Re: checking for changes in file size/permissions since installation?
Carl, On Feb 13, 2011, at 2:42 PM, Carl Johnson ca...@peak.org wrote: You might have been thinking of debsums. That will check the MD5 sums of installed packages, but I don't know whether it covers all installed files. The checksums are stored in /var/lib/dpkg/info, so they might also be corrupt. This looks like what I had used initially - at the very least I can apt-get --reinstall the packages which show problems. Looks like the md5s are not corrupted. Thank you! -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/10528c70-5953-4d3b-931d-5eebaf298...@sonous.com
Re: checking for changes in file size/permissions since installation?
Brian, -lev On Feb 13, 2011, at 3:09 AM, Brian a...@cityscape.co.uk wrote: On Sat 12 Feb 2011 at 23:17:12 -0800, Lev Lvovsky wrote: Does apt/dpkg keep track of permissions and file sizes of the files which belong to a package? If so, how can this information be retrieved so as to compare to existing files on the file system? Is dpkg -c package.deb of any use? Sorry, I should have been more specific - I need to audit all of the installed packages on my system (massive file system corruption). I'm assuming that dpkg knows the size and permissions (not as important) of a file after it's been installed, so need to compare the existing file to what it was originally when installed. I'd found a command-line piped together a week or so ago that does this, but I can't for the life of me find it again. Thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/371f5150-5786-4f97-b26c-7f004dffc...@sonous.com
checking for changes in file size/permissions since installation?
Does apt/dpkg keep track of permissions and file sizes of the files which belong to a package? If so, how can this information be retrieved so as to compare to existing files on the file system? thanks! -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1e383eb5-5595-49b4-8aae-c2cb56d4c...@sonous.com
grub2 not displaying boot menu on serial redirect
Hello, I run a headless box in conjunction with serial redirect. I posted about a year ago in regards to grub2 not displaying the boot menu to the serial output: http://www.mail-archive.com/debian-user@lists.debian.org/msg565218.html Since that post, I'd just waited for upstream to fix the bug, and didn't think much of it. Grub2 had since then been upgraded, and somehow my problem went away. Yesterday I ran update-grub, and the problem has resurfaced. The menu appears on video out, but no longer to the serial. The boot process redirects to serial just fine (at which point grub hands over control the the kernel iirc). any advice would be appreciated - below is a snippet of my grub.cfg: ### BEGIN /etc/grub.d/00_header ### set default=0 set timeout=20 serial --unit=0 --speed=9600 --word=8 --parity=no --stop=1 terminal serial ### END /etc/grub.d/00_header ### ### BEGIN /etc/grub.d/10_linux ### menuentry Debian GNU/Linux, linux 2.6.26-2-686.21 { set root=(hd0,6) search --fs-uuid --set 7abda901-2a56-4917-96af-c6dc983d015c linux /vmlinuz-2.6.26-2-686.21 root=UUID=25ed25b7-64bb-4949-ba5e-932093b1fcef ro console=tty0 console=ttyS0,9600 initrd /initrd.img-2.6.26-2-686.21 } thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/57cadb85-3121-4c06-823b-ed8f18743...@sonous.com
Re: grub2 + serial redirect
Andrew, On Feb 5, 2010, at 7:21 AM, Andrew Malcolmson wrote: On Thu, Feb 4, 2010 at 12:44 PM, Lev Lvovsky lis...@sonous.com wrote: Hello, [snip. This works fine once the kernel is booted (it redirects output to my serial port). I am not able to see the boot menu with this new configuration however - it appears on the video portion of the screen, and I'm able to select a kernel that way, but the serial ouput gets nothing. With Grub v.1 on other machines with serial redirect, I've been able to get the menu. Is there some option that I'm missing? I was working on this a few weeks ago and AFAICK I needed the following in grub.cfg to serial the grub menu over serial: terminal_input serial terminal_output serial Thanks, I'll give this a shot tonight, if it works, I'm assuming that this is an upstream patch to the update-grub script? -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
grub2 + serial redirect
