Problem with HP XW9300
I'm mainly sending this in the hopes it saves someone else the frustration I had to go through. If you're getting messages like: PCI: Multiple domains not supported or PCI-DMA: high address but no IOMMU try adding acpi=off to your kernel's boot parameter. I EVENTUALLY found this in a HP document for installing Red Hat. It fixed all of my booting issues. I have no idea what the consequences of turning ACPI off are, but if I can boot, and see both my processors, then I'm happy. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Night mare to set day light savings time
Jigga Man [EMAIL PROTECTED] writes: Sorry to directly send you an email but i dont know how to reply to the mailing list. when i click the reply to link it gives a dialog box saying mailto protocl not registered. Anyways Thank you very much for your time to help me with my problem I tried the su -c tzconfig and when thru the the proces s..it only asked me 2 things first select a country in which i selected option no 3 which US timezones and then in option 2 i selected Eastern and here is the output Your default time zone is set to 'US/Eastern'. Local time is now: Thu Nov 13 11:44:48 EST 2003. Universal Time is now: Thu Nov 13 16:44:48 UTC 2003. this time is still one hour off ..its says local time is 11:44 when its actully 10:44 here ... i didnt find any specific option for daylight savings time... did i do anything worng ? I'm a bit confused. Are you in a location which is still on daylight savings time (DST)? Most of the US switched back to standard time a couple of weeks ago and if you're on the East coast and your clock is set correctly then the time zone is correct. EST = UTC - 5, which is what the info you give above says. If your clock still thought you were in DST your time information would be EDT = UTC - 4. Perhaps the issue isn't the timezone, but the fact that your clock is set incorrectly? You can reset your clock, as root, using the command: date 11131317 for example, to set the time/date to Nov. 13, 1:17pm (see man date). This is all apart from the hardware/BIOS clock. The hardware clock only matters at boot time and shutdown time. Once you get the time set correctly in Linux you can synch your hardware clock using the command: hwclock --systohc Depending on how you have your hardware clock set up you'll want to add either --localtime or --utc to the command line. You can find out which of these Linux thinks your hardware clock should be set at by reading /etc/default/rcS and looking at the variable UTC. If it's set to no then run: hwclock --systohc --localtime if it's set to yes then run hwclock --systohc --utc If your hardware clock is set to local time then the issue may be that your hardware/BIOS clock didn't reset when we came off DST? Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt-get unstable question
Rick Weinbender [EMAIL PROTECTED] writes: I'm running the current stable version and I need to install a package from an unstable source (tmda ver.0.86). (the unstable package does work with woody). * Is there a way to install an unstable package on stable distro from the command line. * Someone suggested 'apt-get -t unstable install tmda'. I tried this but it didn't work. Does anyone know if this is possible and the proper syntax. Do you have a /etc/apt/sources.list file that contains pointers to the unstable distro? Here's a excerpt from mine: deb http://ftp.debian.org/debian woody main contrib non-free deb http://ftp.debian.org/debian sarge main contrib non-free deb http://ftp.debian.org/debian sid main contrib non-free I list all 3 releases, by name (woody = stable, sarge = testing, sid = unstable). Once you have that you'll want to tell apt the main release you want. Do that by editing /etc/apt/apt.conf and putting a line like: APT::Default-Release stable; in it. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: cvs over ssh with non standard port
Micha Feigin [EMAIL PROTECTED] writes: I am trying to get cvs to access the repository through a ssh connection when the sshd is listening on a non standard port. I tried using cvs -s CVS_RSH=ssh -p port -d :ext:cvs:/var/lib/cvs co package but cvs insisted on trying port 22. Is it possible to do this? I haven't tried it, but perhaps you can twiddle the config file, ~/.ssh/config. In there try a block like: Host=cvs Port=1028 And see what happens. This is on the client. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Crash in fork.c on everything = install hosed
Malcolm Box [EMAIL PROTECTED] writes: Paul Scott wrote: Malcolm Box wrote: snip symptoms Near as I can tell, this is an assertion somewhere in the bowels of libc - a nasty place for things to go wrong. The error seems to be affecting anything trying to do certain type of fork() operations. Among others, it has taken out apt-get and dpkg :-( I just posted the same thing two days ago. Thanks to Jacob Lell. Bug 218546. It's fixed in the latest libc6. OK, that looks like it might fix things. But can anyone tell me how to get the new version onto my machine given that both dpkg apt-get don't work at the moment? Is there some way to manually unpack the files from the .deb into the right places? I've never done it, but there's nothing magical about *.deb files. They're simple ar archive files, usually containing a data.tar.gz and control.tar.gz file, among other things. For your purposes you could try and untar the data.tar.gz file, which would contain the actually libc binaries and go from there. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Weird DNS problem
I just spent a couple of hours diagnosing a DNS lookup problem, but I have no clue what's causing the problem. If I do a host junker.whatever.gov it's trying to do a lookup over the loopback interface. This times out eventually. If I do a host junker.whatever.gov 134.xxx.xxx.xxx where 134.xxx.xxx.xxx is our local DNS server everything works fine. The request goes out over the eth0 interface and immediately returns a valid IP for the given name. Does anyone know why it's trying to send the request over the loopback interface if I don't specify the IP address of a DNS server? /etc/resolv.conf looks fine, it has a domain whatever.gov line and then our two primary DNS servers. Looking at strace output of host junker.whatever.gov it seems to be reading /etc/resolv.conf Every other network transaction works fine as long as I specify an IP address, including routing correctly. I'm stumped!! TIA, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Weird DNS problem
Gary Hennigan [EMAIL PROTECTED] writes: I just spent a couple of hours diagnosing a DNS lookup problem, but I have no clue what's causing the problem. If I do a host junker.whatever.gov it's trying to do a lookup over the loopback interface. This times out eventually. If I do a host junker.whatever.gov 134.xxx.xxx.xxx where 134.xxx.xxx.xxx is our local DNS server everything works fine. The request goes out over the eth0 interface and immediately returns a valid IP for the given name. Does anyone know why it's trying to send the request over the loopback interface if I don't specify the IP address of a DNS server? /etc/resolv.conf looks fine, it has a domain whatever.gov line and then our two primary DNS servers. Looking at strace output of host junker.whatever.gov it seems to be reading /etc/resolv.conf Every other network transaction works fine as long as I specify an IP address, including routing correctly. I'm stumped!! Well, it never fails to amaze me how I can stare at something for an hour and not see the obvious. In /etc/resolv.conf nameserver was misspelled nameservr. I looked at that file dozens of times and only when I did a diff of the /etc directory with that of a working system did the problem slap me in the face. Man I'm a dumbass sometimes! Sorry for the waste of space on the mailing list. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Memory
Eicke [EMAIL PROTECTED] writes: I have a machine with 6Gb of memory. I installed the last version of Debian and linux kernel 2.4.21. I am trying to run the cap3 (ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), not stripped), but the following error occurs: Ran out of memory: -874931512 bytes requested I monitored and when the software reach 1.9 Gb the error occours. Then I test another application all_align.pl (perl script text executable), and when the script reach 3.0 Gb the following error occours: out of memory In the first case I think is a software problem but in the second I guess there is something in Operating System or Kernel configuration. I ran as root and as common user in both cases. In general you're not going to be able to allocate more than 2GB of RAM on a 32-bit system like the Pentium. While Intel played some tricks with the hardware and actually implemented a 36-bit address bus (I think it's 36 bits anyway), applications generally use 32-bit pointers on a 32-bit CPU and they're assumed to be signed so that limits you to 2^31 bytes of memory, or 2048MB (2GB). What the 6GB of RAM buys you is that you could run 3 separate processes each using 2GB of RAM and never hit your swap space, but a single application can't use more than 2GB at a time, in general. There may be low-level things in the kernel that would allow you to use more than 2GB of RAM, but I'm not familiar with them and it certainly wouldn't be portable. If you need an application to have access to more than 2GB of RAM then you need to get a 64-bit system like an Alpha, Sparc or Itanium. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Memory
Eicke [EMAIL PROTECTED] writes: Hi Gary I appreciate your response. My machine is a Dual Xeon 2.8 GHz 64-bit. Should be happens the errors anyway? Clear up some things for us Eicke. First, I may be behind on my Intel marketing, but aren't the Xeon line of processors just good ole Pentium 4's (32-bit) with extra cache? Is Intel marketing the Itanium (64-bit) under the Xeon name? Second, did these applications ever work under Linux on your system? If indeed Intel is selling 64-bit Xeons(TM) using Itanium processors then I'll have to pass on answering your questions as I have no experience with Linux on the Itanium processor. My no-experience guess would be that you'd need to compile the kernel with Itanium support, which I don't see as an option in my 2.4.21 kernel configuration, and install 64-bit Itanium versions of the applications you're trying to use. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: pdf readers
Greg Folkert [EMAIL PROTECTED] writes: On Tue, 2003-08-12 at 13:17, Vincent Lefevre wrote: On Tue, Aug 12, 2003 at 13:10:21 -0400, MJM wrote: If anyone else has ghostview vs xpdf vs acrobat vs ??? stories or comments I'd like to read them. I use xpdf on my PPC laptop since gv renders some PDF documents very badly and acrobat doesn't exist here. Begging to differ with you: duke:~# apt-cache policy acroread acroread: Installed: 5.07-woody0.0 Candidate: 5.07-woody0.0 Version Table: *** 5.07-woody0.0 0 500 http://marillat.free.fr testing/main Packages 500 http://marillat.free.fr unstable/main Packages 100 /var/lib/dpkg/status Not that it is as current as I'd like (v6) but it'll have to do. All the other renderers crap out on most things some things that are simple. Acroread doesn't play nice with FontConfig... so it is a bit choppy at times. He says right there that has a PowerPC laptop, not an i386-based system. Unless I'm mistaken you've provided a site that has i386 Debian software, NOT PPC. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: finding the cause of a lock-up
techlists [EMAIL PROTECTED] writes: I have a slight problem with xmms, and freeamp. Both, on the start-up of playing the first song lock-up for approxametly 5 minutes...then start playing fine. after the initial lock-up the programs run with no problems what so ever. I've run the programs from the command line, but neither program seems to have a verbose mode. What can I do to find out what is causing these programs to freeze like they are? Use strace (Debian package strace). Just run them like: strace xmms take a peek at the man page if you need more info (man strace). Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: mail gathering
Kirk Strauser [EMAIL PROTECTED] writes: At 2003-08-04T22:54:34Z, vinz [EMAIL PROTECTED] writes: fetchmail only allows me to download mail to root but i want to put the mail into my pop3 account how do i go about doing this and what program can i use ? Run fetchmail as you? If I understand the Powermail stuff he may not have an account on his server. Vinz, take a look at the --smtpname option of fetchmail. You can also do the same in a fetchmail configuration file. Look at the man page for fetchmail and the CONFIGURATION EXAMPLES where they specify separate user IDs for the remote system (pop3 server) and the local machine. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Linux firewall vs Windows and Hardware based firewalls
Andre Volmensky [EMAIL PROTECTED] writes: I have to put forward an argument to management regarding setting up a firewall on some of our clients networks. What are the advantages of a linux firewall over something like Windows with WinRoute on it, or even a hardware based firewall. What are the disadvantages etc. I know I am asking on a linux users mailing list, but I would also like reply's not to be too bias. Just to add to what others have said, I run an software firewall (OpenBSD 3.2 soon to be 3.3) and one of the things I like about it is being able to log intrusion attempts and contribute those to DShield.org (http://www.dshield.org). Gives me a sense that I'm contributing to the overall security of the net. Probably a false, since a lot of the fightback reports just get ignored by ISPs, but nonetheless... Now I believe some hardware firewalls allow you to capture logs via something like a remote syslog capability, but then you'd have to have a server to process them anyway. With my OS/software-based firewall I can have it do everything necessary to send the logs off to DShield. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: why I don't want CCs
martin f krafft [EMAIL PROTECTED] writes: [snip] top - 19:32:46 up 104 days, 4:57, 1 user, load average: 0.01, 0.02, 0.05 Tasks: 299 total, 1 running, 295 sleeping, 3 stopped, 0 zombie Cpu(s): 0.2% user, 2.7% system, 0.0% nice, 97.1% idle Mem: 2068748k total, 2043068k used,25680k free, 471920k buffers Swap: 498004k total, 4176k used, 493828k free, 1160540k cached Do tell why your top shows two CPU lines and mine only one -- this is also an SMP system, and it works fine, utilising both CPUs... Man, I *REALLY* wanted to avoid this thread! ;) But a legitimate question deserves an answer... Hit 1 while in top and it'll display the CPU info seperately. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: why I don't want CCs
