way to measure the number of debian installs would be for security.debian.org to count unique IP addresses. While lots of people won't have popularity-contest installed, a large majority of them will be getting security updates... Of course this would not count users of testing or unstable, which don't have security updates. And it won't properly count people using apt-proxy, etc. or behind NAT firewalls. But it would be a start. >From what I remember of using the new debian-installer a few months ago, it really encourages people to use popularity-contest, so maybe the number of installations of popularity-contest will grow. Of course its purpose is mainly to determine which packages should go on which CD images of Debian, so as long as the set of people with popularity-contest installed is reasonably representative, it doesn't matter if the numbers are not large. As someone said, debian servers are probably under-represented, as good sysadmins won't install unnecessary software on production machines. I hope the people determining the CD package order take this into account... Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help with Woody Install
On Mon, 2004-08-09 at 19:38, Ben Alex wrote: > Hi everyone > > I'm trying to install Woody (my first Debian installation) Hi Ben, I can't offer any advice on your SCSI/RAID issues. But are you aware that Debian is pretty close now to releasing a new version that will make Woody obsolete? The current "testing" (aka beta) debian has a vastly improved installer that might just solve all your problems. So unless you intend to use this install in production within the next 2 months, you may well be better off trying to install the current "testing" release, called "Sarge". See: http://www.debian.org/devel/debian-installer/ Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Upgraded to unstable - lost network connectivity
Hi,
I recently added unstable to my sources.list, and did a dist-upgrade.
After some mucking around, I now have a working system again - except
for networking.
I'm on a LAN with DHCP etc. Normally, interface eth0 comes up
automatically. However now when I boot I get no network connection, and:
su -
ifup eth0
just gives me:
pcsimon:~# ifup eth0
Internet Software Consortium DHCP Client 2.0pl5
Copyright 1995, 1996, 1997, 1998, 1999 The Internet Software Consortium.
All rights reserved.
Please contribute if you find this software useful.
For info, please visit http://www.isc.org/dhcp-contrib.html
eth0: ERROR while getting interface flags: No such device
Bind socket to interface: No such device
exiting.
Failed to bring up eth0.
Fortunately I backed up the previous system onto a separate partition,
and am able to boot into that to get this working system.
I have tried:
diff -r /etc/network {badsystem}/etc/network
and this shows that everything in the /etc/network directory of the
"bad" system is identical to the scripts in the "good" system, except
the "ifstate" file which is as expected, ie they weren't touched by the
update:
pcsimon:~# diff -r /etc/network /mnt/hda6/etc/network
diff -r /etc/network/ifstate /mnt/hda6/etc/network/ifstate
1,2d0
< lo=lo
< eth0=eth0
Does anyone have any ideas how I can get network connectivity back
again? Where might I start diagnosing this problem?
One thing I would like to do is try an apt-get dist-upgrade again, in
case there are some updated packages to fix this. But of course that is
difficult without network connectivity. Is there a way I can upgrade the
"bad" OS install which is on a separate partition?
Thanks,
Simon
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Upgrading XFree86
On Sat, 2004-06-05 at 14:40, Bradley Alexander wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> I did a dist-upgrade on one of my sid boxes yesterday, and I can't get
> xserver-xfree86 to install. Because of that, x-window-system-core and
> x-window-system cannot install. When I run apt-get dist-upgrade, apt-get -f
> install or manually dpkg -i, I get the following (this one is from an apt-get
> - -f install):
>
> After unpacking 0B of additional disk space will be used.
> Setting up xserver-xfree86 (4.3.0.dfsg.1-1) ...
> dpkg: error processing xserver-xfree86 (--configure):
> subprocess post-installation script returned error exit status 127
> dpkg: dependency problems prevent configuration of x-window-system-core:
> x-window-system-core depends on xserver-xfree86; however:
> Package xserver-xfree86 is not configured yet.
> dpkg: error processing x-window-system-core (--configure):
> dependency problems - leaving unconfigured
> dpkg: dependency problems prevent configuration of x-window-system:
> x-window-system depends on x-window-system-core; however:
> Package x-window-system-core is not configured yet.
> dpkg: error processing x-window-system (--configure):
> dependency problems - leaving unconfigured
> Errors were encountered while processing:
> xserver-xfree86
> x-window-system-core
> x-window-system
> E: Sub-process /usr/bin/dpkg returned an error code (1)
>
> Can someone point me in the right directon on getting this installed? I
> checked bugs.d.o, but did not see a similar bug.
Just FYI, I got exactly the same problem ("script returned exit status
127" when running dpkg --configure for the xserver-xfree86 package). And
either X or GDM then refused to start (I'm not sure which it was).
I previously had a debian-testing setup, and added "unstable" then did
dist-upgrade [mainly so I can play with gnome-2.6].
Eventually I marked the xserver-xfree86 package as "HOLD" and did "dpkg
--configure -a", which seemed to at least get graphics working again.
However I'm fairly new to debian, so I really don't know what I'm doing
yet.