Hello, So after 12+ years of using lilo on my server, I've decided that it might be time to upgrade my boot loader. With lilo I'm using the standard serial redirect since I'm running a headless box (more specifically it's a guest of a VirtualBox host). With Grub2, I've been able to get the standard serial redirect working with the following in my /etc/default/grub: --- GRUB_CMDLINE_LINUX=console=tty0 console=ttyS0,9600 GRUB_TERMINAL=serial GRUB_SERIAL_COMMAND=serial --unit=0 --speed=9600 --word=8 --parity=no --stop=1 --- This works fine once the kernel is booted (it redirects output to my serial port). I am not able to see the boot menu with this new configuration however - it appears on the video portion of the screen, and I'm able to select a kernel that way, but the serial ouput gets nothing. With Grub v.1 on other machines with serial redirect, I've been able to get the menu. Is there some option that I'm missing? Additionally, IIRC, the GRUB_CMDLINE_LINUX option above is conditional, meaning that if I want video vs. serial redirection while booting, I need to swap the two 'console' directives. Given that there's only one way I can specify the linux command line option, how would I go about doing this? thanks! -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: grub2 + serial redirect
Hi Tom, On Feb 4, 2010, at 12:14 PM, Tom H wrote: Here's a WAG: Change GRUB_CMDLINE_LINUX=console=tty0 console=ttyS0,9600 to GRUB_CMDLINE_LINUX=console=ttyS0,9600 and run update-grub and reboot to see if you have the serial output (and still have the console output). I'm not at the machine right now (don't have access to see how the video looks), but this should only change the operation of the booting of the kernel, rather than the availability of the Grub menu. On machines which use grub1, I have the following: terminal --timeout=1 serial console currently, for grub2, I have: terminal serial no timeout value, and no option to pass both serial and console in - apparently grub2 does not support the timeout option for that line. thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: grub2 + serial redirect
Hey Tom, On Feb 4, 2010, at 3:18 PM, Tom H wrote: I'm not at the machine right now (don't have access to see how the video looks), but this should only change the operation of the booting of the kernel, rather than the availability of the Grub menu. On machines which use grub1, I have the following: It was just a WAG... and much appreciated - usually if I write something out like that, it might help me sort out where else to look (http://c2.com/cgi/wiki?RubberDucking) thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: grub2 + serial redirect
Matt, On Feb 4, 2010, at 1:28 PM, Matt Zagrabelny wrote: Note the GRUB_CMDLINE_LINUX line, it includes console=ttyS0,9600. Perhaps it is time to file a bug report against grub-pc? After your explaining how to do this privately, I dug through the existing bug reports, and found this: http://www.mail-archive.com/debian-bugs-d...@lists.debian.org/msg685989.html This looks sort of similar to the related problem, although oddly enough even specifying serial doesn't get the menu on the serial console. I have to see what the video output looks like once booting... thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: concurrent installs of previous + current kernels
Hi Chris, On Feb 1, 2010, at 11:44 PM, Chris Jones wrote: Unless there's some pre or post magic that goes on, these are the same files which are currently owned by the pre-existing (debian release 17) kernel package: This is odd. I keep an up-to-date ubuntu partition on the side, which I boot into every now and then, and every couple of weeks or so, the update manager installs a new version of the kernel. I can't look now but I believe I have something like two or three different versions of 2.6.31 at present. Say, 2.6.31-15, 2.6.31-16, and 2.6.31-17, complete with modules, headers and all. There would be more, if I hadn't removed a few older versions manually to keep my grub menu to somewhat manageable lengths. As Boyd pointed out the versions you're refering to above are all different from one another, and can coexist. However as Stefan mentioned, sub-revisions of the same kernel cannot. In my case, any sub-revision past 17 on the latest stable kernel causes the system to hang on boot. Tracking down the problem will be easier now that I don't have to jump through too many hoops to try later sub-revisions. Not much help, I guess, but maybe worth taking another look at the manuals. Oddly, there's not a whole lot of discussion about this (unless I'm using the wrong search terms). thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
concurrent installs of previous + current kernels
What if any is the generally accepted way of maintaining multiple versions of kernels? I'm currently experiencing boot problems with the latest version of the kernel, and must keep version 2.6.26-17 installed. I'd like to install this, and the latest concurrently so that I can experiment with booting to the newer version, as well as jump back to the older version if I need to. This is also to keep 'apt-get upgrade' from finding anything else pending. thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: concurrent installs of previous + current kernels