Nathan E Norman [EMAIL PROTECTED] writes: On Thu, Jul 17, 2003 at 12:26:36PM -0600, Gary Hennigan wrote: [snip] Man, I *REALLY* wanted to avoid this thread! ;) But a legitimate question deserves an answer... Hit 1 while in top and it'll display the CPU info seperately. Er, what version of procps? Doesn't work here; I've got 2.0.7-8 (and a non-i36 arch but I hope that doesn't matter). I'm on a i386 arch and using procps 3.1.9 so one of those is likely the culprit. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: keystroke ctrl+s freezes terminal/console
Joerg Johannes [EMAIL PROTECTED] writes: On Tuesday 15 July 2003 13:51, Rogier Wolff wrote: On Tue, Jul 15, 2003 at 07:33:47AM -0400, Shawn Lamson wrote: Hey all - I have noticed that pressing ctrl+s in console or xterm suspends input... I have not figured out how to successfully get out of that situation. Is there a graceful way? What is the purpose of ctrl+s ( i believe I have the emacs style command line editor )? How can I disable ctrl+s or alter it so that it does not freeze my console/xterm? Note that control-S doesn't suspend input. It suspends OUTPUT! control-Q lets the output go again. As Roger already asked: Is there a way to disable this Ctrl-S shortcut? Ctrl+S/Ctrl+Q is actually the XON/XOFF protocol. If you've been around long enough you'll be familiar with that. It was used a lot for slow terminals and for slow modems. You can disable it in most terminal types using stty. For example: stty -ixon will disable Ctrl+S. See man stty for more details on XON/XOFF. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: keystroke ctrl+s freezes terminal/console
Joerg Johannes [EMAIL PROTECTED] writes: On Tuesday 15 July 2003 17:08, Gary Hennigan wrote: As Roger already asked: Is there a way to disable this Ctrl-S shortcut? Ctrl+S/Ctrl+Q is actually the XON/XOFF protocol. If you've been around long enough you'll be familiar with that. It was used a lot for slow terminals and for slow modems. You can disable it in most terminal types using stty. For example: stty -ixon will disable Ctrl+S. See man stty for more details on XON/XOFF. OK, thanks. This supend is now turned off. Now, Ctrl-S wnats to start an incremental search like in emacs. That's your shell, bash. By default it uses Emacs-style keybindings. Look at the Bash man page under READLINE to alter this behavior. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT? www.donotcall.gov problem
Paul E Condon [EMAIL PROTECTED] writes: Yesterday, I attempted to use the new federal (USA) web site for registering my telephone numbers in the 'do not call' database. I did not succeed, and I am wondering what is wrong. Let me describe my experience: The web site displays a form on which I am requested to fill in the telephone numbers that I want added to the list, AND a space where I am to enter my email address. I followed the instructions. I entered three telephone #s, and one of my four email addresses. In a few seconds, three emails arrived in my inbox. Each of them, requested that I click on a hotlink to visit a web site that would complete the registration process. I did this for each email. In all three cases there was an almost immediate response that the data was incorrect, and no indication as to how I might make it correct. I am using Galeon web browser and mutt MUA. Does gov require Windoze software on my computer? Has anyone succeeded in registering with a Debian machine? I saw the same thing when I didn't paste the link in the email correctly. I use emacs+gnus for email and when I selected the link in the email, via X's selection mechanism, it included the \ character, which is used in emacs to indicate a line wrap. After I removed that character from the link I had pasted into Galeon I was able to register all three of the phone #'s I requested. So, perhaps you didn't paste the link correctly. I did succeed in registering by using the telephone and calling 1-888-382-1222. When I did this, the hardware of the telephone system was able to verify that I was calling from the telephone instrument that I was registering. But I have a question about this process: I was required to enter the telephone # manually, and then my manual entry was checked against what the hardware said. Why? Why not just have the hardware grab the telephone # and put it into the database? Perhaps you can register numbers other than the one you're calling from, with some other type of verifaction? Dunno. You'd really have to ask them. And what relation does my email address have to my telephone #? What are they checking in my email that has any relation to my telephone? (Yes, I use ppp to my ISP, but surely they can't query the ISP server to find out which telephone I'm using to call the ISP, and surely it should be possible to register if ones internet connection is entirely independent of ones telephone service. ???) It's probably just another traceability issue. I'm sure the tele-marketing industry was able to at least convince the FTC to have some means to verify that you actually signed up for the DNC list and just didn't sign up every number in your local phone book for the list. What's to stop someone from writing a little script to query every number on record, via something like switchboard.com, and entering all the numbers into the database? The tele-marketers probably wanted some way to ensure that this didn't happen and so the software probably has a check that disallows more than a certain number of phone numbers from being registered via the same email address. Again, you'd have to ask the government if you want anything more than just guesses. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[OT]: Linux poll at Novell
I don't normally pay attention to things like online linux distribution polls. They're hardly scientific and so mean little. But I've seen some folks crowing about how Gentoo is smashing the competition because of the results of this poll and so I thought I'd bring the existence of the poll to the attention of Debian folks here. The poll is here: http://www.novell.com/linux/ Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: HOWTO layout formatting
Nori Heikkinen [EMAIL PROTECTED] writes: i want to write a mini-howto on all the black magic i had to go through to get debian onto this dell inspiron 8000 (i know there are lots out there; i want to write another). is there any specific txt2html or latex2html or some formatting tool that is standard for the HOWTOs? I'm just going with LaTeX for the time being, but it's not quite the style that they have (i.e., http://qref.sourceforge.net/Debian/reference/reference.en.html). form as much as function, Try going to their home page (http://qref.sourceforge.net). Right there at the bottom are instructions for getting the document source and generating html. Looks like they're using SGML and the debiandoc-sgml package. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Tab-Completion in gnuplot
Joey Hess [EMAIL PROTECTED] writes: Joerg Johannes wrote: Last week I helped a friend out with some gnuplot graphs on his SuSE machine. He asked me why I was typing in the whole filenames in gnuplot instead of using auto-completion with the Tab key. I said, because gnuplot does not support it. BUT - On SuSE it does!Back home I tried it on my sid box, and there is really no auto-completion. Does anybody know how I can turn that on? What did SuSE do to gnuplot??? Quoting /usr/share/doc/gnuplot/README.Debian: libreadline --- Yes, the built in readline of gnuplot is bad. However, libreadline cannot be used instead because it is licensed under the GPL, whereas gnuplot has special licenses (patches only). Linking those programs together is forbidden by the GPL. Please don't file bugs telling me to use libreadline in gnuplot... For this reason I generally build gnuplot from source and add the necessary config stuff to link to libreadline. Steps, roughly, are: apt-get install libreadline4 apt-get install libreadline4-dev cd /usr/local/src apt-get source gnuplot cd gnuplot-3.7.3/debian # Edit the file named rules and change --without-gnu-readline TO --with-readline=gnu cd .. dpkg-buildpackage cd .. dpkg -i gnuplot*3.7.3-1*.deb Then I go into dselect and put a hold on gnuplot so it doesn't get upgraded automatically during an apt-get dist-upgrade. Of course I don't fully understand all the GPL implications. I believe what the Debian gnuplot maintainer means is that it's ok to use GNU readline library with gnuplot, gnuplot just can't be distributed that way and so (s)he doesn't distribute it that way with Debian. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Kernel will not compile
Patrick Wiseman [EMAIL PROTECTED] writes: On Tue, 17 Jun 2003 at 1:42am, Roberto Sanchez wrote: : --- dhobner [EMAIL PROTECTED] escribió: : I did the following: : : apt-get install gcc : apt-get install kernel-package : apt-get install kernel-source.2.4.18 : apt-get install libc6-dev : apt-get install tk8.3 : apt-get install libncurses5-dev : : tar -jxf kernel-source-2.4.18.tar.bz2 : : make-kpkg clean : : :I beleive that you are missing a 'make config' (or menuconfig or xconfig). No, I'm having the same problem and I made config first. :Check your kernel-source directory and make sure that the .config exists :and has your selected options. : : make-kpkg --append-to-version=.030320 kernel_image : : The following error occurs: : : In file included from ksyms.c:50 : /usr/src/Kernel-source-2.4.18/include/asm/checksum.h:72:30:Missing : terminating character : ***other errors after this within the same header file : : Anyone know why this is not compiling? No, but I grabbed the 2.4.20 source and had all sorts of other problems. I'm on a 'testing' system; you too? I suspect it's the testing gcc which is causing the problem (from Googling another error I was getting trying to compile 2.4.20). Can anyone confirm? I think, in general, it's best to stick with 2.95 compilers for the kernel. There are exceptions, but the easiest route is to use gcc-2.95 for compiling kernel source. You can do this by editing the kernel-source-2.4.20/Makefile and setting HOSTCC=gcc-2.95 and CC=$(CROSS_COMPILE)gcc-2.95. If you have external modules you're building then you'll also want to add CC=gcc-2.95 to the environment. Like: CC=gcc-2.95 make-kpkg --revision foo.1 --bzimage kernel_image and then CC=gcc-2.95 make-kpkg --revision foo.1 --bzimage modules_image Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Why isn't there a cron.hourly?
Kevin McKinley [EMAIL PROTECTED] writes: You can set up ntp to use a time server inside the firewall, the same as it would use a public server (just in case you didn't know that already). Also note that a lot of the network routers that I've run across act as NTP servers. Their accuracy, of course, is up to whoever manages the router, but in my own case I'm not as interested in absolute accuracy as I am in keeping all my machines synchronized to the same clock. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Server - no video card
Mike M [EMAIL PROTECTED] writes: I want to save money as I expand my collection of Debian servers by running without a video card. On the rare occasion that I need console access to a machine I would pull it off the shelf and insert an AGP video card. I just tried it on one of my servers and the only thing I observed was a beep during the boot-up sequence. Several questions come to mind: Is this an acceptable mode of operation? Are others running in this mode? Should most motherboards being recently produced be expected to run without a video card? (Maybe it's a BIOS thing?) The BIOS on a lot of machines, particularly older ones, won't allow you to boot without a video card. In some cases you can tweak a BIOS setting to tell it to ignore the lack of a video card and boot anyway, *but*, you have to install a video card in order to set that in the BIOS, usually. Is it a valid approach? Yeah, I suppose there's nothing wrong with the approach, if your machines BIOS supports booting without a video card, but it's a MIGHTY pain to have to add a video card when something goes wrong. For most cases like this I generally just get a computer with built-in video, set it up while connected to a monitor and then disconnect the monitor and throw it in the closet. Either that or I shell out $30 for a cheapo VGA card. If I tried setting up a server without a video card I know I'd be cursing myself every time something went wrong and I had to open up the case and install a video card just to diagnose the problem. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Partitioning advice, 13.9 GB HD
Ben Kal [EMAIL PROTECTED] writes: [snip] AFAIK the guide to the size of swap is the amount of RAM: make it equal to or twice that amount. By that standard you can cut down swap at least to half what you now plan to make it. I don't know if Linux would like to fill swap space with the iso image when a cd burner is busy but I doubt it. [more snippage] AFAIK this isn't true of Linux. Some Unices required this in the past. The one I remember was HP-UX. If it did a kernel dump it copied memory to swap (for analysis (did anybody ever analyze a 128MB kernel image dump?)) and so you had to have enough swap space to hold the contents of the memory in case the kernel dumped. For Linux the amount of swap is like the size of /usr/local (or /opt). It's dependent on how you use the machine. I think it's generally true that it's a good idea to have a bit of swap (128MB is my typical minimum), but after that it's dependent on what you do with your machine. If you're running X + Gnome + Openoffice on a 64MB system you'll need a LOT of swap. If you're running in console mode and use the same system for C app-development, for example, you probably won't need much swap. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: blocking icmp...
Paul Johnson [EMAIL PROTECTED] writes: On Tue, May 27, 2003 at 10:51:15AM -0600, Gary Hennigan wrote: [snip] I've been blocking all incoming, non-stateful, ICMP for a number of years on my cable-connected LAN and have never had a problem, but I don't run any type of globally accessible server. Don't do this! If you were on @Home, you are one of the people who damaged me for life by doing this. 8:oP I was on @Home for about a month. Anyway, blocking ICMP is the least of my worries if I have to talk to support. Next time, for fun, call your support line and tell them you're running an OpenBSD firewall that's connected to your cable modem and prepare for, if you're lucky, a profound silence. In most situations I'd expect to hear that's not supported. In general, if I have to talk to support I swap cables in my wiring closet and hook my Win2k box up directly to my cable modem and then call support. Personally, I'd rather make my presence on the 'net as hard to discover as possible. If you allow echo requests it's a simple matter for someone to run nmap, for example, to find out that a particular IP address is valid. If you block such messages any cracker will likely just move on to the next poor slob when your IP address doesn't show up on his nmap scan. Better idea: Keep patched instead of relying on obscurity. Best idea: Use any and all tools at your disposal for security, *including* obscurity. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: blocking icmp...