Cheers,
Simon
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Upgraded to unstable - lost network connectivity
On Sat, 2004-06-05 at 15:42, dircha wrote: > Simon Kitching wrote: > > I recently added unstable to my sources.list, and did a dist-upgrade. > > After some mucking around, I now have a working system again - except > > for networking. > > > Does anyone have any ideas how I can get network connectivity back > > again? Where might I start diagnosing this problem? > > You sound like someone who has probably thought of this as the possible > cause of the problem, but did you upgrade your kernel as well? > Nope. > Or even if you have not upgraded your kernel, but if the kernel driver > for your nic was compiled as a module, have you checked whether the > module is being loaded (lsmod), and if so, is it? Bingo! Running "lsmod" in the working vs non-working config shows that a whole bunch of drivers are no longer being loaded after the dist-upgrade. Running modprobe to force the network drivers to be loaded restores network connectivity. Thanks, dircha! So the question now is: why did dist-upgrade from testing to unstable mess around with the list of modules that are loaded at boot time? That's sort of a rhetorical question; I don't hugely care as I now have a working system [at least I can manually force the necessary drivers to be loaded on boot]. But presumably other people will be bitten by this too... Presumably the list of modules to load on boot is just a config file floating around somewhere like in /etc. Or is it dynamically determined during booting [in which case the dynamic detection has been broken]? Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Upgraded to unstable - lost network connectivity
On Sat, 2004-06-05 at 18:24, Paul Scott wrote: > Simon Kitching wrote: > > >On Sat, 2004-06-05 at 15:42, dircha wrote: > > > > > >>Simon Kitching wrote: > >> > >> > >> > >>You sound like someone who has probably thought of this as the possible > >>cause of the problem, but did you upgrade your kernel as well? > >> > >Nope. > > > > > But did the dist-upgrade upgrade the kernel? If you had previously > installed a kernel-image it might have been upgraded. Sorry, I should have been more explicit. Here's the output of "uname -a" on the "bad" version (linefeed in the middle inserted by me): Linux pcsimon 2.4.26-1-686 #1 Sat May 1 18:04:05 EST 2004 i686 GNU/Linux and here's the output I see after booting into the backed-up (pre-upgrade) version of debian. It's identical: Linux pcsimon 2.4.26-1-686 #1 Sat May 1 18:04:05 EST 2004 i686 GNU/Linux I can also confirm that command diff -r /boot /mnt/hda6/boot reports no differences except the tweaks I manually applied to get the OS booting on the backup [/mnt/hda6 points to the partition containing the "bad" upgraded version]. I think that between them, these tests (/boot contents and uname output unchanged) demonstrate that the kernel of the good and bad (upgraded) systems are identical, right? Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Upgraded to unstable - lost network connectivity
On Sun, 2004-06-06 at 14:22, dircha wrote: > > Bingo! Running "lsmod" in the working vs non-working config shows > > that a whole bunch of drivers are no longer being loaded after the > > dist-upgrade. Running modprobe to force the network drivers to be > > loaded restores network connectivity. > > > > So the question now is: why did dist-upgrade from testing to unstable > > mess around with the list of modules that are loaded at boot time? > > That's sort of a rhetorical question; I don't hugely care as I now > > have a working system [at least I can manually force the necessary > > drivers to be loaded on boot]. But presumably other people will be > > bitten by this too... Just for the record: I did a dist-upgrade today and rebooted, and all the appropriate modules are now loading automatically. Whatever the problem was, it seems to have been fixed in the latest unstable. Dircha: thanks for the additional info on modutils. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Possible convert to Debian
On Tue, 2004-06-08 at 15:10, John Fleming wrote: > Consider me a newbie, but with enough experience to be dangerous. I started > learning Linux with RH8.0 about a year ago. I later installed Fedora Core 1 > and have upgraded to FC2, but I'm not happy with the upgrade process (they > recommend a fresh install of each upgrade, and of course, Fedora has a > relatively frequent planned upgrade cycle). For my intended uses, I don't > care about support for the latest whiz-bang USB devices, cameras, > Centrino/ProWireless etc - I just want easy to install and maintain basic > server functions with decent security. With my FC2 server (Dell 600SC), I > have these things going: pop and imap mail, Squirrelmail, Mailman, > SpamAssassin, Webmin, and Apache with several virtual hosts. I'm fairly new to Debian myself (but with experience with commercial unix flavours and other linux distros) so probably know where you're coming from. > > I've toyed with Debian and Knoppix a bit, but haven't tried to fully > implement the above services yet and have a few questions. > > 1. It sounds like Sid is actually pretty stable, I'm guessing especially > for the basic mail and webserving things I use. Would you recommend I go > with Sid, or with testing or stable? Personally I would recommend the "testing" distribution. Sid/unstable really can be unstable at times. I upgraded last week and lost all network connectivity for a while. Others have reported losing all graphics, etc. Unless you're fairly happy poking around in config files from a console the unstable distribution can be nasty. I ran testing for several months, though, and had only the occasional minor inconvenience in return for access to software *reasonably close* to the latest releases. Stable is really old at the moment - though hopefully a new release will be out within a few months. It's really more appropriate to servers than desktop systems. > > 2. It seems that installing from the Knoppix CD might be a bit easier for a > newbie? I've never tried something like woody 3.0r2 on new technology - > only on an older limited computer, and I had problems with NIC being > recognized and others. I have installed Sid using the Knoppix CD on my Dell > laptop after first partitioning with Partition Magic and have a dual boot > going with Sid and that other OS. I have done some preliminary > experimenting using apt-get to install Squirrelmail and Webmin, tinkering > with the sources list and so forth, so I'm a little familiar with how that > works. Lots of work has gone into a new debian installer over the last year or so. It is at "beta 3" or similar at the moment. I used it to install my current system and found it pretty good; it's not as friendly as RedHat's Anaconda installer but from what you describe of your background I think you would be comfortable with it. BTW, I'm not criticising the new debian-installer by comparing to anaconda; there are lots of things it can do that anaconda cannot. A lot of work has been done on auto-detection of hardware & peripherals, so you might be pleasantly surprised. > 3. If I start with the Knoppix CD, I really do end up with the same Debian > Sid distro as if I started with the unstable CDs themselves? Not quite, according to the reports I've seen. Yes, you can point a Knoppix installation at standard debian repositories to get updates. But apparently Knoppix sets up some of the system scripts etc. differently from debian so there can be surprises (unexpected breakage) later on. NB: all this is just hearsay. > > 4. Correct me if I'm wrong, but as I understand it, if I use Sid and keep it > updated, I should never have to do a complete reinstall of the OS as > recommended for Fedora upgrades. And if I use the stable version, I can do > an apt-get dist upgrade and upgrade to the new version(s) as available > without a complete reinstall, right? I believe this is the case no matter whether you run stable, testing or unstable. Running dist-upgrade just pulls down the appropriate files. I've not been using Debian long enough to go through a major release upgrade yet, but I've heard only good things about the process. > > 5. I am interested in software RAID 1 and have 2 identical HDDs. Is there > an option during the install from Debian CDs (didn't see it in the Knoppix > HD install) to setup RAID? If not, any recs on the easiest way to get RAID > 1 going after the initial installation? No idea. > > Any other comments for this relative newbie that's old (>50) and not a > sysadmin by trade? Thanks a bunch for your time! I have used RedHat8 and 9, Mandrake9 and 9.1, and debian. My early experiences of debian never got past the amazingly complex installation process. However using the new debian installer was quite reasonable and I am now sold on debian. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Good News Server
On Thu, 2004-06-10 at 18:29, s. keeling wrote: > Incoming from Loren M. Lang: > > Does anyone know of a good news server I could use? I'd like to post A "good news" server sounds like a fine idea - none of this depressing war, famine and SCO stuff :-) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Basic, but what the hell am I doing wrong here?
On Fri, 2004-06-11 at 14:49, Kevin Mark wrote:
> On Thu, Jun 10, 2004 at 09:36:04PM -0400, Stephen Touset wrote:
> > I've opened and read files hundreds of times in my life. What gives now?
> >
> > --
> > Stephen Touset <[EMAIL PROTECTED]>
>
> > #include
> >
> > using namespace std;
> >
> > int main(void)
> > {
> >
> > ifstream fin("test.cc");
> > char* str;
> >
> > if (!fin.is_open())
> > {
> > exit(1);
> > }
> >
> > while (!fin.eof())
> > {
> > fin.getline(str, 80);
> > }
> >
> > fin.close();
> >
> > return 0;
> >
> > }
> Hi Stephen,
> I can understand your confusion. you have a perception issue that is
> clouding your understanding. either that or you have been looking at
> this for umteen hours and too much caffine x-). The issue is the
> understading of memory allocation versus pointer definition.
>
> 'char * str' defines a 'char *' called str. it allocates memory for this
> pointer. so, str has an address of 0xabcd (for example). It has not been
> initialized with anything, so it contains undeterminate data (eg.