Hi Boyd, On Feb 1, 2010, at 1:14 PM, Boyd Stephen Smith Jr. wrote: On Monday 01 February 2010 14:00:07 Lev Lvovsky wrote: What if any is the generally accepted way of maintaining multiple versions of kernels? Just install each of their packages separately. Since the kernel team does support concurrent installs, the upstream version number is part of the package name: I'll admit ignorance on this one - I'm unable to find anything other than just linux-image-2.6.26 and its variants. More specifically, I'm unable to glean any more information than just the major versions of the package as it is to be installed. Upon performing something like 'apt-get install linux-image-2.6.26-2-686', I won't seemingly be able to keep the *currently* installed version of the kernel, and instead, it will simply do an upgrade. This is totally understandable for most package installs, however with a kernel, keeping the previous version installed is useful (obviously). I'm not sure if the procedure would involve only downloading via apt-get, and then running dpkg on the .deb file itself. thanks! -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: concurrent installs of previous + current kernels
Hi Stefan, On Feb 1, 2010, at 2:11 PM, Stefan Monnier wrote: What if any is the generally accepted way of maintaining multiple versions of kernels? Hmm... well, the wayu I do it is: I install multiple kernels. That's all there is to it. Any reason you're wondering about it? Have you tried something and bumped into problems? Are you doing this simply with 'apt-get', or are you using other commands as well? I've only got the stable and security apt repos in my sources.list file, so I'm note sure if I'm opening myself up to all of the versions of the kernel available. thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: concurrent installs of previous + current kernels
Hi Stephen, On Feb 1, 2010, at 2:00 PM, Stephen Powell wrote: One must be careful, though. As an example, consider the following Debian package file names: linux-image-2.6.26-2-686_2.6.26-19_i386.deb linux-image-2.6.26-2-686_2.6.26-19lenny2_i386.deb These package files have different file names, but they are different revisions of the same package. When installed, they will both have the package name linux-image-2.6.26-2-686 And the boot images, initial RAM disk images, and library modules will have the same names in the same directories. The only way I know to install two different revisions of the same stock kernel package name on the same machine at the same time is to have two different boot partitions. If you're building custom kernels you can get around this problem by using something like the --append-to-version flag of make-kpkg. But when using stock kernel images this is not an option. I think this is the problem that I'd be running into. In Red Hat, you get a list of kernel package names which have the distro release version in the name of the package so that you can have multiple packages of the same kernel installed at the same time. I'd be shocked if it were so hard to do on Debian, but I can't for the life of me figure it out... -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: concurrent installs of previous + current kernels
On Feb 1, 2010, at 3:13 PM, Boyd Stephen Smith Jr. wrote: This is totally understandable for most package installs, however with a kernel, keeping the previous version installed is useful (obviously). It's usually not a big deal when the kernel ABI hasn't changed. If you have a strong desire to keep every kernel version around, you'll have to go outside the official packages. From what I understand, the Debian provided tools to build kernel packages can handle this easily, but I've not tried it myself. Well, currently the only desire I have is to keep the version of the kernel which I currently have installed, as well as whatever debian has available as the latest. If I understand, this requires that I build each of these kernels? I'm assuming also, that I could make copies of the binaries dropped into /boot, upgrade, and tell my boot loader to also reference the old files. This of course assumes that modules and such are compatible in between the different versions (which is dubious). There has to be a better way... thanks for your help! -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: concurrent installs of previous + current kernels