[EMAIL PROTECTED] writes: On Sun, May 25, 2003 at 01:09:29PM -0400, Kevin McKinley wrote: On Sun, 25 May 2003 07:31:02 -0700 Paul Johnson [EMAIL PROTECTED] wrote: On Sun, May 25, 2003 at 09:56:07PM +0800, Hanz wrote: In setting up a firewall will there be any negative side effects if i block icmp? Well, other than it breaking the TCP/IP standard and making some servers think you don't exist (some ping back), no. How would declining to answer pings break the TCP/IP standard? That's like saying if you don't answer the telephone you're breaking the telephone standard. It's anti-social and hamfisted. Some CPAN servers are blocking icmp now and that makes it difficult to tell if they are even up. There are more precise and reasonable means, using netfilter (iptables) to protect against icmp DoS attacks. At the very least one shouldn't do something that one doesn't want others to do. Set a good example and all that. For a server I'd agree. For a home system I'm not sure there's any issue. I've been blocking all incoming, non-stateful, ICMP for a number of years on my cable-connected LAN and have never had a problem, but I don't run any type of globally accessible server. Personally, I'd rather make my presence on the 'net as hard to discover as possible. If you allow echo requests it's a simple matter for someone to run nmap, for example, to find out that a particular IP address is valid. If you block such messages any cracker will likely just move on to the next poor slob when your IP address doesn't show up on his nmap scan. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: default run level
Irish, Jon D BAE SYSTEMS [EMAIL PROTECTED] writes: All, This is my mistake (as usual ;-). I did not provide enough information in my post. I need to temporarily disable booting into a graphical environment (i.e. X) so that I can install the new Nvidia display drivers for X. After the install, I want to re-enable booting into X. Does this make more sense? Your original question made perfect sense, I just can't read! ;) Anyway, to temporarily shut XDM down you can simply Ctrl+Alt+F1 to get a console. Then log in as root and do /etc/init.d/xdm stop and install away. When you're done you can either reboot or /etc/init.d/xdm start Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sawfish 1.3 + Gnome2 questions?
Florentin Ionescu [EMAIL PROTECTED] writes: Does sawfish work with gnome2 ? Except for the edge-flipping issue, which is solved thanks to the respondents, it hasn't given me any trouble with Gnome2. I'm using sawfish version 1.3 from unstable. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: is there a i586 distro?
Ray [EMAIL PROTECTED] writes: or others that are compiled/optimized for the later i386 PCs? if so, where? if not, how should i go about mass source compiling debs for my computers? (i have a couple P1 laptops, P2s and a P4) and is it worth it? My suspicion would be that you wouldn't see much difference. Even if there was *some* difference I bet it'd be hard to justify the slight performance increase you'd see considering the amount of time a job like recompiling a significant portion of your installed base would take. Plus, if you're talking about something like the testing or unstable Debian distribution you'd be compiling/recompiling almost daily as packages were updated. The only thing that I vaguely recall someone mentioning would see a significant performance boost from recompiling for later processors were the encryption-based packages that need to do a lot of number crunching, but I would hazard that even there you're talking about, maybe, a few tens of percentage points performance increase, not orders of magnitude. If you're really Gung-Ho about doing it though, I'd probably consider a source-based Linux distribution instead of Debian. Something like Gentoo (http://www.gentoo.org). Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: default run level
Irish, Jon D BAE SYSTEMS [EMAIL PROTECTED] writes: Here is a newbie question: Which default run level do I change inittab to so that the PC boots to a VGA console instead of X? On a clean install, level 2. Look in /etc/inittab and there you'll see a line like id:2:initdefault: That means 2 is the default level. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: default run level
Dave Sherohman [EMAIL PROTECTED] writes: On Thu, Apr 03, 2003 at 03:06:34PM -0700, Gary Hennigan wrote: Irish, Jon D BAE SYSTEMS [EMAIL PROTECTED] writes: Here is a newbie question: Which default run level do I change inittab to so that the PC boots to a VGA console instead of X? On a clean install, level 2. Bzzt! Wrong answer... Your answer applies only to Red Hat (and Red Hat-derived distros). It does not apply to debian. Right answer, wrong question! :) Actually, I did read the question incorrectly. But you certainly can use runlevels to decide whether or not you want to start XDM. I set mine up that very way. RL 2, does not start XDM, RL 3 does. Easier for me to boot it the way I need it, especially on my laptop. And yeah, the way you customize your run levels globally is with update-rc.d. But, and I may be wrong, *again*, if you want to customize a specific run level, for example, remove xdm from RL 2, you just go into /etc/rc2.d and remove the appropriate link manually. Debian is also smart enough, I believe, so that it won't overwrite the changes you make if you do things manually like this. And isn't there still an issue if you globally do update-rc.d -f xdm remove for example, the next time xdm gets updated via apt/dselect/etc. it will reinstall the xdm links if you have your configuration priority high enough? Or did that get changed? Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Galeon 1.3.3 and URL completion
Scott Henson [EMAIL PROTECTED] writes: check the debian-gtk-gnome archives. This was discused maybe a week ago. Found it! Under the Debian galeon packages thread. In summary, bring up the gconf-editor and go to apps-galeon-snapshot-Browsing-History and select completion_enabled. Voila! It's back to doing things the Right Way(TM) in terms of url auto completion. Thanks for the pointer Scott. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Sawfish 1.3 + Gnome2 questions?
A couple of questions for those running sid with Gnome2 and Sawfish. 1) I can't figure out how to turn edge-flipping on. That's the ability to drag a window between workspaces. I use that feature quite heavily and am loathe to use the little teeny things in the pager to try and accomplish this. I read some discussion on the edge-flipping capbility in the sawfish bugzilla database and my interpretation was that it was removed, but the discussion was from last year and a poll showed overwhelming support for the feature so I'm hoping it didn't disappear. 2) Using Nautilus, I can't seem to get the sawfish configurator. If, for example, I click on preferences:///Windows I get: --- Cannot start the preferences application for your window manager Window manager Sawfish has not registered a configuration tool --- Of course I can middle-click on the background to get to the sawfish configurator, but why isn't it working via Gnome2? Relevant (maybe) version info (via apt-show-versions): sawfish/unstable uptodate 1:1.3-1 nautilus/unstable uptodate 2.2.2-1 Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sawfish 1.3 + Gnome2 questions?
Gary Hennigan [EMAIL PROTECTED] writes: A couple of questions for those running sid with Gnome2 and Sawfish. 1) I can't figure out how to turn edge-flipping on. That's the ability to drag a window between workspaces. [snip] I managed to get this working by manually copying the appropriate lines out of a ~/.sawfish/custom file on a machine on which edge-flipping does work. Here are the relevant lines: (custom-set-typed-variable (quote edge-flip-only-when-moving) (quote t) (quote boolean)) (custom-set-typed-variable (quote edge-flip-type) (quote workspace) (quote (choice viewport workspace))) (custom-set-typed-variable (quote edge-flip-enabled) (quote t) (quote boolean) (quote sawfish.wm.ext.edge-flip)) After I added those and restarted sawfish I can drag windows between desktops. As far as I can tell there is NO way to set this from the sawfish customization tool. The only reason I had the lines above in my custom file, I believe, is because that machine was upgraded from a previous sawfish installation instead of starting from scratch. Version wise the machines are identical. #2 is still a mystery, i.e., why no interface from Nautilus to the sawfish customization tool? Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Galeon 1.3.3 and URL completion
I'm running galeon-snapshot in sid which is version 1.3.3. One of the things about Galeon that I really like(d) is that while you manually type in a URL it would progressively find the best match for what you were typing as you typed. If you satisfied the uniqueness you could just hit enter without typing in the rest of the URL. Well, that's the way it used to work. Now it appears to be working like MS Exploder and instead of allowing you to type enough to uniquely specify a previously used URL and hitting the Enter key, it wants you to select from the URL pulldown using the mouse. I'm an old fashioned kind of guy and I use the mouse to switch between windows and scroll windows. I hate having to take my hand off the keyboard to select a URL. I'm hoping that you can set the latest Galeon to behave like the old, but if it's there I can't find it. Anyone else notice this behavior, and hopefully know how to set it back to the old way it did things? Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Mozilla 1.3 and Java plugin
gaumer [EMAIL PROTECTED] writes: On Fri, 2003-03-28 at 10:28, Pedro I. Sanchez wrote: I can't get Mozilla 1.3 to recognize the java plugin that I install in my PC. I'm running Sid, and I've tried the Blackdown Java and Sun's Java run times, both 1.4.1. With both Java implementations, after doing the install, I go to /usr/lib/mozilla/plugins and do a soft link to the libjavaplugin_oji.so file that comes with the Java distribution. I verified that all the libraries are present by running the ldd command against this .so file and everything looks OK. However, when I relaunch Mozilla I don't get any java plugin listed in the about:plugins page. I however see all other plugins that live in the /usr/lib/mozilla/plugins directory. Suns version of Java does not work on Mozilla builds that used GCC 3. http://plugindoc.mozdev.org/whichjava.html http://plugindoc.mozdev.org/linux.html#Java Get the correct version and all will be fine. [note: top-posting whacked] I had a similar problem with the Moz 1.2 I'm using on my unstable/testing system. The solution that was suggested to me here my a kind user was to get the 1.4 jdk compiled with gcc 3.2 and then I just copied over the java plugin. You can find the a Java 1.4 distribution, compiled with gcc 3.2, at ftp://metalab.unc.edu/pub/linux/devel/lang/java/blackdown/JDK-1.4.1/i386/01/ get the file j2re-1.4.1-01-linux-i586-gcc3.2.bin. Run it (it's a self-extracting shell-script) and then go into j2re1.4.1/plugin/i386/mozilla and copy javaplugin_oji.so to your mozilla plugin directory. Again, this is for Moz 1.2, not sure if it'll work with 1.3. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] Backup solutions
Bob Paige [EMAIL PROTECTED] writes: I've been running my debian system without a backup for about a year now. I understand the need for backup (hence this message) so please, no assaults for not using one. Questions: 1) what is better for backup, tape, or CD? (I already have a CD burner) 2) if tape, what is a good/inexpensive product to use? 3) what software works best? I did some googling and found Linux-Backup.net (and others). It appears the state of backup for Linux is to use tar or afio and backup entire images. I haven't found examples of backing up multi-GB archives that span multiple tapes/CDRs. I'm imagining a baseline backup with periodic (maybe weekly, maybe more often) incrementals. I can easily copy my relevant files (3GB+) to a bunch of CDs, but then how do I keep track of newer backup copies that should replace older versions? I am looking for software smart enough to do this. Opinions? First opinion, don't start a new thread by replying to an already existing thread. In this case you replied on the debian 2.0: some intruder broke in thread and changed the Subject. Just about any reasonable mail client (MUA) is threaded these days and I saw your topic only by chance because I was reading the original thread. If I, or anyone using a threaded mail client, had been ignoring the original thread your topic wouldn't have appeared. This is independent of the Subject, mostly, and uses the References header, which you didn't modify. Back to the question... Tape is the best method, IMO, but you have to have some spare money laying around if you don't already have a tape drive. If you do decide to get a tape drive check out eBay for pretty reasonable 4mm drives, and probably other formats. Of course a lot of them are SCSI so if you don't already have a SCSI adapter that's another consideration. DDS2 tapes are pretty cheap. Given that, I like afbackup. Like amanda it can do network backups, and as far as I can tell everything else Amanda can do. When I first looked at Amanda it had some, to me, weird limitation. Something like having to fit all of the data for a run onto one tape? That's fine if you've got a big tape drive, but my little DDS2 4mm drive won't fit a full backup of my primary system onto one tape. afbackup has no such limitation. The one drawback of afbackup, and it's getting better all the time, is documentation. When I first started using it it was horrid. I don't think the author was a native english speaker, or was just a bad writer, and it showed. However it's gotten better with each release and is now not too bad. I'm not complaining, because I didn't do anything to help make it better. The software, in contrast, works fantastically. Never a problem, and I have had to do plenty of recoveries. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Memory leak - somewhere :)
Jeetu Golani [EMAIL PROTECTED] writes: I have a Debian 3.0 installation on my P4 2.2 512MB with 256MB Swap system. I use KDE 3.1. For sometime, I've been feeling that there's a memory leak in one of the apps because after quite a few days of use KDE becomes slow. Apps I typically use are KMail,Konqueror,KMerlin,Konsole,K3B. I've noticed that memory seems to be consumed at an alarmingly huge rate. Sometimes about 300MB+ gets used up overnight inspite of no additional programs being run. For e.g. if with a clean boot into KDE I have 300MB of free Physical memory left (as displayed by free) and the whole swap free then the next morning I may see just 5MB of Physical memory free and the Swap used to a degree. I thought it could be KDE or artsd or some such component/app and I kept rebooting my system with one less program running each time i.e. i'd turn on KDE and not switch on KMail and see how it performs and next time wouldn't switch on KMail and K3b,etc.did this till i basically had a bare KDE session running but the memory usage issue continued. I therefore thought it may be X and I just put a # in the default-display-manager file in /etc/X11, therefore since there's no default manager I just get a message stating that at bootup and X doesn't come up. If I do a free in succession with just this much I notice that my memory is going down at a rate of 8KB each successive reduction.this is most of the time...although sometimes reductions can be of different sizes. I can't make much of it.is this normal?? doesn't look like it should.since I'm basically not turning on any programs...therefore something seems to be up...any help on either clarifying my understanding on the above or tracing the rogue program would be appreciatedhas neone else noticed something similar. This has got to be in a FAQ somewhere. In all my years of using Linux and reading mailing lists and Usenet I've seen this question more than any other. Anyway... Linux allocates all free memory for disk cache. If an application requests memory and all of it is being used by the disk cache then Linux frees up some of the memory being used by the disk cache for the application. On a healthy system, that you've been using for a while, you should have very little free memory. That doesn't mean you don't have something that's leaking memory, it just means you'll have to look harder to find out if there is indeed something sucking up memory. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Gnome2 user-defined menu?