> garbage). lets say it the contents of address 0xabcd is 'Z'. so, *str =
> 'Z'. What does *(str + 1) contain? undeterminate data. As does *(str +
> 2) ... *(str + 79). So you define str, don't initialize it and then read
> data into *(str) ... *(str + 79). Well not really, becase you have no
> right to write to those areas. Why? because you have not allocate memory
> at those addresses. How do you do this? well in c you'd 'malloc. But
> this is c++, so, you'd use 'new', I think.
Yep. or:
char str[80];
And in c++, there probably is a way to read directly into a string or
stringstream object or similar, which is much safer (and easier) than
calling malloc or new.
(damn; it's been about 5 years since I last seriously used c++ and it's
all forgotten...)
Regards,
Simon
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How Do I Update to Gnome 2.6?
On Tue, 2004-06-15 at 17:12, Benjamin Sher wrote: > Dear Ed: > > I don't use Debian. I use the Debian-based Xandros Linux. But I would advise > you very strongly to read some of the reviews of Gnome 2.6. They are > horrendous. Just look for them at LinuxToday. The consensus is that if you > really want Gnome 2.6, you should wait till your distro has upgraded to it. > Otherwise, you'll be facing one hell of a nightmare both installing and > running it. The above is not true at all. I'm running Gnome 2.6 right now. It's very nice, and perfectly stable. Add the following to the file /etc/apt/sources.list: deb ftp://ftp.nz.debian.org/debian/ unstable main contrib [Note: instead of the download URL above, select a suitable mirror for your location] then as the root user type: apt-get update apt-get install gnome-desktop-environment Note that pointing apt at "unstable" can have nasty side-effects. You really should run a system backup before doing this, just in case. However I can say that I updated about 1 hour ago and everything is fine. I typically comment out the "unstable" line above when I'm happy with my system, so that I don't download any more "unstable" packages. There is probably a better way to manage this, such as "pinning" etc. Some day I hope to have time to learn the finer details of apt. Note; I am running x86 not PPC. I don't expect that to make any difference, though. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: How Do I Update to Gnome 2.6?
On Tue, 2004-06-15 at 17:54, Benjamin Sher wrote: > Dear friends: > > Here is an example of one of the reviews of Gnome 2.6 from LinuxPlanet: > > Gnome 2.6: Two Left Feet? > http://www.linuxplanet.com/linuxplanet/reviews/5428/1/ All this review points out is how much Slackware sucks. Or how badly the slackware install package for gnome-2.6 sucks. Or that the reviewer is too dumb to be trusted with a keyboard. Choose one or more of the preceding options. The Debian people who built the gnome-2.6 package for Debian appear to have done a much better job - it just installs. Well done - and thanks. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Pros/Cons Kde vs Gnome?
On Tue, 2004-06-15 at 18:14, CaT wrote: > On Mon, Jun 14, 2004 at 02:50:43PM -0700, Steve Lamb wrote: > > s. keeling wrote: > > > I gave up on both of those; they're equally uncontrollable, and far > > > too fat to leave any room for actual applications to run. ymmv. > > > > Could've fooled me. > > > > KDE + Squid + Addzapper + other stuff... > > > > [EMAIL PROTECTED]:~} free > > total used free sharedbuffers cached > > Mem:775556 767612 7944 0 131368 392300 > > -/+ buffers/cache: 243944 531612 > > Swap: 655344 26600 628744 > > > > 531Mb's not enough? Hmph. > > It's more of a case of 'Isn't 240Mb (or 200 cos of squid) a bit much for > a pretty desktop?' ;) It's always real hard to measure actual memory usage of an app. This 240MB is presumably actually the memory taken by the kernel plus disk cache + all sorts of other stuff too, like SSH servers. But assuming all 240MB are used by the desktop, thats what- US$50? I'm willing to pay that for the chance to run a pretty desktop for the lifetime of that PC. And I live in a country where the US$ is about twice that value in real terms. Of course some people live places where that *is* an unacceptable amount of money. So it's good that they have options, like XFce or others [see the RULE project for details on running a truly "light" linux install...]. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Gnome AddressBook?
On Wed, 2004-06-16 at 16:44, Ed Sutherland wrote: > What is the Gnome application serving as an addressbook? I've tried to > apt-get gnome-pim and it no longer exists. I need something like KDE's > addressbook. Thanks. > AFAIK, in the next version of gnome (2.8), the module "evolution-data-server" will be included to provide this functionality. Currently I don't believe gnome provides a centralised addressbook, ie KDE is ahead of gnome in this area. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: programming editor
On Fri, 2004-06-18 at 11:18, Steve Lamb wrote: > Anders Karlsson wrote: > > For vim, you can always install Gvim to give you more of a GUI version. > > Or kvim for KDE... Or even vim-part so that it can be embedded in KDE > apps that support the editor part. :D > > > Best editor I have ever used though was CygnusED on the Amiga. 65kB in size > > and everything you'd expect an editor to do and then some (still have not > > found an editor that will do a rectangular area cut/paste in the middle of > > a text). Second to that is FrexxEd (Amiga, OS/2) that had a very nifty > > virtual disk you could drag/drop files in/out of that opened/saved files > > in the editor itself. But I digress... > > You mean like vim? > > Do this: > 5i123454kl2j2lxp > > Or, long form. 5 rows of "12345", go to 2nd row on the 2, hit CNTL-V, > down 2, over 2. You'll now mark the middle block of 9 digits. x deletes it, > p pastes it back in after the 5 so now the middle 3 rows read "15234". Ahh vim :-) The jedit editor does rectangular cut&paste. It is in fact an *extremely* good general-purpose text editor. And has the "filesystem pane" you were looking for, among many other "plugins". Because it's a java app, though, I suspect it takes at least as much memory as Kate. And it has the clunky swing look-and-feel (though with java1.5 coming out, I hope the gnome l&f for swing has been much improved). It's also nearly as configurable as emacs; almost every part of jedit can be accessed via BeanShell (a java scripting language). There is a debian source for jedit, or you can download & run the gui installer. See www.jedit.org for more info. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: moving to the 2.6 kernel?