Hi Stefan, On Feb 1, 2010, at 3:11 PM, Stefan Monnier wrote: What if any is the generally accepted way of maintaining multiple versions of kernels? Hmm... well, the wayu I do it is: I install multiple kernels. That's all there is to it. Any reason you're wondering about it? Have you tried something and bumped into problems? Are you doing this simply with 'apt-get', or are you using other commands as well? For example, yes. So assuming that I only have stable + security in my apt sources.list config, how would I manage to keep the older version of the kernel package, as well as the newest version? 'apt-get install' will remove the binaries from the previously installed kernel package. thanks, -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: concurrent installs of previous + current kernels
Stefan, On Feb 1, 2010, at 6:09 PM, Stefan Monnier wrote: So assuming that I only have stable + security in my apt sources.list config, how would I manage to keep the older version of the kernel package, as well as the newest version? Huh... you install the new kernel. My thought too, but that doesn't work - the older version is removed. 'apt-get install' will remove the binaries from the previously installed kernel package. What makes you think so? Have you even tried it? Of course I have - otherwise I wouldn't be asking the fine people on this list how to go about this. As an example, the contents of the following linux kernel image deb: linux-image-2.6.26-2-686_2.6.26-21_i386.deb Among other things, is the following structure in /boot: --- % ls -la /tmp/deb/boot total 2492 drwxr-xr-x 2 lev lev4096 Dec 26 04:28 . drwxr-xr-x 5 lev lev4096 Dec 26 04:28 .. -rw-r--r-- 1 lev lev 928295 Dec 26 04:28 System.map-2.6.26-2-686 -rw-r--r-- 1 lev lev 91715 Dec 26 04:28 config-2.6.26-2-686 -rw-r--r-- 1 lev lev 1506448 Dec 26 04:28 vmlinuz-2.6.26-2-686 --- Unless there's some pre or post magic that goes on, these are the same files which are currently owned by the pre-existing (debian release 17) kernel package: --- % dpkg -L linux-image-2.6.26-2-686 |grep /boot /boot /boot/config-2.6.26-2-686 /boot/vmlinuz-2.6.26-2-686 /boot/System.map-2.6.26-2-686 --- So like I said in my initial email, *concurrent* installs of kernel packages doesn't seem feasible by just installing the next kernel handed to me by apt-get. To be explicit, by concurrent, I mean that I could boot into both - release 17, and 21 - via the boot loader. As it stands, if I were to install the linux-image-2.6.26-2-686_2.6.26-21_i386.deb package, I would be only able to boot to release 21. -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: concurrent installs of previous + current kernels
Hi Boyd, On Feb 1, 2010, at 7:20 PM, Boyd Stephen Smith Jr. wrote: New upstream versions, or any version that changes the kernel ABI will be retained (or at least can be simply retained), since the kernel ABI will be in the new package name--it is a replacement only as far as relatively weak dependencies are concerned. For example, your Lenny kernel will be retained when/if you get a kernel from Squeeze (e.g. during a release-time upgrade). You can install the Lenny, Squeeze and Sid kernels side-by-side with some ease currently. This makes perfect sense. If there is a Lenny-and-a-half release, it might have a kernel with a new ABI. (The Etch-and-a-half release did.) If that is the case, installing the Lenny-and-a-half kernel and the Lenny kernel concurrently will also be quite easy. Well, in my case, the difference between the kernel images provided by the following two debs: linux-image-2.6.26-2-686_2.6.26-17_i386.deb linux-image-2.6.26-2-686_2.6.26-21_i386.deb don't differentiate by a release as I understand it, correct? You could. Since the kernel ABI determines module compatibility, modules for the new kernel would work with the old one it is was not assigned a package name. OK, this is good to know. thanks for your help! -lev -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: libc relocation error
If anyone is going through this problem, hopefully my resolution might help them - it looks like in the upgrade of libc-dev, something got hosed. The interim solution to the problem while I fixed the install was to set the LD_LIBRARY_PATH=/lib, as it seemed that the problem was arising from the symbols in the files in /lib/tls. After the full install, things started working, and I unset the environment variable. -lev On Aug 17, 2007, at 3:44 PM, Lev Lvovsky wrote: after further digging, I've found this thread on the debian glibc list: http://lists.debian.org/debian-glibc/2005/03/msg00145.html I'll be looking into how I can fix the problem - their suggested fix does not work. -lev -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
libc relocation error