I've mostly migrated over to Gnome2, in unstable, and most things went relatively smoothly. The one exception, so far, is user-defined menus. How do I add such a beast? Under Gnome 1.x I had a menu that sat in one of my panels where I defined commands, via a Launcher, like gnome-terminal -x ssh remotecomp so I could pop up a window connected to, for example, remotecomp. I can't for the life of me figure out how to do this under Gnome2! It's been a long time since I did it under the old Gnome, but I coulda sworn it was as easy as clicking on the Panel and selecting Add Menu. That doesn't appear to be the case with Gnome2. I've tried navigating through the documentation for vfolder, .desktop, .directory, etc., but unless it's absolutely necessary to read all that gunk to add a simple user-defined menu I'd prefer not having to learn it. I also tried using a Drawer and that would almost work, but I don't want it to display an icon for the items, it needs to display a simple user-defined name, like remotecomp instead. Any suggestions? Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Gnome2 user-defined menu?
Gary Hennigan [EMAIL PROTECTED] writes: I've mostly migrated over to Gnome2, in unstable, and most things went relatively smoothly. The one exception, so far, is user-defined menus. How do I add such a beast? Under Gnome 1.x I had a menu that sat in one of my panels where I defined commands, via a Launcher, like [snip] Ok. I finally found something that explains it. It's not the easiest thing to understand, but it's definately easier than trying to learn the whole of the Gnome VFS to add a menu! ;) http://www.geocrawler.com/archives/3/263/2002/8/0/9453293/ Shoulda known. The answer for just about everything in Gnome2 is Nautilus! :) I'm still all ears if anyone has something easier than this. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: using diff command
Curtis Vaughan [EMAIL PROTECTED] writes: I must be really stupid, so forgive me, but I can't figure out for the life of me how to use the diff command. I've read the man pages and looked at some stuff on the internet, but I can't get it to do what I want it to do. But then maybe it doesn't do what I want it to do. Basically, the issue is this, I have 2 files: File1 File2 Both are text files. File1 was File2 a day ago. Since then File2 has had additional information tagged on to it (it's a log file). All I want to see is what information has been added since yesterday. So, I would think that diff File2 File1 should provide me with that information. But all I ever get is a message that the files differ. I have tried with various arguments, but no luck. Hmm. You got somethin' funky going on! What you describe should only happen if you're specifying the -q or --brief option to diff. What does diff --version give you? Make sure you're running the actual diff binary, and not via some alias. Do /usr/bin/diff File2 File1. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: using diff command
Gary Hennigan [EMAIL PROTECTED] writes: [snip] Hmm. You got somethin' funky going on! What you describe should only happen if you're specifying the -q or --brief option to diff. What does diff --version give you? Make sure you're running the actual diff binary, and not via some alias. Do /usr/bin/diff File2 File1. One other thing that might cause this is if diff thinks the files are binary. Could be they have a binary character in them somewhere and this is throwing diff off. Try: diff --text File2 File1 Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Unstable/Java/Mozilla
Florentin Ionescu [EMAIL PROTECTED] writes: There is blackdown java compiled with gcc3.2 - but there is no debian package AFAIK and that should solve the problem. ftp://ftp.mirror.ac.uk/sites/ftp.blackdown.org/java-linux/ JDK-1.4.1/i386/01/j2re-1.4.1-01-linux-i586-gcc3.2.bin [snip] This will work until *.deb files are available. I simply downloaded the j2re compiled with 3.2, unpacked it and copied the included javaplugin_oji.so over the one installed by the previous debian package (j2re1.4). Java is now playing with Mozilla again. Many thanks! Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Unstable/Java/Mozilla
I just realized that due to the fact that Mozilla, and it's descendents, in unstable, is now being built with the 3.2 compilers I'm stuck without a useable Java plugin. Anyone have a solution to this? I've been using the 1.4 *.debs from Blackdown, but those appear to have been compiled with 2.9x compilers and so I'm getting unresolved symbols when I try to load up pages that require the Java plugin, and Galeon crashes There's a resolved bug filed against Mozilla which seems to indicate that the maintainer doesn't consider it a problem. Any word on a Java2 JDK being compiled with 3.2 compilers? Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: X-over cables [was Direct cable connection]
Hugh Saunders [EMAIL PROTECTED] writes: on the subject of crossover cables, do you know where you can get adapters that go on the end of normal patch cables to convert them to crossover? The reason i ask is because i only want to carry one network cable in laptop bag but would be useful to have a crossover cable sometimes.. Aside from the fact that, as another respondent pointed out, you'll still need another cable to plug in to the other side of the adapter, and so you're probably ahead to just have two cables, google is your friend: http://www.google.com/search?hl=enie=UTF-8oe=UTF-8q=ethernet+crossover+dongle%7CadapterbtnG=Google+Search Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Nvidia driver
Jack Pistachio [EMAIL PROTECTED] writes: Just comment out the corresponding line for GLcore and dri. ie: # LoadGLcore # Loaddri Then change the driver line in Section Device to: Driver nvidia Also, make sure you have the NVIDIA-GLX package installed to. That should work for you. If you do this make sure you tell/told debconf that you want to handle the XF86Config manually, otherwise whatever you do will likely get overwritten if a new version of xserver-xfree86 gets installed. Just do: % dpkg-reconfigure -plow xserver-xfree86 and answer No in response to the question about managing XFree86 server configuration with debconf. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Kernel-sourcecode directory
Willem-Jan Meijer [EMAIL PROTECTED] writes: Hello for the 3rd time today, I've downloaded the nvidia-kernel-src and nvidia-glc-src and read the documentation. You have to give a command wich gives this result: We do not seem to be in a top level linux kernel source directory tree. Since we are trying to make a kernel package, that does not make sense. Please change directory to a top level linux kernel source directory, and try again. (If I am wrong, and this is indeed a top level linux kernel source directory, then I have gotten sadly out of date with current kernels, and you should upgrade kernel-package) So I downloaded the kernel-source-2.4.18 and installed that package. I searched my computer but I can't find the correct directory. I still get this message. Does someone know where to find this directory/how to install the nvidia drivers? cd /usr/src tar xjvf kernel-source-2.4.18.tar.bz2 cd kernel-source-2.4.18 make config|xconfig|menuconfig make-kpkg --revision mymachine.1 --bzimage kernel_image make-kpkg --revision mymachine.1 --bzimage modules_image cd ../nvidia-glx-1.0.2880 dpkg-buildpackage cd .. dpkg -i kernelimage-2.4.18_mymachine.1_i386.deb dpkg -i nvidia-glx_1.0.2880-3_i386.deb nvidia-glx-dev_1.0.2880-3_i386.deb The above assumes you have the prereqs required to install all that stuff, and, of course, you have the kernel-package and dpkg-dev packages installed. There may be more elegant solutions, but that's how I do it, to the best of my recollection anyway. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: New Kernel - No Network
[EMAIL PROTECTED] writes: I just compiled and installed linux-2.4.20. I had been running linux-2.2.22 without network problems. The new kernel does not configure the network. At startup, the new kernel says it can't find module 3c59x, the driver for the 3COM network card (it also says it can't find the unix module). If I boot the older kernel, I have no network problems at all. I'm relatively new to Linux and don't know how to troubleshoot the new kernel. The modules and modules.conf files in both kernels look the same. Please advise. Thanks. Vince You compiled the kernel yourself? If so you forgot to select the 3Com driver for your ethernet adapter. You'll have to recompile the kernel with that module. Using make xconfig in the kernel-source directory, go to Network device support-Ethernet (10 or 100Mbit) Click y for 3COM cards and then select m for [3c590/3c900 series (592/595/597) Vortex/Boomerang support]. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] what are you running? (was Complaint)
On Wed, Feb 26, 2003 at 05:07:16PM -0600, Ray wrote: what are people on the debian-user list using for there primary machine(s)? the last few messages i see are from users of (going by User-Agent or X-Mailer header) kmail mozilla on windows mozilla on debian xemacs mutt I'm using gnus, via GNU Emacs. It's a bear to set up, but I've been using it since the Epoch (I exaggerate, slightly) and when I used to read Usenet it was nice to have a common front-end for mail and news. That and mail expiry, which I believe Mutt can now do, were my main motivations for using it. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: copying out of openoffice
martin f krafft [EMAIL PROTECTED] writes: I don't seem to be able to copy text out of an OpenOffice.org text document. I select the text, I even choose 'Copy' from the 'Edit' menu, but when I try standard X pasting, it doesn't work. Why? Is OpenOffice just not behaving like a proper X application??? I can't help you, but I can say that it works fine for me. % apt-showversions|grep openoffice openoffice.org-bin/unstable uptodate 1.0.2-1 openoffice.org-l10n-en/unstable uptodate 1.0.2-1 openoffice.org/unstable uptodate 1.0.2-1 openoffice.org-debian-files/unstable uptodate 1.0.2-1+1 % apt-show-versions |grep xfree xserver-xfree86/testing uptodate 4.2.1-3 xfree86-common/testing uptodate 4.2.1-3 At this point, I'm running a mostly unstable system, but with some testing mixed in, as you can see. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: sit (compressed files)
nate [EMAIL PROTECTED] writes: Emma Jane Hogbin said: [snip] perhaps it's simply not installed? try dpkg -L macutils from the looks of the package there is a README.unsit in /usr/share/doc/macutils that may give some hints .. The info seems dated. unsit functionality has, apparently, been incorporated into macunpack. Unfortunately, there's are bugs filed against the package http://bugs.debian.org/cgi-bin/pkgreport.cgi?pkg=macutils that indicate that it doesn't handle at least one of the newer compression algorithms used in a sit archive. I can't believe there's not something out there to do this, but, at the very least, I think you'll have to look outside Debian to find such a utility. So you can try macunpack, which is in macutils, but if it fails you'll have to hunt further. That, or tell the person who sent it to you to use Zip or something for which there are Linux utilities. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Memory usage on debian
Nils-Erik Svangård [EMAIL PROTECTED] writes: My system use about 95% of my 512 mb ram, but ps aux and top doesent show which process that eats all the memory. Can anyone figure out what to do this could be a kernel issue or something? Linux uses unused memory for a disk cache. This is dynamic and so if you start a process that needs the memory the kernel will shrink the disk cache. Use free (/usr/bin/free) to see how much disk cache is allocated. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NVidia source compiling problems
Joris Huizer [EMAIL PROTECTED] writes: I've been trying to get the packages nvidia GeForce 2 MX video card working before before christmas but I gave up then and temporarily installed the redhad 8.0; :-S Today I installed Debian woody again. And, unfortunately, I still don't see what's going on and why it's not working - I installed the kernel source and I point to that but the nvidia-kernel-source doesn't compile - still complaining about a missing .h file :-( I'm sure someone could be of a lot more assistance if you posted the full error message you're getting. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NVidia source compiling problems
Joris Huizer [EMAIL PROTECTED] writes: Sorry - because lynx couldn't log in on yahoo.com I have to reboot in windows to read mail which is why I didn't add it before. I've got the nvidia-kernel-src and nvidia-glx-src both in /usr/src. I've used the following commands (as suggested by Ochronus somt time ago) cd /usr/src/ tar xvfz nvidia-kernel-src.tar.gz export KSRC=/usr/src/ - or - export KSRC=kernel-headers-2.4.18-k7/ export KVERS=2.4.18 cd modules/nvidia-kernel-1.0.3123 debian/rules binary_modules Because the kernel is pre-compiled this gives problems (cc mismatch) - therefor I edited the Makefile in the NVIDIA subdir, so it won't stop right away. Then when I call the 'make' prog I get this error output: cc -c -Wall -Wimplicit -Wreturn-type -Wswitch -Wformat -Wchar-subscripts -Wparentheses -Wpointer-arith -Wcast-qual -Wno-multichar -O -MD -D__KERNEL__ -DMODULE -D_LOOSE_KERNEL_NAMES -DNTRM -D_GNU_SOURCE -DRM_HEAPMGR -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE -DNV_MAJOR_VERSION=1 -DNV_MINOR_VERSION=0 -DNV_PATCHLEVEL=3123 -DNV_UNIX -DNV_LINUX -DNVCPU_X86 -I. -I/usr/src/linux/include -Wno-cast-qual nv.c In file included from nv.c:14: nv-linux.h:24: linux/modversions.h: No such file or directory make: *** [nv.o] Error 1 How do I get around this ?? Do you have kernel-headers-2.4.18-k7 installed? I've never built the nVidia stuff for use with a precompiled kernel. I've only done it as part of compiling my own kernel. At a minimum, I'm guessing, you need to install the kernel-headers package that corresponds to your running kernel in order to compile the nVidia drivers. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian as a Personal Video Recorder??