On Mon, 2004-06-21 at 16:28, Jules Dubois wrote: > On Sun, 20 Jun 2004 10:47:10 -0500, Kent West wrote: > > > Just install a Debian box like you normally do, then "apt-get install > > kernel-image-2.6.whateverfitsyourarchitecture". > > Thanks for the advice, Kent. A foolish question: There's really nothing > more to it than installing a kernel-image package (and, in my case, > updating initrd-tools)? I had no problems. Once the install had completed, I could boot into either the 2.4 or 2.6 kernel. Not that the 2.4 option gets much use ;-). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: The boot sector and it's backup
On Sat, 2004-06-19 at 03:07, Keith O'Connell wrote: > Hi, > > I just installed sarge from scratch onto a machine here and I got a > message during the subsequent startuo sequence that was several screens > full of 8 digit numbers followed by the message; > > "there are differences between the boot sector and it's backup" > > It went on to load happily, but I don't get this on other machines and > the thing was installed only half an hour prior to this. > > What would cause this, and what is the correct tool to deal with it? > Better would be an indication of what I should be reading to learn about > this. > > Can anybody advise? I'm not an expert in this field, but it sounds like this message is coming from the bootloader - lilo or grub. Is this message appearing before any kernel boot messages? I'm guessing that when lilo/grub installs itself it makes a copy somewhere safe for some reason, and that either the boot sector or the "save" sector of your hard-drive has been corrupted somehow. The bootloader is checking that all is well on boot, and has (correctly) determined that either the boot sector or its saved copy has changed. >From the fact that bootup works, I suspect it is the backup that is corrupted. Reinstalling your bootloader might fix this; try "man lilo" or "man grub" respectively. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: USB memory stick?
On Mon, 2004-06-21 at 08:43, jakob bratkovic wrote:
> stan wrote:
> > I was given a USB memory stick, as a promotional giveawau by a vrndor,
> > Friday.
> >
> > How can I use this with my Debian laptop?
> >
>
> If I understand correctly you're talking about an USB drive. If this is
> the case, Linux will see it as an SCSI disk and probably assign it to
> /dev/sda1 if you otherwise use IDE drives. Adding a line such as:
>
> /dev/sda1 /mnt/usbvfatnoauto,user 0 0
>
> and creating a directory:
>
> /mnt/usb
>
> should allow you to mount it with a command:
>
> mount /mnt/usb
I found it necessary to do this when using kernel-2.4:
modprobe usb-storage
mount /proc/bus/usb
where /etc/fstab has:
none/proc/bus/usb usbfs defaults 0 0
This can be a script linked from rc{n}.d. Or maybe added to
/etc/modprobe.conf? Not sure; I just run these commands manually before
I use USB - and of course it's only needed after a reboot.
Then as Jakob says, the entry in the /etc/fstab file described above
allows any user account to type:
mount /mnt/usb
... do stuff
umount /mnt/usb
You could even make desktop entries to do the mount/umount :-)
Cheers,
Simon
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: USB memory stick?
On Mon, 2004-06-21 at 11:59, Jon Dowland wrote: > On Sun, 20 Jun 2004 15:32:54 -0400, stan <[EMAIL PROTECTED]> wrote: > > > > I was given a USB memory stick, as a promotional giveawau by a vrndor, > > Friday. > > > > How can I use this with my Debian laptop? > > You need to use something called hotplug and ensure that you have > various SCSI modules built for your kernel - sd_mod.o is necessary. > some very out-of-date details at http://jon.dowland.name/unix/twinmos/ I'm pretty sure this is not necessary, at least for basic functionality, with recent systems. Maybe someone read Jon's notes and built it in ;-). I'm using a 256MByte USB memory stick fine, with a debian testing system running kernel-2.4, and also with a debian testing system running kernel-2.6. Yes, the USB drive appears as a SCSI device, but that all seems to work out-of-the-box for me. Now I *would* like to see an icon pop up on the desktop when the USB device gets inserted. Maybe the hotplug stuff is related to that? I can live without it, though... Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: USB memory stick?
On Tue, 2004-06-22 at 02:03, Jon Dowland wrote: > On Mon, 21 Jun 2004 20:22:17 +1200, Simon Kitching > <[EMAIL PROTECTED]> wrote: > > > I'm pretty sure this is not necessary, at least for basic functionality, > > with recent systems. Maybe someone read Jon's notes and built it in ;-). > > > > I'm using a 256MByte USB memory stick fine, with a debian testing system > > running kernel-2.4, and also with a debian testing system running > > kernel-2.6. Yes, the USB drive appears as a SCSI device, but that all > > seems to work out-of-the-box for me. > > Are you using a pre-built debian kernel or a custom one? According to > ftp://ftp.debian.org/debian/dists/woody/main/disks-i386/3.0.23-2002-05-21/bf2.4/kernel-config; > the bf24 stock kernel has the relevant SCSI modules built-in. This may > be the case for > all the debian stock kernels. I'm using the prebuilt debian kernel-image packages (2.4 and 2.6) from the standard debian mirrors. Just did "apt-get install kernel-image-xxx". And yes all the necessary SCSI stuff is just there. Running modprobe -l | grep scsi reports heaps of stuff. > > > Now I *would* like to see an icon pop up on the desktop when the USB > > device gets inserted. Maybe the hotplug stuff is related to that? I can > > live without it, though... > > Me too - or similar such things. Someone said in another thread that > `udev' works > towards these goals, but I haven't investigated it. I had a memory of seeing an article about "project utopia", maybe on lwn.net, so I did some googling and found it: Today, Project Utopia is an umbrella project including multiple other open source components such as udev, hotplug, HAL, the kernel, D-BUS, and GNOME. It's is managed using use cases, as opposed to functional and technical specifications See: http://conferences.oreillynet.com/cs/os2004/view/e_sess/5195 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: sudo lock-up
On Tue, 2004-06-22 at 13:56, Tom Allison wrote: > "mynamehere" is not allowed to run sudo on localhost. This incident > will be reported > > > I used the wrong password too many times. > How do I "clear" this? > This is new to me; I am not aware of any feature of either "sudo" or "pam" that allows an account to be locked out after N login failures (though a pam module could probably be built for this fairly easily). Are you absolutely certain that the lockout is due to repeated login failures, and not some other cause like accidentally deleting some sudo-related file? I believe that when sudo wishes to validate a password, it just uses the standard Linux-PAM module. So you may also wish to check the contents of /etc/pam.d/sudo to see if there are any funky authentication rules in there. All I have (which I believe is normal) is: authrequiredpam_unix.so ie sudo authentication simply checks /etc/passwd and /etc/shadow, with no facility to perform lockout at all. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Hard Drive Limitations
On Tue, 2004-06-22 at 12:45, Greg Madden wrote: > On Monday 21 June 2004 03:21 pm, Ryan Nielsen wrote: > > I would like to use a couple of 250GB Western Digitals to backup > > data. I set up a Debian machine (I have 4 others for other purposes), > > to use as a backup server. I am having difficulties with the drives. > > I tried two Maxtors previous and have swapped them out thinking it > > was a bad set of drives. I get various different errors copying > > files to these hard drives. I have tried different partitioning > > schemes, rsync instead of cp, and a few other methods, but I still > > get some form of error when I get into the process of copying. > > > > My only success has been with Knoppix. I used Knoppix and was able > > to copy and rsync just fine to these disks. I would like to use > > Debian, and not have to install another version of Debian or Linux. > > > > What I am getting at is, is there any problem with having a big > > partition or HD on a Debian system, or anything else someone might > > know? Any help would be greatly appreciated. > > This just sounds weird to me. Managing disks and filesystems is done in the kernel; there is very little a *distribution* can do to screw up access at this level. Maybe it is something to do with the mount options in /etc/fstab? Though even that seems unlikely to me... When you say "some form of error", what exactly do you get? eg "bad blocks", "invalid superblock", ??? What kernel are you running with your debian distro (what does "uname -a" report)? Did you compile it yourself? Apply any custom patches? How did you partition the drives? How did you create the filesystems on these drives? And what type of filesystem are you using: ext2/ext3/reiserfs? When you say "Knoppix works just fine", did you try this *on the same machine*, or did you move the disks to a different server to do this test? Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: moving to the 2.6 kernel?