[EMAIL PROTECTED] writes: For playing, transcoding, and yes recording, I recommend mplayer. For the unattended part, use with good old reliable, at or maybe even cron. Here's simple command to record a TV ^^ show using mjpeg codec: aumix -i 50 -l 50 ; mencoder -tv on:driver=v4l:norm=ntsc:chanlist=us-cable:channel=36:width=352:height=240:forceaudio:forcechan=2:adevice=/dev/dsp -ovc lavc -lavcopts vcodec=mjpeg -oac pcm -o tvcapture@`date +%Y`-`date +%m%d`-`date +%H%M`.avi Don't have anything relevant to add, but simple is not a word I'd use to describe the command line shown above! ;) Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: XEmacs and Gnome Terminal copy/paste
Jorge Santos [EMAIL PROTECTED] writes: It seems like I can't paste with the middle button to the Gnome Terminal (2.x) text selected inside XEmacs, however pasting to other programs work as does pasting the other way around. Anyone can confirm this and/or suggest a solution? BTW, I'm using Gnome Terminal 2.1.0 and XEmacs 21.4 (patch 6) both from testing. It doesn't solve your problem, but it works fine for me. Whenever I have weird problems with Emacsen I always start by renaming/moving my ~/.emacs (or ~/.xemacs) to some scratch name and try whatever it is without an Emacsen init file. Does that show you anything? In other words, make sure you're not putting something in your xemacs init file that is causing the problem. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: cups and hpijs driver
Phil Reardon [EMAIL PROTECTED] writes: How do I get the hpijs driver for cups on my debian sid box? I am trying to configure an hp photosmart 7350 printer, but when I went to http://localhost:631/ and tried to set up the printer, there were only three hp drivers listed, one each for deskjets, laserjets and new deskjets. Thanks in advance. I went through this not long ago and posted a summary of the steps I took. Google Groups is your friend! http://groups.google.com/groups?q=group:*debian*+insubject:cups+author:henniganhl=enlr=lang_enie=UTF-8oe=UTF-8selm=nhxwun02tn4.fsf%40sadl12238.sandia.govrnum=1 Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How insecure are cable connections, versus dialup?
Dennis G. Wicks [EMAIL PROTECTED] writes: Sorry, Gene. That link is now a petition to abolish the DMCA! Do you have an alternate link? Google it. Here is the first result of such a google: http://naughty.monkey.org/~dugsong/dsniff/faq.html Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: switching between two network configurations
Jörg Johannes [EMAIL PROTECTED] writes: With my laptop, I'd like to access two different networks (depending on where I am): My home network and university network. At home, my IP is 192.168.0.99 and hostname shold be horchloeffel, at university the IP and hostname are set to other fixed values. What I'd like to do is choosing at boot-time which network configuration to load. Is it possible to write a shell script that is invoked by a lilo parameter? It would have to set the hostname, the nameservers and the IP adress according to predefined values (should be simple to write two configurations, say /etc/hostname.uni and /etc/hostname.home, one of which is copied to /etc/hostname and so on). Is this possible, or are there other (better) solutions for doing this task? It pays to know how to use Google Groups! http://groups.google.com/groups?safe=imagesie=UTF-8oe=UTF-8as_ugroup=*debian*as_usubject=laptop%20environment%20detectionlr=lang_ennum=100hl=en This should net you a thread started by Martin Krafft on this very subject about a month or two ago. He surveyed quite a few utilities with the capabilities you desire, and more. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] spam advice
[EMAIL PROTECTED] writes: at the bottom of many spam-type emails there is a get out clause or something to remove my name from the database. Surely clicking on this link/sending an email to the adress specified simply verifies that your address is valid? Is it adviseable to attempt to use these measures to be removed from lists? or just keep on writing filitering rules? I've generally avoided doing this for the very reason you state, it might just verify for the spammers that your email address is valid and it'll get bumped up to a premium list and they can charge more for it. I'm paranoid though! ;) I've had great luck with bogofilter (it's in unstable and testing I believe). It uses Bayesian statistics to try and determine if an email message is spam. Of course the drawback is that you have to train it by feeding it spam, and non-spam, to analyze. In my experience this is a short duration process. After feeding it maybe 20 spams it was getting it right about 95% of the time. It also seems to be getting more and more accurate as time goes by. It's easy to setup a procmail rule (examples in the man pages) so that all mail passes through bogofilter and thus it improves itself with time. When necessary, you can feed it back mistakes to reverse what it learned before. Anyway, I've been very happy with it. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: In a bind about named and network printing.
Paul Lewis [EMAIL PROTECTED] writes: I have been playing with my local networks 'named' and finally got it working more or less the way I think it should. Evidence for this is that I can ping local machine name and it returns the fully qualified machine name plus the ping responces. This works on all machines except the server 'named' is actually running on. Where the ping goes off along the cable modem never to resolve (my network is strictly a local affair). So far so good. How do your internal machines resolve external addresses? Does each of them have a resolve.conf that points to your ISP nameservers, as well as your internal nameserver, or do you have your internal name server query your ISPs nameservers? You really want the latter to have everything work the Right Way (TM). I'm assuming you don't have your local bind configured to query your ISP name server? That's probably what you need to do to fix things up completely and then put into /etc/resolv.conf on your box running named: domain whatever.you.call.your.home.domain search whatever.you.call.your.home.domain nameserver 127.0.0.1 nameserver ISP nameserver #1 IP nameserver ISP nameserver #2 IP The 127.0.0.1 will cause your server to query itself first, then, if necessary go to your ISP nameservers. You'll have to look at the bind docs to see how to set up your bind server to forward requests for external IPs to your ISPs nameservers. In Bind 4 (my server ran OpenBSD), after you defined your local zones in named.boot you put: forwarders 10.x.x.x 10.x.x.x options forward-only replacing the two 10.x.x.x addresses with your ISPs nameserver IP addresses What is really causing me to scratch my head is that suddenly ( at the same time ) the printer located on my server (the same server which runs named) no longer works for print jobs across the network. The print jobs are all queued on my debian machine but never get printed on my redhat server. I am pretty sure it is related to my tinkerings but just can't see where the problem lies. Any suggestions? Fix the first prob, then the second may disappear. If not you can ask again after the first problem is resolved! Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: X problem
Willi Dyck [EMAIL PROTECTED] writes: On Tue, Dec 03, 2002 at 01:17:18PM -0500, Bruce Park wrote: Hello Debian users, I was wondering how I can fix this problem regarding the root user and X. Normally, I log into the machine as a regular user then use the su command to become root. Whenever I need to utilize X(xmms, emacs), I get an error stating the following: [snip] Your X server is refusing any connection from clients as long as authentication is enabled. Disable by 'xhost +'. Or as root, copy over the .Xauthority file from the users home directory. Cookies are stored in that file, without which, access to the X server is not allowed. Don't forget to set the DISPLAY varible 'export DISPLAY=:0.0'. For more info, read the Remote-X-Apps-HOWTO at linuxdocs.org. The best solution I've seen for this problem, and the one I use, is after you su: xauth merge ~user/.Xauthority ;export DISPLAY=:0 replace user with whatever your login ID is. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: installing new kernel-image
Rodrigo Agerri (ylbaggar) [EMAIL PROTECTED] writes: in order to install a new kernel in stable, I did apt-get install kernel-image-2.4.18-686 I configured /etc/lilo.conf keeping the older kernel (2.2.20) such as: image=/vmlinuz label=linux read-only image=/vmlinuz.old label=kernel-2.2.20 read-only I then executed lilo and got this output: Warning: Int 0x13 function 8 and function 0x48 return different head/sector geometries for BIOS drive 0x80 Added linux * Added kernel-2.2.20 if I try to boot with the new kernel, I get kernel panic (I still can boot with the old one) what did i forget? I build my own kernels from source so I might be off, but don't the stock 2.4.x kernels require you to boot with initrd? That could explain your kernel panic. The drivers necessary to get to your hard drive, or the file system of your root partition, may not be built in to the kernel you installed. With initrd the kernel will boot with initrd, load the modules necessary to continue booting your system and then proceed to boot from your root file system on the hard drive. Unfortunately, since I don't use initrd, I don't know how to fix your problem. In all likelihood the best way to go is to let the kernel-image installation modify your lilo.conf itself and then make the modifications you desire after the installation script for kernel-image has done its thing. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: CUPS Samba print server under Debian
Thanks for the reply Derrick. Next time please try to be a little quicker though would ya? :) I ended up spending a few hours figuring everything out. Derrick 'dman' Hudson [EMAIL PROTECTED] writes: On Tue, Nov 19, 2002 at 01:55:16PM -0700, Gary Hennigan wrote: | I'm in the process of setting up a Linux server for my Home LAN and | have some questions regarding the setup. First question is in regards | to CUPS. I see that under cupsomatic-ppd it says that you should try | foomatic-bin and foomatic-db first. Just install cupsomatic-ppd and select the driver for your printer. (it's not really a driver; rather it is a config file that tells cups what your printer can and can't do and how to talk to the printer) This works, but a lot of the cupsomatic-ppd drivers for my printer (HP 932C) have unmet dependencies in Debian, or just flat out don't work. The one's I can remember were the stp driver and the hpdj driver. The hpdj driver tries to call ghostscript with a hpdj device, which isn't in ghostscript and was probably superseded by the hp ijs driver. There are lots of hpdj drivers built in to ghostscript, but the PPD file was trying to invoke gs with one that didn't exist. I don't remember the details on the stp driver, but it had similar problems. | The next question relates to using this printer via Samba from a | Windoze client. I understand the Windoze client needs the native | Windoze driver for my printer installed, Yes and no. The Windows architecture requires the clients to do all conversion to printer-native data streams. The secret here is that you can set up a generic postscript driver on the windows machine and your cups configuration will handle the translation to printer-native just as it does for all of your unix applications. I use this for a Canon BJC-610 which doesn't have any good windows drivers (except for the one that came with it which doesn't support networking). That would work, except I typically print photos from my Windoze box and you typically lose something going through the CUPS driver if you do this, since it doesn't have some of the enhancements that the Windoze drivers have for printing photos on inkjets. [snip] | Will I lose that capability by attaching the printer to the Linux | box, or does it basically send raw printer command over the Samba | link and effectively bypass the Linux driver, which may not have all | the capability of the Windoze driver? If you set up the print spool in cups as type raw then you can use the windows driver to generate the printer-native data stream. In raw mode cups will merely queue and then deliver the data to the printer. (See section below on how I FINALLY got this to work) You can configure the printer with 2 (or more) queues directed to it. One use for that is to have one configuration with the cupsomatic driver for your *nix apps to use and one configured as raw for the windows clients to use. I was aware of this. I do it all the time with my PS printer at work. One queue for single-sided, one for double, etc. Here's my summary for those that are in a similiar situation: Linux side drivers: 1) HP now has a sourceforge project for drivers for a lot of their printers, including the ink jets, which is what I have. Just go to http://hpinkjet.sourceforge.net/ if you want to read up on the ink jet drivers. 2) Install the hpijs Debian package. I think this might recommend or require the gs-esp package. Just pull in any other packages it recommends. 3) Go to http://www.linuxprinting.org/show_driver.cgi?driver=hpijs. Near the bottom you'll see the PPD-O-Matic. Select your printer, for example, HP DeskJet 932C and hit the Generate PPD file button. Save the result to your system as a text file in the directory /usr/share/cups/model. At the top of the file will be a line that tells you what to name the file. In my case that was HP-DeskJet_932C-hpijs.ppd. I don't believe the name is critical, as long as you give it the .ppd extension, but use the recommended name if possible. Also note the field *ModelName: in the PPD file. You'll use that in #5. 4) Restart cups with /etc/init.d/cupsys restart 5) Now go through the CUPS interface to Add the printer. Make the obvious choices and when you come to Model select the field as it was shown in the PPD file under *ModelName:. For me it was HP DeskJet 932C, Foomatic + hpijs 6) Through the CUPS interface Configure the printer with the options of your choice. 7) The only problem I have with my printer using the above is that the margins are off. I need to tweak the *ImageableArea Letter/Letter field in my PPD file to fix this. Raw printing with CUPS via Samba from a Windoze box: This took me too long to figure out! It should be easy, you just configure everything on the Linux box as Raw via CUPS and install
Re: Determining what packages are installed