On Tue, 2004-06-22 at 18:01, Jules Dubois wrote: > On Mon, 21 Jun 2004 01:39:16 -0500, Kent West wrote: > > > Jules Dubois wrote: > > > >>There's really nothing more to it than installing a kernel-image package > >>(and, in my case, updating initrd-tools)? > >> > > If you're not already running an initrd kernel, you may have to add one > > line to /etc/lilo.conf, but you're warned about it during the install. > > I installed kernel-image-2.6.6-2-k7. I get lots of error during boot; > they go by quickly, I don't seem to be able to stop them, and they're not > recorded in any log file. If I interpret the situation correctly, every > attempt to load a module fails with a "/lib/modules: read-only filesystem" > error. Have you tried using "dmesg"? See "man dmesg" for more info. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Problem with CVS and group ownership.
On Wed, 2004-06-23 at 12:53, Vincent Hallberg wrote: > Package: cvs > > Version: 1:1.12.9-1 > > > > In order to share files between normal users on CVS I’ve added them to > the ‘src’ group using ‘gpasswd –a someuser src’. Next I’m using > Tortoisecvs to login as someuser and upload to the repository I’ve > created. It all goes up ok and this has been working without any > problem for the last 4 days between 5 different users but now all of a > sudden all ‘commits’ preformed by any user are changing the group > ownership away from ‘src’ to the user’s group name. Once that happens > other users can no longer overwrite or checkout this data. > (‘someuser’ is of course to be replaced by the login name of a real > user on the server.) > > > > Obviously I don’t understand a great deal about all the configuration > options available in CVS but does any body have any idea what might > have changed or how I can get this back to my original configuration? > > > > I did do an ‘apt-get update’ and ‘upgrade’ over the weekend which > might have something to do with what’s happened here but I guess a > good question might be, “How do other people set this sort of system > up for this kind of use.”? Maybe the problem is related to the default "umask" for users? Anyway, I find using CVS via shared files very nasty to maintain. I recommend you set up a CVS pserver (either as daemon or via inetd), and force everyone to use the :pserver: protocol to checkin/checkout. That way, all the files in the CVS repository are then owned by the "cvs" user, and all file permission problems go away forever. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Problem with CVS and group ownership.
On Wed, 2004-06-23 at 15:21, Vincent Hallberg wrote:
> The default umask on a fresh Debian install is:
> UMASK 022
>
> I've been thinking of adding users to the system this way to solve the
> problem:
> adduser -no-create-home --ingroup src | username
>
> I am running the pserver which starts up automatically but am not
> familier with how to add user's to the CVS that are not already normal
> user's on the system. Tortoisecvs's default connect method is
> pserver.
>
> Does this make any sense or should I be working this monster an
> entirely different way?
Hi Vincent,
Just a note; the convention in most email lists is to add your comments
mixed with or *below* the comments that are being responded to. This is
called "bottom posting", rather than adding comments at the top of the
email ("top posting"). The "bottom posting" style means an email makes
sense when read top-to-bottom...important in any situation, I think, but
critical when people may be picking up a "conversation" part-way
through.
And please reply to the list rather than direct. The email archives are
a very useful resource for other people to search, but that doesn't work
if all emails other than the first exchange are off-list. And if I screw
up and give bad advice, there won't be anyone watching to correct me :-)
When using "pserver", CVS keeps its list of valid users & passwords in
the file CVSROOT/passwd within the CVS repository. To add a CVS "user",
you just edit this file. There is no connection between "cvs user ids"
and users on the local system, nor is there any connection between a
user's CVS password (if any) and their OS login password.
For a cvs repository that is "internal" to a company, I usually just set
everyone's CVS password to "password" or similar; the aim of CVS user
accounts in this case is simply to provide distinct login accounts so
that changes are marked against users, not to apply security. But you
can have "real" passwords if you want. Just don't make them the same as
the OS login passwords, because when using pserver, CVS stores passwords
in a .cvspass file in each user's home dir, in plain text.
If you *do* care about strong authentication of your users, you should
probably not go the "pserver" route. But in that case I can't offer any
advice other than to allow plenty of time to get your access permissions
right! Or use SubVersion (subversion.tigris.org), which is far better
than CVS (though somewhat more resource-hungry).
Regards,
Simon
> -Original Message-
> From: Simon Kitching <[EMAIL PROTECTED]>
> To: Vincent Hallberg <[EMAIL PROTECTED]>
> Cc: [EMAIL PROTECTED]
> Date: Wed, 23 Jun 2004 13:40:06 +1200
> Subject: Re: Problem with CVS and group ownership.
>
> > On Wed, 2004-06-23 at 12:53, Vincent Hallberg wrote:
> > > Package: cvs
> > >
> > > Version: 1:1.12.9-1
> > >
> > >
> > >
> > > In order to share files between normal users on CVS Ive added them
> > to
> > > the src group using gpasswd a someuser src. Next Im using
> > > Tortoisecvs to login as someuser and upload to the repository Ive
> > > created. It all goes up ok and this has been working without any
> > > problem for the last 4 days between 5 different users but now all
> of
> > a
> > > sudden all commits preformed by any user are changing the group
> > > ownership away from src to the users group name. Once that
> > happens
> > > other users can no longer overwrite or checkout this data.
> > > (someuser is of course to be replaced by the login name of a real
> > > user on the server.)
> > >
> > >
> > >
> > > Obviously I dont understand a great deal about all the
> configuration
> > > options available in CVS but does any body have any idea what
> might
> > > have changed or how I can get this back to my original
> configuration?
> > >
> > >
> > >
> > > I did do an apt-get update and upgrade over the weekend which
> > > might have something to do with whats happened here but I guess a
> > > good question might be, How do other people set this sort of
> system
> > > up for this kind of use.?
> >
> > Maybe the problem is related to the default "umask" for users?
> >
> > Anyway, I find using CVS via shared files very nasty to maintain.
> >
> > I recommend you set up a CVS pserver (either as daemon or via
> inetd),
> > and force everyone to use the :pserver: protocol to
> checkin/checkout.
> > That way, all the files in the CVS repository are then owned by the
> > "cvs" user, and all file permission problems go away forever.
> >
> > Regards,
> >
> > Simon
> >
>
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: moving to the 2.6 kernel?
On Wed, 2004-06-23 at 18:07, John Summerfield wrote: > Aaron Robertson-Hodder wrote: > > >Hi, > > > >Just wondering if all you people are moving to 2.6 from Woody, Sarge or > >sid? I have been trying to compile the 2.6.6 kernel on my Woody distro > >and keep getting depmod errors. > > > >I've tried using a number of different configs as a starting point > >(including the knoppix one) and still can't seem to get it right. > > > >I have now read that there may be issues with Woody going to 2.6. Is > >that true? If so, I guess I'll try Sarge if that sounds like the best > >option... > > > > > > > > > There's certainly more to going to 2.6 than just building the packages. > There _is_ a 2.6 kernel and everything else you need around, I found it > and downloaded it a while ago. > > Support (security fixes) may be a problem with them, but that applies to > Sarge too. > > Me, I went to Sarge. Sarge has 2.6 kernels up to 2.6.6. > If I remember correctly, Kernel 2.6 has a whole new module-loading system that requires updated modutils stuff. This could explain your depmod problems. I believe the new tools are backwards-compatible, ie the new versions work with 2.4 kernels too. But from sarge it is *so* easy to do: apt-get install kernel-image-2.6.6-1-686 and have it all just happen. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Lots of problems - help please
On Thu, 2004-06-24 at 10:27, Daniel Klein wrote: > Hi, > > I'm having a whole damn lot of problems here, pretty much all of a sudden. > > Ever since a good friend helped me install debian (let's say he's a > little on the hectic side.. I didn't see everything he did), there was a > problem with getting X to like my mouse (IntelliMouse Explorer).. I had > to type startx twice to get X and KDE up and running, but then things > were fine. Now everything's come crumbling down. > > First of all, I had the alt-tab bug for some time (see my other mails in > this list). That behaviour has stopped now. Alt-tab stopped working > altogether, I get the taskbox and NOTHING ELSE. I can close it with ESC. > Great. Next, my mouse has become sluggish like hell. We're talking maybe > 3 or 4 fps when I move the cursor around. > > Also, I fail to remember what dpgk-reconfigure I have to run to do > anything to my X Server. It is absolutely foggy to me, why X runs KDE > (I've looked around the X config files and found no mention of KDE at > all). I wanted to try fvwm, just to see if that'd run bearably fast > (we're talking Athlon TB 1200 / 512 MB RAM / very fast 200 gigs hd / > Geforce 3). I have absolutely no clue how to do that in debian. > > I love apt-get for the easiness of installing things, but do any of you > have any idea what a headache it is to find out what a certain package > is called? If there was a simple 'list everything that's installed' (and > I'm sure there is, I just can't find it) command, that would help > already. There's xf86config, xf86cfg, dpkg-reconfigure > whateverIhavetotype here and the settings in KDE. This is a damn > headache, and I'd be infinitely grateful if someone could help me with this. > > The alt-tab behaviour is totally unbearable right now. I want to solve > this somehow, but I have no idea what to do. I know this is the linux > mailinglist joker, but I'm seriously considering going back to Windows > for good, and that'd be after 3 years of using Linux and trying my best > to manage it. I had such high hopes for debian as well, it seemed perfect :/ > > Thanks for any help you can offer. Hi Daniel, Maybe you should just make a clean install of debian on a different partition of your machine. That will leave you with a working system with the least messing around with debian internals - and it's what Windows users typically do every couple of years to clean up cruft. Yes, I'm sure that it's possible to repair your existing system, but that could be more difficult.. You don't say what Debian version you are using; if you are using "sarge", aka "testing", then there is a very nice new installer that makes the process much easier than it used to be. And the new installer's hardware-detection is very good. To find what packages you have installed using the command-line, you can use: dpkg --list See "man dpkg" for more info. Or you could use the GUI package manager "synaptic". There is then a nice drop-down box that allows you to say "show me all installed packages". Though maybe with your mouse problems a GUI app is not so easy to use at the moment... NB: Can anyone point me to a reference that describes what the flags on the left of the "dpkg --list" output (eg "ii", "rc") mean? I know "ii" means "installed".. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Another "testing" vs "unstable" question
On Thu, 2004-06-24 at 12:58, John Summerfield wrote: > > > >>fwiw I was much amused when I first tried Knoppix (it was, I think, a > >>3.2 beta but it might have been 3.1). The hardware detection is done > >>with Red Hat's tools. > >> > >> > > > >Why be amused? If RedHat licenses their stuff such that other systems > >can use it, and it works, it would be foolish to reinvent the wheel. > >That's one of the reasons people like open source so much. > > > > > > I know. It was the thought of using RH tools to configure what is > essentially a Debian system. Given how well RH detects hardware, I think > it a great shame that the Debian project didn't use some of the RH tools > to help installing Woody. > > I can (and used to) install RHL 7.3 on arbitrary local-computershop > hardware in fifteen minutes, fully automated. > > I gather the name Ian Murdock has some significance here, and that he's > connected to Progeny. Here's what Progeny says, "Red Hat's® Anaconda is > the standard installer among Linux distributions. Our port of Anaconda > to Debian brings the familiar installation experience of Anaconda to the > rest of the Linux world." > > See http://platform.progeny.com/anaconda/ And those comments also point out that Progeny's Anaconda port is x86-specific. Debian supports a much wider range of hardware than Red Hat does. So Anaconda for debian (+RH hardware discovery) is nice for people with x86 hardware, but *everyone* can use the new debian-installer and its hardware-discovery framework (which also happens to be largely developed by Progeny: see http://platform.progeny.com/discover/index.html). Personally, I think support for alternative hardware is really important; x86 chips aren't bad, but opening the door for other chip makers and designers to compete using alternative architectures is good for everyone. Imagine how much slower our CPUs would be if Intel didn't have AMD and Transmeta pushing them...and imagine how much harder they will have to work when PowerPC and others can compete fairly (ie when the dominant computer OS is not one that is limited to x86 architecture only). The debian-installed FAQ has more info (esp. item 4): http://wiki.debian.net/?DebianInstallerFAQ Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Debian installation
On Fri, 2004-06-25 at 08:30, Gabriel Ebner wrote: > On Thu, Jun 24, 2004 at 03:23:35PM -0500, Kent West wrote: > > Last I knew, the sid installer was very very broken. Most folks use the > > Sarge installer (which needs testers anyway) to do the base install, and > > then upgrade from there to Sid. > > Interesting. I've always used the sid installer and never experienced > _any_ problems (except for mkreiserfs blocking while waiting for input). > It even set up GRUB instead of LILO. I used the sarge installer last night (beta-4) to set up a new system on a pretty old PC. It worked beautifully - and installed GRUB. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
OT: Re: Internet Explorer 6 on Debian Unstable
On Thu, 2004-08-12 at 13:12, John Summerfield wrote: > As for the advisability of running Internet Exploder at all, see the > recent advice at slate.msn.net. Slate is _owned_ by Microsoft, and for > those who don't wish to check, the advice is "use firefox." Actually, I think this is something Microsoft management should get some respect for. Journalists being allowed to actually express their own opinions rather than the opinions of the owner of the media in question is all too rare these days. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: componentized linux and "sarge"