Joe Riel [EMAIL PROTECTED] writes: How do I determine what packages are installed? Or determine whether a particular package is installed? I can use deselect to check a particular package, but there probably is a better way. Please don't start a new thread by replying to an existing message. In most mail readers there is a capability called threading that puts mail in a tree with the original question at the highest level and all replies beneath it. If you reply to such a message and simply change the Subject your message will still be placed within the thread of the original message. This is bad for a couple of reasons, 1, people trying to follow the original thread will become annoyed that your polluting it with a totally new subject and, 2, if people are reading mail in a threaded mail reader and they're not interested in the topic of the message that you replied to they'll likely never see your message, even if they actually might be interested in the new Subject you started. So it's bad for the readers and bad for the writers to the list. Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Questions before installing w2k or XP Dual-Boot
Michelle Storm [EMAIL PROTECTED] writes: I'm planing on installing Windows 2k Pro or XP, on a 15gb HD. But I want to know the proper way to do this before I try it, as I can't really afford to reformat and start over with my entire system. Currently I have the following setup. FilesystemTypeSize Used Avail Use% Mounted on /dev/hda3 reiserfs 24G 1.5G 22G 7% / proc proc 0 0 0 - /proc devpts devpts 0 0 0 - /dev/pts /dev/hda1 ext37.6M 4.8M 2.5M 67% /boot /dev/hda5 reiserfs 49G 15G 34G 31% /home /dev/hda6 reiserfs2.4G 33M 2.3G 2% /tmp /dev/hdb1 reiserfs 15G 33M 15G 1% /data usbdevfs usbdevfs 0 0 0 - /proc/bus/usb /dev/hdb1 will be what I'm going to install Windows 2k Pro or XP on. What I need to know is, when I install Windows 2k Pro or XP, what will I need to do to ensure that my Linux is still bootable. Is there something I need to do PRIOR to installing windows? Only reason I am installing windows is that there are some programs that I just can't get running on linux (games mostly, my stress relief), but also this is a learning experience for me. I'm in no rush to install windows. If I could play all my games I like on Linux, I'd never go back to windows. It's been a while since I installed Windoze from scratch, especially installing it *after* Linux. Generally it's best to install Windoze first. With that said, the one thing I do remember is that Windoze will overwrite the MBR. This means lilo (or grub) will be overwritten. That means you need to have a boot floppy handy to boot into Linux. So, once you have Windoze installed, pop the boot floppy in and boot Linux. Then modify LILO to boot Linux or Windoze and then rerun lilo, putting the boot record on your MBR. That should do it, assuming Windoze install hasn't changed since the last time I did this. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] Love you guys
deFreese, Barry [EMAIL PROTECTED] writes: Sorry for the goofy OT post but I really wanted to put Debian on a Compaq Proliant 1600 and not have to put RedHat on it. I couldn't get it to see the array controller and being the newbie moron that I am I had no idea how to get it to work. A quick google pointed out to me that Disk2 has the array module on it and BAM I am now installing it!! You guys ROCK!! Wait until it's installed and running before letting the compliments fly. Especially with a Compaq! :) Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] Love you guys
deFreese, Barry [EMAIL PROTECTED] writes: What do you mean especailly with a Compaq? I personally think that the Proliant servers are some of the best PC-based servers out there. My experience with Compaq is colored by experiences I had with them way back in the early 90's (92 or 93 I think). They used to play all sorts of tricks with their hardware to overcome some of the limitations, as they perceived them, in the hardware on the PC. Unfortunately this made them a bear to install Linux on. Similiar to a lot of laptops in more recent times. You know, DSP modems, winmodems, proprietary sound chips, etc. Things along those lines. Perhaps they've since moved away from that practice and I'm being unfair to the modern Compaq. But I haven't bought a Compaq since then because of all those problems. Of course Linux 0.99 wasn't exactly the picture of perfection that it is today either! ;) Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
CUPS Samba print server under Debian
I'm in the process of setting up a Linux server for my Home LAN and have some questions regarding the setup. First question is in regards to CUPS. I see that under cupsomatic-ppd it says that you should try foomatic-bin and foomatic-db first. When I install the foomatic-* packages I only see a couple of printers that seem appropriate for my HP DJ932C. Namely: HP DeskJet Series CUPS v1.1 (en) HP New DeskJet Series CUPS v1.1 (en) On the other hand, if I install cupsomatic-ppd I see 5 drivers specific to the 932C. Am I losing anything by using what appear to be the generic drivers in foomatic over the cupsomatic-ppd drivers that seem to be specific to my printer? If I stick with foomatic which one of the above do I use? Just try 'em and see which works best? If I go with cupsomatic-ppd the same question applies, but increases by a factor of 2.5 because now I'll have to choose from among 5 drivers instead of 2. The next question relates to using this printer via Samba from a Windoze client. I understand the Windoze client needs the native Windoze driver for my printer installed, which isn't a problem since that's what the printer is currently attached to. The question is will I lose any capabilities by attaching it to the Linux box running Samba? I use the printer to, among other things, print near photo-quality pics and I know on the Windoze side that usually involves some special driver tricks (in the HP's case I think it's called Photo ret, or something similar). Will I lose that capability by attaching the printer to the Linux box, or does it basically send raw printer command over the Samba link and effectively bypass the Linux driver, which may not have all the capability of the Windoze driver? I imagine I can just hook it up and try all the combos, but I thought maybe someone here had already fought that fight and I could, once again, rely on another's experience. Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: bind8 vs bind9
[EMAIL PROTECTED] writes: Tim Dijkstra [EMAIL PROTECTED] [2002-11-18 17:14:38 +0100]: Is there any reason to stick with bind8 other then convenience? I'm asking this because bind9 seems pretty mature, but the default bind is still bind8 I think... What is convenient about bind8? That is a serious question and not meant to be snide. Please educate me. With the many security advisories that have been posted against bind8 recently I find bind9 much more convenient since I have not needed to update those servers as often. Perhaps you did not know that bind9 uses the exact same configuration file as bind8 in almost all cases? Unless you are doing something strange you don't need to change your configuration files at all. Hmm. Searching CERT: bind4: 10 results bind8: 19 results bind9: 21 results tinydns+dnscache: 2 results This was with only Advisories and Vulnerability Notes checked and, admittedly, I didn't read the results in any detail. I personally use tinydns+dnscache, but that's for a small home LAN and I didn't have too much invested in BIND when I made the switch, but certainly these results seem to indicate that if security is *the* major concern in a DNS installation it's probably a good idea to stay away from BIND altogether. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Using Lilo to select different Kernels
Doug MacFarlane [EMAIL PROTECTED] writes: [snip] I'm in the process of learning to compile my own kernels, and since I do everything the Debian way, i'll be a kpkg guy and build .debs of my custom kernels, which will install in identical fashion, as the prepackaged kernels. make-kpkg is a *very* nice piece of software. Makes maintaining your own kernels relatively painless. That said, as near as I can tell from looking in /boot, when one installe a kernel in this fashion, everything is still there from the previous kernel except the symlink from vmlinuz to the kernel or initrd image, so . . . Actually, I think the default behavior for Debian is to set up /etc/lilo.conf to keep two kernels in the boot menu. The current one you just installed and the old one. This is done using the links /vmlinuz and /vmlinuz.old. So if you were running 2.2.20 (from a default Debian install) and then installed a prepackaged kernel-image-2.4.19 the new kernel package would likely have reset /vmlinuz.old to point to /boot/vmlinuz-2.2.20 and set /vmlinuz to point to /boot/vmlinuz-2.4.19 and if you let the package script run lilo for you, and assuming you have a default /etc/lilo.conf file, you can boot either of those two kernels from the LILO prompt. For a new install, of stable, I think your choices are Linux and LinuxOLD. Of course you can hit TAB (I think) at the LILO prompt and it'll give you the list of possibilities. Is it possible to have multiple LILO entries to boot different kernels off of the same /boot partition? My replay above should give you a hint. Yes, you can have as many kernels, of different kernel versions (like 2.2.20, 2.4.18, 2.4.19, etc.), as you want. If you want different kernels of the same kernel version, i.e., you want to install multiple copies of the 2.4.19 kernel, you'll have to look at kernel flavors (--append_to_version in newer make-kpkg speak) I believe. Otherwise it will install over any previous kernel of the same kernel version. If you want to have more than two kernels in LILO you'll probably have to set that up yourself by editing /etc/lilo.conf. It'd be easy to do by just copying the already existing Linux and LinuxOLD sections and replacing the appropriate fields to coincide with your kernel, including specifying the kernel image in /boot. And as a followup, what's the difference between booting a kernel, and booting an initrd image? Why have the prepackaged kernels moved to initrd? Why would I want to build my custom kernels to boot as initrd images? Why wouldn't I? initrd allows pretty much any and all drivers to be built as modules. The bare-bones kernel is then booted, via initrd, and it loads the modules necessary to continue booting your computer, for example your disk controller (IDE, SCSI). This allows you to build a single generic kernel that will boot on pretty much any hardware supported by Linux. In contrast, distributions used to have to build kernels for different drivers. You'd have kernels like kernel-image-2.2.x-idepci, kernel-image-2.2.x-scsi, etc., so that you could install on a specific computer with the specified hardware. They had to do this because building all the necessary drivers into the kernel, so that you could boot on any conceivable hardware, made the kernel too big to fit in memory at boot time. I, personally, don't like using initrd for my own custom kernels. Just adds another layer to the boot process that I'd rather avoid. Since you, in theory, know what you need to compile into your own custom kernel in order to boot your machine you shouldn't need to use initrd. Now if you were maintaining a custom kernel for a large variety of machines it might be worth it so that you could avoid compiling custom kernels for each machine, but that's the only reason I am aware of for using initrd for custom kernels. All of this is to the best of my understanding, and I'm hardly an expert on initrd since I don't use it, so take it with a grain of salt. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: bind8 vs bind9
Kirk Strauser [EMAIL PROTECTED] writes: At 2002-11-18T18:12:13Z, Gary Hennigan [EMAIL PROTECTED] writes: ...if security is *the* major concern in a DNS installation it's probably a good idea to stay away from BIND altogether. I'd disagree for one main reason: BIND is Open Source, and tinydns is not. More security compromises have been found in the former, but it has perhaps 100 times the number of users testing it and programmers examining it. Frankly, as a programmer, I would not waste my time auditing tinydns when I could make a bigger contribution to BIND. It's not Open Source perhaps by some formal definition, but you can certainly look at the source code and publish patches and, according to Bernstein, modify your own copy. So if you're hard-core Open Source then you will probably want to avoid all the controversy surrouding Bernsteins licensing. But, he does stand behind the security of his software with cash. Granted, not a lot of cash, but it's a refreshing change from the big companies. Even open source companies that put out sendmail and bind don't do that. Also, getting away from the the licensing controversy, as someone who didn't cut his teeth on BIND, I find the configuration of tinydns much easier. Those BIND config files are not easily understandable and the fact that losing a trailing period (.) can cause the whole thing to come crumbling down was a bit frustrating. Please remember too that I'm coming from the perspective of a small home LAN sysadmin that doesn't have to do things like zone transfers and worrying about root servers. I just have a handful of computers, all locked behind an OpenBSD firewall, using 192.168.0.0/16 addresses, that I want a DNS server for. tinydns suited me much better than BIND. As an aside I also tried qmail. Uggh, what a nightmare to configure! All those small config files hanging out in weird places! After about 5 hours I surrendered and reinstalled Exim. Maybe it's not as secure but I had it up and running in less than an hour with Anomy and bogofilter for my families email. Just my opinion. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: bind8 vs bind9
Nathan E Norman [EMAIL PROTECTED] writes: On Mon, Nov 18, 2002 at 04:28:13PM -0600, Kirk Strauser wrote: At 2002-11-18T21:11:23Z, Nathan E Norman [EMAIL PROTECTED] writes: Uh, of course tinydns (sic; it's really djbdns) is open source. Perhaps you meant to say _DFSG Free_ ? No more than Microsoft's shared source is open source. I can't redistribute either of them, regardless of what bugs I patch or features I add. You can distribute patches to djdns all day long. You can even distribute a patched version if you don't call it djbdns. What you can't do (and this is why djb stuff fails the DFSG) is distribute a pre-compiled version with your patches applied. Comparing DJB to MS is disingenuous and you know it. Oh well, a little demagoguery seem the status quo on debian lists. Just curious, on which list(s) isn't a little dg. status quo? ;) Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Cheap SCSI controller?