On Thu, 2004-08-12 at 17:46, Ritesh Raj Sarraf wrote: > Hello all, > was just reading about componentized linux, Progeny etc. > Found the idea of comoponentization excellent. I just got > amamzed/confused with the following statemnt from Progenty. > > "Progeny Debian 2.0, A componentized linux distribution is based > and closely related to Debian Sarge". > > So, I'm happy with sarge. Is sarge componentized ?? As I understand it, Progeny are building some "meta-packages" (aka "components") which reference carefully selected groups of standard debian packages. The idea (I think) is that you choose the appropriate Progeny meta-package and get a set of normal debian packages installed which Progeny has selected for quality and tested as a group. I imagine they'll offer things like a "mailserver" metapackage, a "firewall" metapackage, etc. Anyway, that's how I understand it. The description of "componentized linux" on the progeny site is rather abstract and difficult to understand. An interview I saw with Ian Murdock wasn't much easier to understand either :-) I do quite like the concept (assuming I've understood it right). You can get the same effect with normal Debian, but only if you know things like which individual packages you need to achieve your task, which of them is the best quality, which have the best security patch response period, and which ones integrate well together. Not that easy to figure out sometimes.. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Deleted root account in passwd
On Sun, 2004-08-15 at 06:14, Pete Clarke wrote: > Hi there, > > I managed to delete the root entry in /etc/passwd whilst playing > on a test box > Is there an easy way of re-inserting, or copying the backup passwd file I > have without physical access to the box? > > I can ssh into it, but only as a regular user - obviously I can't su to root > as it doesn't exist anymore :-) I presume it's still possible to boot into "single user mode" with a broken /etc/passwd. Here are instructions I found by googling for "linux boot single user mode": http://www.redhat.com/docs/manuals/linux/RHL-7.3-Manual/ custom-guide/s1-rescuemode-booting-single.html It should then be possible to edit the /etc/passwd file. Alternatively, boot using a rescue disk, mount the filesystem containing the /etc/passwd file, then edit it to reinsert a root entry. I think the Debian install CD will function as a "rescue disk" for this purpose. Or Knoppix/Gnoppix/etc will do the job. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Deleted root account in passwd
On Sun, 2004-08-15 at 11:55, Simon Kitching wrote: > On Sun, 2004-08-15 at 06:14, Pete Clarke wrote: > > Hi there, > > > > I managed to delete the root entry in /etc/passwd whilst playing > > on a test box > > Is there an easy way of re-inserting, or copying the backup passwd file I > > have without physical access to the box? > > > > I can ssh into it, but only as a regular user - obviously I can't su to root > > as it doesn't exist anymore :-) > > I presume it's still possible to boot into "single user mode" with a > broken /etc/passwd. Here are instructions I found by googling for "linux > boot single user mode": > http://www.redhat.com/docs/manuals/linux/RHL-7.3-Manual/ > custom-guide/s1-rescuemode-booting-single.html > It should then be possible to edit the /etc/passwd file. > > Alternatively, boot using a rescue disk, mount the filesystem containing > the /etc/passwd file, then edit it to reinsert a root entry. I think the > Debian install CD will function as a "rescue disk" for this purpose. Or > Knoppix/Gnoppix/etc will do the job. Sorry, I didn't read the original email properly, and missed the bit about not having physical access to the machine. Are the security patches on the server up-to-date? If not, you might be able to download code for one of the local-root-exploit holes that have been found within the last 6 months :-) Otherwise, I think you're stuffed. Rather like I was when I misconfigured iptables on a remote server :-( Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Websphere MQ on Debian ??
On Mon, 2004-08-16 at 01:38, John Foster wrote: > On Saturday 14 August 2004 07:41 pm, Mark D. Hansen wrote: > > Has anyone here ever installed WebSphere MQ (MQ Series) on Debian Linux? > > Any special tricks or pitfalls? Thanks. > - Hi, Yes I have installed MQ (v5.3.0) on Debian Sarge, and it works fine. I have been using it just about every day for a year or so, without problems. There are a few minor hassles getting it installed, but not too bad. I've got a short document around somewhere with some instructions I wrote on how to install MQ on debian, but can't find it just at the moment. I'll have another look for it later. But basically the steps, as I remember, are: * create user accounts * install using "alien" Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: On downloading Debian ISOs
On Mon, 2004-08-16 at 20:59, Rakhesh Sasidharan wrote: > Hi! > > Have been wanting to try out Debian for so long, but never managed to > get my hands on it. Finally I land up at this client site where we have > a good net access, and so am thinking of downloading the ISOs to give > them a shot on my home machine. Hi Rakesh, Just as a side-note, I hope you didn't download the "Woody" (Debian 3.0) files. The new release is due out *very* soon, and is a major improvement over the previous version. Unless you are are intending to set up a production server, you will probably be best off downloading the "testing" version, ie what will become the stable version shortly. In particular, the installation process is vastly improved. The current "testing" version is called "Sarge", and will probably be released as Debian 3.1, though I think it deserves to be Debian 4.0. > Unfortunately the proxy and firewalls here do not allow me to use jigdo As noted earlier, Jigdo just uses HTTP, so it should work for you. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Quick question: Choice of release
On Tue, 2004-08-17 at 10:33, Anders Breindahl wrote: > I and a friend are playing with the idea of convincing our gymnasium to > using Debian GNU/Linux as their main server OS. > > As it comes to this issue, which release we would want; stable, testing or > unstable? > > Making the choice, it would matter that the focus is on reliability. We > would want it to be absolutely reliable and secure. No security flaws would > be accepted. On the other hand, we wouldn't want it to be outdated, so any > performance-fixes are not included. What is the best compromise? Hi, I would recommend waiting until the next Debian release. This is currently scheduled to happen on Sept 19th, though deadlines are rather flexible in Debian. You can then go with that "stable" release. If you want to get started now, you can install "Sarge". Sarge is currently in testing. Sarge will become "stable" at the september release date. It's currently pretty good, perfectly fine for doing the setup work. And after the release date, a simple "apt-get dist-upgrade" command will ensure everything gets upgraded to the official release versions. See "http://www.debian.org/devel/debian-installer"; for Sarge installation info. You *could* install the current stable version, then after the next release upgrade to the new stable version. But I wouldn't recommend it, because the upgrade will want to install *huge* amounts of updates. And in addition, the new installer is much easier to use than the old one. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Publishing program?
On Tue, 2004-08-17 at 11:59, stan wrote: > I remebet using a publishing layout application in Debian a while > ack, had a cutesey sort of name. But I can't rememner what it was. > > Can anyone jog my memory? Scribus (http://web2.altmuehlnet.de/fschmid/) appears to be a publishing layout program. I have also heard people say that the gnome "gLabels" program is actually very powerful and useful for many more layout tasks than just labels. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Serial terminal in testing?