Can anyone recommend a cheap SCSI controller? I'm putting together a little network server for my home LAN and want to attach a SCSI DDS2 DAT drive I have for backups. It's an older DDS2 drive so I think just about any PCI controller will be able to keep it's channel filled, and that's the only device that will be on the SCSI bus. I see Adaptec has just about driven everyone else out of business, but I do see some Tekram cards pretty cheap. I see one place has 'em for $16!! Now that's what I'm looking for! Anyone have any experience with a Tekram DC-315 SCSI controller? It's BIOS-less, so no booting but I don't need that anyway. Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: X Windows and Gnome Problem - HELP!!
Craig Dickson [EMAIL PROTECTED] writes: Tim Woodward wrote: I have installed Debian 2.2r4 on my system and have so far got to the point where X Windows is running with the Gnome version that comes with the Debian pack. The problem i have now is that i have changed my graphics card and installed the XF86_SVGA server, and when i boot up it goes straight into X Windows but the mouse is disabled. The screen is in 640x480 so looks crap, and i can't exit back to the original prompt screen. If i knew how to bypass X Windows upon boot-up i could re-run the XF86Config which may solve the problem. Can anyone help me sort it? Also, (i am new to all this), i've downloaded KDE from the Debian website, and would like to know: Is it just a window manager or an entirely independant X Windows type program? For what it's worth, Debian 3.0 has been out for several months now. You might want to consider upgrading once you get your current problems worked out. If you want the machine to boot up to console mode, rather than entering X, you could disable or uninstall whatever graphical login manager you're using. Or you could just hit ctrl-alt-backspace until X gets the idea that you really don't want it to run (c-a-bs exits X, but usually if you're running a graphical login manager, it will automatically restart X when you do that. But if you do it several times, usually the login manager will get the hint). [snip] Probably the best way to do this is to go to a console. There are 6 text-mode consoles available, by default. From within X you get to them by hitting the key sequence: Ctrl+Meta+F[1-6] The Meta key is either Alt or that Windoze key next to it on most newer keyboards. F[1-6] means you can go to one of six consoles by using the F1,F2,F3,...,F6 keys. Once at a console you can log in as root and reconfigure X. So just hit: Ctrl+Meta+F1 And you'll go to a text-based console. BTW, that means you press those three keys simultaneously, just in case you're not familiar with the notation. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[Semi-OT]: Dual Head Laptop?
I've got a buddy that wants to switch to using just a laptop for all his needs, instead of the laptop/workstation combo he currently uses. One thing he's decided he doesn't want to do without though is a dual-headed setup when he's in the office. My initial thought was that surely someone made a docking station with an AGP slot and he'd just tell his laptop to use the video card in the DS and disable the adapter in the laptop. My Dell laptop seems to be able to do the disable on-board VGA trick. Unforunately it doesn't appear to me that AGP is a feature in any of the big-boys (IBM, Dell, etc.) docking stations. Now my thought is he'd be stuck with a PCI display card, like the Matrox G450 PCI. In theory I think that would be fine, but theory and practice often diverge, especially where linux and laptops and X are concerned. Is anyone using a setup like this? I know the G400 AGP works fine in a workstation using X in Xinerama, which is my current setup, but I'd like to be more confident it'd work with a G450 PCI in a docking-station application before I recommend. Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [Semi-OT]: Dual Head Laptop?
nate [EMAIL PROTECTED] writes: Gary Hennigan said: I know the G400 AGP works fine in a workstation using X in Xinerama, which is my current setup, but I'd like to be more confident it'd work with a G450 PCI in a docking-station application before I recommend. how about getting a laptop with a good amount of video memory(I don't know whats standard these days, my T20 has 8MB i think) in combo with Accelerated X LX : http://www.xinside.com/Pages/LX/LX%20ProductPage.html price is $149 and you get(in general) top notch drivers, good quality software. it supports running the LCD and an external monitor at the same time on some laptops. demo is freely available. They mention a Zoom mode which sounds like some sort of display spanning, but I have never heard of the term, reccomend you ask them if your interested. It might work more transparently too, not sure. Hmm. I've seen this capability mentioned, and it might be interesting, but in this case he doesn't want to use the small LCD on the laptop and a nice 20 external LCD simultaneously. He wants to use 2 nice 20 external LCDs when he's in the office. And the internal laptop LCD when he's on the road. But I'll certainly look at the commercial X servers and see what they say! Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Calendars
Michael Montagne [EMAIL PROTECTED] writes: I need a good calendar. Evolution is OK, but all I want is a calendar, amd I want it to be speedy. Gnomecal is my current chice but I can't figure out how to add holidays and other events that aren't appointments. I also like the command line interface to gnomecal that allows me to query the calendar very quickly. But it doesn't work from cron. The old unix calendar program offers me what I need from cron but it is too hard to get information from when looking for free time to schedule an appointment. What is the popular choice out there? I use plan. I don't use the command-line varation, but it claims to have that capability. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Ughh. What happened to /bin/echo?
It appears that /bin/echo is gone. In stable it was contained in shellutils and shellutils was replaced by coreutils, which doesn't include /bin/echo. I realize echo is a builtin in Bash, but my Makefiles don't seem to realize this and are now completely broken. Did echo move to a different package, hopefully? apt-cache search doesn't pull up any obvious candidates, except coreutils and it's definately not in there. I'm running testing. coreutils 4.5.1-2 Something similar seems to have happened to /bin/cat and /usr/bin/tr in what was textutils under stable. Anything on this? Thanks, Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: error upgrading woody to sid (passwd)
Sebastian Schinzel [EMAIL PROTECTED] writes: Hi! I have trouble upgrading from woody to sid: The error after #apt-get dist-upgrade is: Preparing to replace passwd 2902-12 (using .../passwd_1%3a4.0.3-3_i386.deb) ... Unpacking replacement passwd ... dpkg: error processing /var/cache/apt/archives/passwd_1%3a4.0.3-3_i386.deb (--un pack): trying to overwrite `/usr/share/man/de/man1/chsh.1.gz', which is also in packag e manpages-de dpkg-deb: subprocess paste killed by signal (Broken pipe) Ok. Finally it worked by removing manpages-de, but thats not very elegant. This bug seems to be reported at: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=162097repeatmerged=yes Its from 09/23 and for me it sounds that they fixed it. Does anyone know a better solution for this? Shall I write a new bugreport about this (never done this before)? I personally just download the offending *.deb file(s) manually and do, for example: dpkg -i --force-overwrite passwd_1%3a4.0.3-3_i386.deb I just last week did a woody-sarge upgrade and had to do this a couple of times. Nothing seems broken so far as a result. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Canonical Way to install Java
Kent West [EMAIL PROTECTED] writes: I can get Java installed and working in my browsers, but at the risk of starting a Holy Way, what's the canonical way to install Java (run-time only needed, not dev. kit)? From Blackdown? From Sun? From Debian's site, which seems to only have JDK1.1 for Sid? I would suspect that Sun's JRE is *the* JRE, whereas Blackdown's might work better with Linux, whereas Debian's is more official. I haven't seen anyone mention IBM's offering (http://www.ibm.com/java). I know at one point the IBM runtime was 20% or so faster on a lot of benchmarks than the other offerings, that was some time ago though. I haven't used it in a while, but it seemed decent to me. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How stable is dual head, twin view feature ?