On Tue, 2004-08-17 at 13:34, Michael D. Crawford wrote: > Is there a serial terminal program in testing? Like what you'd use to > log in to another host over a serial line? > > Not something like xterm or gnome-terminal which is for login shells > over a network login. I need something that communicates over serial > ports, and will let me set the baud rate and flow control. > > I'm developing the firmware for an embedded device that sends debugging > messages out a serial port, and into the serial port on my PC. When I > was using Windows I used Hyperterminal for this. But I just got set up > to develop the firmware under linux instead. Browsing through dselect I > couldn't find anything in the way of serial terminals for debian. > > I'd be shocked if there wasn't something so I must be overlooking it. The command apt-cache search serial returns a number of likely-looking candidates. Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: X resolution gnome/wmaker
On Tue, 2004-08-17 at 14:40, Lance Hoffmeyer wrote: > Hello, > > I just installed Sarge from Netinst (very cool) and have > X11 running. The problem is that the resolution is > incredibly high (1600x1200) and I cannot see the icons > or fonts very well. I am using Wmaker and Gnome as the > desktop. I have tried to reconfigure using both gnome-conf > and wmakerconf. Nothing seems to happen. > > Using Ctl-Alt+ allows me to change resolution but then the > icons don't fit the desktop. > > Any help on how I can change resolution to 1280x968 and > increase font/icon size? Well, for gnome-2.6, you just double click on the "start here" icon on the desktop, select "desktop preferences" then "screen resolution". There are also options for setting the font size in there. Right-clicking on the gnome-panel allows the icon size to be tweaked. Under the hood, the screen-resolution control is calling the X "R-and-R" (Resize and Rotate) extension. There is an "xrandr" commandline tool to do this too, though you don't get the nice confirmation dialog, and I don't know if this tool makes the change permanent. I don't know if the "xrandr" tool saves the changed resolution for next time. I hope this helps, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: urgetn help date/time settings on debian
On Tue, 2004-08-17 at 18:51, Vijaya S wrote: > Hi all, > The mail server 's date and time has gone one day ahead and it has > debian unstable installed on it. > it shows Wed Aug 18 10:25:59 IST 2004 > but the actual time and date now is Tue Aug 17 12:21:04 IST 2004 > how do i get it right > man date -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: alternative to cdrecord?
On Tue, 2004-08-24 at 10:36, Brian Pack wrote: > Haven't you heard? IIRC the kernel 2.6.8 plugged a security hole that > cdrecord used to function. Once the hole was closed, users could no > longer run cdrecord as they could in previous kernels. With the old > kernel, a user could potentially wipe a drives firmware. > The emphasis is on the word "users". There are some scsi commands that can toast a drive; the kernel now has a blacklist of commands that non-root users are forbidden from sending. Unfortunately, cdrecord uses some of these commands, and hence cannot record when run as a non-root user. Root users can still send any scsi command they like to a drive. Unfortunately one comment I saw indicates that making cdrecord suid won't help, as cdrecord deliberately drops back to the real user id before burning the cd, for "security". I guess kernel hackers are working on a solution.. Subscribers to Linux Weekly News can find more info here now: http://lwn.net/Articles/97552/ Non-subscribers have to wait until thursday to access this article. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: CPU temparature display??
On Tue, 2004-08-24 at 12:50, Ishwar Rattan wrote: > Is there a utility (X11 or text based) that will display > the cpu temperature? I think you want the "lmsensors" module. See: http://www2.lm-sensors.nu/~lm78/ Regards, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Java on Debian
On Wed, 2004-08-25 at 12:24, Michael Satterwhite wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > Having just installed Sarge on my laptop, I tried downloading and installing > the Sun J2SE (from Sun - I didn't see it - or expect to see it! - in apt). > Unfortunately, it wouldn't install because of a missing library. > > I *KNOW* there are many people writing Java on Debian. What tool do you use. > If Sun (what I'd prefer) what's the magic to getting it to install? Missing library? I've never had that problem, and I've installed java a bunch of times. I just download the linux ".bin" bundle from Sun and run it to unpack it. Then ensure that the location I've unpacked it into is in my $PATH. Or links to the most important binaries could be created manually from /usr/local/bin etc. Because I need to have multiple different java versions installed concurrently (I've got 3 sun versions and 1 ibm version installed now) I prefer to install manually rather than from an apt repository. Which version were you trying to install, and what error message did you get? Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Real Time monitoring/alerting utility..
On Thu, 2004-08-26 at 10:24, Tim Kelley wrote: > On Thu, Aug 26, 2004 at 08:14:50AM +1000, Michael Bellears wrote: > > > No - He wants to be notified immediately if an FTP or SSH connection is > > established. > > Using snort and tailing the logfile, it doesn't get much more real > time than that. Just modify the config files to treat all accesses as > alerts. Use acidlab with it and you have a history of every access, > ever. > Another option might be to use a PAM module. I don't know if there already exists a suitable pam module, but if not then writing one shouldn't be too hard. Then just add it to /etc/pam.d/ssh. Cheers, Simon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
RE: Websphere MQ on Debian ??
Hi Mark, I'm running a 2.4 kernel. I think that MQ doesn't like the new threading model in 2.6 (MQ also has problems on later RedHat 2.4 kernels, which backport the new threads). Setting environment variable LD_ASSUME_KERNEL should resolve the issue. Google should tell you what to set it to... You'll also need to install fixpack CSD06 or later... And sorry, I looked for that "install mq under debian" doc but couldn't find it. Regards, Simon On Fri, 2004-09-17 at 02:25, Mark D. Hansen wrote: > Which kernel are you running 2.4 or 2.6 ? I'm running 2.6 and have been having > problems getting MQ v5.3 to run. Apparently, kernel 2.6.x isn't officially > supported. > > I you find that document, and could pass it along, I'd appreciate it very much. > > Thanks, > > Mark > > -Original Message- > From: Simon Kitching [mailto:[EMAIL PROTECTED] > Sent: Monday, August 16, 2004 12:56 AM > To: [EMAIL PROTECTED] > Subject: Re: Websphere MQ on Debian ?? > > > On Mon, 2004-08-16 at 01:38, John Foster wrote: > > On Saturday 14 August 2004 07:41 pm, Mark D. Hansen wrote: > > > Has anyone here ever installed WebSphere MQ (MQ Series) on Debian Linux? > > > Any special tricks or pitfalls? Thanks. > > - > > Hi, > > Yes I have installed MQ (v5.3.0) on Debian Sarge, and it works fine. I > have been using it just about every day for a year or so, without > problems. > > There are a few minor hassles getting it installed, but not too bad. > I've got a short document around somewhere with some instructions I > wrote on how to install MQ on debian, but can't find it just at the > moment. I'll have another look for it later. > > But basically the steps, as I remember, are: > * create user accounts > * install using "alien" > > Regards, > > Simon > > > > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