tvn1981 [EMAIL PROTECTED] writes: I have a Woody system (kerne 2.4.18) with XFree86 Version 4.1.0.1. I also have a Radeon 7000ve with twinview, and an Nvidia gforce2 quadro with twinview. However, I just can't get dual monitor run well on this system. For the Radeon card, the twinview feature doesn't work at all, I was told that it doesn't even work w/ this Radeon card even with XFree 4.2.x. The Nvidia card, using the latest nvidia's driver, works! But the problem is X freezes so often, the whole computer just froze, I have to physically reboot. I like the dual monitor setting so really want to know if there's any setting that make this work on my system. So just wondering if it's just me that has problems with twinview stuff ? Or is there any other settings that you would recommend - how well is Matrox G400 or Agp+PCI settings ? they run well ? I have a Matrox G450 card and it has been absolutely rock solid with Xinerama since I got the thing (currently running X 4.1.0.1). In the 8-12 months I've had the system I think I remember, maybe, one X crash. And who knows if that was the applications fault or the driver? With Matrox you do lose DRI if you enable Xinerama, so there is that drawback if DRI is important to you. Gary P.S. Please wrap your lines at about 72 characters. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Time(zone) problems
Mike Mimic [EMAIL PROTECTED] writes: Hmm. The only other thing I can think of is the setting in /etc/login.defs. ENV_TZ is typically commented out though. If it's not you might try commenting it out, although I would think root would be affected if that variable were set. The root account doesn't have any TZ environment variables set, right? No, there isn't set any variable. And I have set the ENV_TZ but it doesn't work (should I restart anything after changing login.defs?). But if I set TZ variable manualy date reports correct time. The other thing to try is to create a new user account, using adduser, and see if that account has the same problem. That might narrow the possibilities somewhat. I have asked other users and they have same problem. Hm, so I still don't know what's wrong but TZ varible solves the problem. But is there anything else to try because I would like to find out what is wrong. I'm sure there's more you could try, since something is obviously wrong, I'm running low on ideas though. All the time info comes from libc6. The only other thing I can think of that could be relevant: % ls -l /usr/share/zoneinfo/localtime lrwxrwxrwx1 root root 14 May 16 11:16 /usr/share/zoneinfo/localtime - /etc/localtime % ls -l /etc/localtime lrwxrwxrwx1 root root 27 May 16 11:16 /etc/localtime - /usr/share/zoneinfo/MST7MDT % ls -l /usr/share/zoneinfo/MST7MDT -rw-r--r--6 root root 877 Apr 28 03:56 /usr/share/zoneinfo/MST7MDT Make sure the owner/permissions on the two links and the file are as given above, and that /etc/localtime points to the correct zoneinfo file for your time zone (mine is MST7MDT). Also, make sure you as a user can read those files. As a user try: % cat /etc/localtime /dev/null You should *NOT* get a Permission denied message. If you do that'd be your problem. I'm not sure how those zoneinfo files get generated but maybe yours got corrupted, if all of the above is ok. You could try, in order: 1) dpkg-reconfigure --priority=low libc6 2) Extracting the appropriate zoneinfo file from the libc6 package and overwriting your existing file 3) A full-blown reinstallation of libc6, e.g., get the libc6*.deb file that you have installed and do dpkg -i libc6_2.2.5-6_i386.deb (that's the version of libc6 on my system. It's location on the debian mirrors is: /debian/pool/main/g/glibc/libc6_2.2.5-6_i386.deb If none of that works I'm afraid I've run out of ideas. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: RAM upgrade == Kernel Panic
Jamin W.Collins [EMAIL PROTECTED] writes: On Wed, 12 Jun 2002 16:51:06 -0400 Tom Allison [EMAIL PROTECTED] wrote: I got a really interesting one. I am running a stock K7 kernel from woody. I recall seeing such things as cramfs, VFS, and initrd during the boot sequence. I upgraded my RAM to replace the current 128M to 768M Now, when I boot it panics with a message that VFS can't mount the root device at 03:02. If I return to the original RAM, it's fine. Have you tested the new memory? Memtest86 is a great tool for doing just this. And there's a Debian package for it, memtest86! I have to agree with Jamin here. While it's not impossible that it's a software issue, it's much more likely to be a problem with your memory or with one of the memory sockets on your motherboard. memtest86 is a stand-alone memory tester (you boot into it) and it does a thorough test that should point out a hardware problem. If memtest86 shows an error you can try inserting the memory sticks one at a time and if each one individually checks out it's probably a problem with your motherboards memory sockets. The BIOS memory test is practically worthless, BTW. It won't catch 90% of memory errors. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Time(zone) problems
Mike Mimic [EMAIL PROTECTED] writes: I have some strange time problems. When I run date as normal user I get time in UTC timezone. And when I create files they also get timestamp with that time (and that's wrong time). But as root I don't have that problems and I have normal CEST time (as it should be). I have runed timezoneconf but it didn't solve the problem. Any ideas? Mike, Do you have any environment variables set that relate to the timezone? Try: % env|grep TZ as both yourself and as root. Sounds like you have a mismatch here somewhere. In particular the TZ environment variable can be set per-user, but generally that functionality isn't used. Also, make sure the file /etc/timezone contains the correct setting for your system and make sure you have the proper setting for UTC in /etc/default/rcS so that it matches what your BIOS clock is set to (either local, UTC=no, or UTC=yes if your BIOS clock is set to UTC time). You can use the hwclock utility to find this out. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Time(zone) problems
Mike Mimic [EMAIL PROTECTED] writes: Do you have any environment variables set that relate to the timezone? No. I haven't any. Also, make sure the file /etc/timezone contains the correct setting for your system Yes, it have. And I have tryed some different cities (in the same timzone) and it's the same. and make sure you have the proper setting for UTC in /etc/default/rcS so that it matches what your BIOS clock is set to I have looked it and it's correct. Is there anything else what should I check? I really don't know what's wrong. Is it possible that some permissions on files (devices) aren't good and so there is a problem (I have checked /etc/timezone)? Hmm. The only other thing I can think of is the setting in /etc/login.defs. ENV_TZ is typically commented out though. If it's not you might try commenting it out, although I would think root would be affected if that variable were set. The root account doesn't have any TZ environment variables set, right? The other thing to try is to create a new user account, using adduser, and see if that account has the same problem. That might narrow the possibilities somewhat. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Palm Pilot
Stephen A. Witt [EMAIL PROTECTED] writes: I'm running woody and have just been handed a Palm Pilot at work. I'm not really sure I want the thing, but thought I would check it out. I was going to install a syncing application with my woody desktop, but found there were several palm pilot apps in the distribution. Is a question like 'which is the best one to use?' relevant? Anyone using Palm's that could provide some advice? I use PilotManager. Mostly because that's what I've used for a long time and it syncs with plan, my calendar of choice, and GNU BBDB, my database of choice (or habit anyway), for storing mail addresses, phone numbers, email addresses, etc. If I were starting from scratch I'd probably look at using gnome-pilot for all this information, or perhaps Evolution. I also use malsync to sync up with my AvantGo (http://www.avantgo.com) account. Avantgo essentially creates, or has companies create, web pages that are Pilot-friendly. Makes some of my meeting bearable! :) Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SMP working
Robert Webb [EMAIL PROTECTED] writes: I just loaded up Woody on a new machine that has two 500MHZ Pentium III processors. I am trying to determine if the build I loaded is recognizing both processors. My problem is I have no clue as to where to check for this . I looked in the /proc directory and did a cat of cpu but it only showed one processor at 500MHZ. Am I looking in the wrong place??? You can use the cpuid module (modprobe cpuid) and x86info to tell you this. Of course x86info simply reads /proc/cpuinfo so you could just cat that. I don't have a /proc/cpu on my system. You can also look at the output during boot and check there. Like: % dmesg | grep CPU . CPU clock speed is 1694.7612 MHz. CPU0T0:996912,T1:664592,D:14,S:332306,C:996918 CPU1T0:996912,T1:332288,D:12,S:332306,C:996918 checking TSC synchronization across CPUs: passed. I have a two-CPU system. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ftp.kernel.org
ben [EMAIL PROTECTED] writes: anyone know what's happening at kernel.org? ftp connections are being rejected. i also can't read the mirrors list. is there an alternate site to ftp kernel sources? Well, there's the debian source packages, e.g., kernel-source-2.4.18. You can also try the country-specific kernel sites, e.g., ftp.us.kernel.org Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: What do I need for email virus checking on debian server?
Motiv8d [EMAIL PROTECTED] writes: Hi Further to my last post. I would also like to be able to virus check email attachments, change extensions on .js etc What would be the suitable antivirus options? and Email package or addon that allows rule based modification of attachments and does the other things my last post mentioned. For a decent free solution you can do what I do on my OpenBSD firewall at home and use the E-mail Sanitizer http://www.impsec.org/email-tools/procmail-security.html. It can't scan binaries for viruses, but it can reject binaries altogether, or rename them so the users can't just click on them to execute such attachments. I have no clue if this particular scanner is packaged under Debian but I see Anomy is, under Debian package sanitizer. It does much the same thing and I understand the author of the procmail sanitizer is going to work on Anomy once he gets a stable 2.x release out. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Setting system time on startup
Ian D. Stewart [EMAIL PROTECTED] writes: Right now, when booting my linux box, the system clock is off by four hours (I'm guessing it is set to GMT). I can reset the clock using date, but this is starting to get a bit tedious. Is there a way to automate this process and/or convince Linux to set the system clock to local time? You can change that by setting UTC=yes|no in /etc/default/rcS. This setting should be yes if your HW clock is set to UTC (aka GMT) or no if your HW clock is set to local time. So first check to see what your HW clock is set to by using the command hwclock --show. You can then use hwclock to either set your system time from your HW clock, or vice versa. This is done automatically at boot by the hwclock* scripts in /etc/init.d The main thing is setting UTC to the appropriate yes or no in /etc/default/rcS, assuming, of course, that the problem is the fact that your HW clock setting doesn't match the UTC setting. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: group equals username.. why?
irado furioso com tudo [EMAIL PROTECTED] writes: I noted that when creating a user, it is assumed a group name with the very same username (user irado, group irado). Is there a way to select a generic (say: users) group when creating new users? Please read the man page for adduser and the configuration file /etc/adduser.conf. All of your questions are answered therein. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Dual Boot with XP
Keith O'Connell [EMAIL PROTECTED] writes: Steve, Better yet, add this to /etc/lilo.conf: other=/dev/hda1 label=WinXP This works splendidly for me. YMMV. I am surprised - I was under the impression that XP was hostile to Linux and you had to be sneaky to get it to work. So, am I right then; I don't have experience with WinXP, but previous windoze versions weren't hostile to Linux. Their installation programs on the other hand had a tendency to just blindly repartition your disks without looking to see if something else (other than another Win OS) was installed on them. So they had a bad tendency to obliterate non-Windows partitions. Once installed though I've never had any trouble, even when doing Windoze upgrades from, for example, 98 - 2k. 1: install XP to hda1 2: install Debian to hdb1-7 3: install lilo to hda 4: edit lilo.conf as above 5: run lilo 6: reboot and select as appropriate Is the right things in the right order? The sequence looks good to me. It's been a while but that's what I did (with 98 and 2k). Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: New Debian user - a bunch of problems
Brian [EMAIL PROTECTED] writes: 2). The mouse is going nuts. When I move it around it does all kinds of weird things, including acting like I've pressed one of the 3 buttons when I haven't, or sen't kill commands to fvwm, etc. I have a Logitech M-C48 wheel mouse. If any of you have the console and X settings for it, I'd appreciate it. I had a similar problem. Do you have gpm running? ( I assume you know Ctrl-Alt-Fi takes you to the ith console to get a shell when X isn't working. X is F7.) Kill gpm if it is running and hopefully your mouse will start working. Brian And then, if you're not going to do much work on the console, i.e., outside of X, do: apt-get --purge remove gpm gpm is a package to allow you to use the mouse on the console. If you think you'll be using console mode equally as much as X you'll have to look at getting X to play nice with gpm. Something about using /dev/gpmdata as the mouse device in X, but that's just off the top of my head so look up the relevant info. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Problems: mutt to Outlook (Express)
Cam Ellison [EMAIL PROTECTED] writes: I guess alittle more information is in order. Some of this is repeated elsewhere in this thread. I have a cable hookup, and am using exim (3.35) as my MTA. Messages I send from mutt go to just about everyone, but there are at least three exceptions, all using an ISP other than my own. Case 1: Correspondent uses Outlook Express 5.00.2615.200 on a dialup connection to a telco. Case 2: Correspondent uses the same version on a dialup connection to an independent ISP. Case 3: Correspondent uses Outlook 2000 (9.0.2416) and has DSL (same telco as above). Case 4: Correspondent uses OE for the Mac (5.02.2022), receives the mail via a 24/7 link to a mail company. This one always works. The mail for the first three never arrives. I do not know whether it is left on the server, or goes to /dev/null. A number of possibilities suggest themselves, but the most likely to my mind is that Outlook thinks it is malformed and assigns a 5xx code, thus killing it. Very strange, and it does nothing to endear me to M$. Personally I don't think the trouble is on your end. You'll more than likely have to bring their ISPs into this to solve the problem. One other thing they can try is to use Netscape mail and see if they can access the mail. And, *please*, for the love of God and country, can you wrap your lines at 70 characters or so?! Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Is there any tool like purify(memory leak check tool) for LI
Paul Smith [EMAIL PROTECTED] writes: We'd buy it at work in a heartbeat if it were available on Linux: Purify _is_ cool. But, it's not available on Linux and I've not heard Rational provide any details on whether it ever will be :(. Insure++ (http://www.parasoft.com) has almost all the same features as purify (Parasoft, of course, claims they have more features) and is available for Linux. I'm sure it's not cheap, but we have a site license so I don't know the cost/seat. It works, but does seem to add even more overhead to the executable than Purify (execution time, memory footprint, etc.), although I've only run purify on IRIX and have never run Insure++ on IRIX and so I can't do an apples-to-apples comparison. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: recommended documentation iptables
justin cunningham [EMAIL PROTECTED] writes: Hey, I did I search for iptables how to and got tons of docs-- does anyone recommend one over another or a debian specific one? The goal is to take advantage of three nics for web and mail services. I'm reading this one now http://www.linuxsecurity.com/resource_files/firewalls/IPTables-Tutorial/ iptables-tutorial/iptables-tutorial.html thanks, justin I would assume the place to start would be the Debian documentation: file:/usr/share/doc/iptables/html/packet-filtering-HOWTO.html file:/usr/share/doc/iptables/html/NAT-HOWTO.html These are in the Debian iptables package (on sid and woody anyway). Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Woody: apt mysteriously wants to pull in 100 MB from sid. Why?
Mario Vukelic [EMAIL PROTECTED] writes: On Tue, 2002-05-07 at 19:56, dman wrote: What if you use testing and unstable instead of woody and sid in your preferences file? There was some discussion on that recently, and some people said using names doesn't work. I've only tried with the description and haven't had problems. Works with testing and unstable (both in sources.list and preferences). Thanks. Any idea why this is? Are you using apt's release capability to determine which version to use? In my sources.list I use woody and sid, but I had to use APT::Default-Release testing; in my apt.conf file, otherwise I saw the same behavior as you, apt trying to pull in a bunch of sid stuff. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Anyone using ulogd?
Jonathan Matthews [EMAIL PROTECTED] writes: On Wed, May 01, 2002 at 04:30:24PM -0600, Gary Hennigan wrote: [snip] I also did something I wasn't sure was/is necessary. I recompiled my 2.4.18 kernel with CONFIG_NETLINK_DEV=y it's in the networking options if you're using xconfig or menuconfig. After booting that kernel I [snip] Silly question maybe - did you compile ULOG target support into the kernel? Do you modprobe/insmod it if it's a module? Just a thought ... That's an important point to make for others. You need this! Of course I always compile all the iptables options as modules so I knew that wasn't my problem. Indeed one of the first things I checked when it wasn't working was that ipt_ULOG was there when doing lsmod. My problem, I'm convinced, is something in the patch that the package developer for ulogd did, because the virgin copy works just fine. I've spent enough time on it at this point that I'm not going to look to find specifics. I'll probably add my info to the bug filed against ulogd though. Gary -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