Hello, it looks like after running apt-get upgrade on my box, my glibc environment has been broken: Updated the following: Get:1 http://http.us.debian.org stable/main libc6-dev 2.3.6.ds1-13etch2 [2717kB] Get:2 http://http.us.debian.org stable/main libc6 2.3.6.ds1-13etch2 [4699kB] Get:3 http://http.us.debian.org stable/main lsb-base 3.1-23.2etch1 [16.6kB] Get:4 http://http.us.debian.org stable/main x11-common 1:7.1.0-19 [338kB] Get:5 http://http.us.debian.org stable/main debian-archive-keyring 2007.07.31~etch1 [9318B] Get:6 http://http.us.debian.org stable/main nano 2.0.2-1etch1 [547kB] Get:7 http://http.us.debian.org stable/main locales 2.3.6.ds1-13etch2 [3967kB] Get:8 http://http.us.debian.org stable/main nfs-common 1:1.0.10-6 +etch.1 [128kB] Get:9 http://http.us.debian.org stable/main apache2-utils 2.2.3-4 +etch1 [341kB] Get:10 http://http.us.debian.org stable/main apache2-mpm-prefork 2.2.3-4+etch1 [418kB] Get:11 http://http.us.debian.org stable/main apache2.2-common 2.2.3-4 +etch1 [962kB] Get:12 http://http.us.debian.org stable/main tetex-base 3.0.dfsg. 3-5etch1 [22.7MB] Get:13 http://http.us.debian.org stable/main xlibs-data 1:7.1.0-19 [18.5kB] dpkg subsequently segfaulted, and I now get the following error when running commands like dpkg, or even 'ls': ls: relocation error: /lib/libpthread.so.0: symbol __libc_stack_end, version GLIBC_PRIVATE not defined in file ld-linux.so.2 with link time reference I'm running kenel 2.6.18, on etch. Any help would be appreciated. thank you, -lev -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: libc relocation error
replying to my initial post - I just subbed for the purposes of finding out about this issue - I'm assuming that if this was widespread, I would've heard a yeah, it's being worked on sort of reply. Given that - can anyone recommend even a place to start from as far as this problem goes? It seems that the linking to glibc has gone awry either on my side or (less likely), on the etch side, is there a glibc package I can use to fix the problem? thanks, -lev On Aug 17, 2007, at 12:42 PM, Lev Lvovsky wrote: Hello, it looks like after running apt-get upgrade on my box, my glibc environment has been broken: Updated the following: Get:1 http://http.us.debian.org stable/main libc6-dev 2.3.6.ds1-13etch2 [2717kB] Get:2 http://http.us.debian.org stable/main libc6 2.3.6.ds1-13etch2 [4699kB] Get:3 http://http.us.debian.org stable/main lsb-base 3.1-23.2etch1 [16.6kB] Get:4 http://http.us.debian.org stable/main x11-common 1:7.1.0-19 [338kB] Get:5 http://http.us.debian.org stable/main debian-archive-keyring 2007.07.31~etch1 [9318B] Get:6 http://http.us.debian.org stable/main nano 2.0.2-1etch1 [547kB] Get:7 http://http.us.debian.org stable/main locales 2.3.6.ds1-13etch2 [3967kB] Get:8 http://http.us.debian.org stable/main nfs-common 1:1.0.10-6 +etch.1 [128kB] Get:9 http://http.us.debian.org stable/main apache2-utils 2.2.3-4 +etch1 [341kB] Get:10 http://http.us.debian.org stable/main apache2-mpm-prefork 2.2.3-4+etch1 [418kB] Get:11 http://http.us.debian.org stable/main apache2.2-common 2.2.3-4+etch1 [962kB] Get:12 http://http.us.debian.org stable/main tetex-base 3.0.dfsg. 3-5etch1 [22.7MB] Get:13 http://http.us.debian.org stable/main xlibs-data 1:7.1.0-19 [18.5kB] dpkg subsequently segfaulted, and I now get the following error when running commands like dpkg, or even 'ls': ls: relocation error: /lib/libpthread.so.0: symbol __libc_stack_end, version GLIBC_PRIVATE not defined in file ld- linux.so.2 with link time reference I'm running kenel 2.6.18, on etch. Any help would be appreciated. thank you, -lev -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: libc relocation error
replying to myself - I just subbed for the purposes of finding out about this issue - I'm assuming that if this was widespread, I would've heard a yeah, it's being worked on sort of reply. Given that - can anyone recommend even a place to start from as far as this problem goes? It seems that the linking to glibc has gone awry either on my side or (less likely), on the etch side, is there a glibc package I can use to fix the problem? thanks, -lev On Aug 17, 2007, at 12:42 PM, Lev Lvovsky wrote: Hello, it looks like after running apt-get upgrade on my box, my glibc environment has been broken: Updated the following: Get:1 http://http.us.debian.org stable/main libc6-dev 2.3.6.ds1-13etch2 [2717kB] Get:2 http://http.us.debian.org stable/main libc6 2.3.6.ds1-13etch2 [4699kB] Get:3 http://http.us.debian.org stable/main lsb-base 3.1-23.2etch1 [16.6kB] Get:4 http://http.us.debian.org stable/main x11-common 1:7.1.0-19 [338kB] Get:5 http://http.us.debian.org stable/main debian-archive-keyring 2007.07.31~etch1 [9318B] Get:6 http://http.us.debian.org stable/main nano 2.0.2-1etch1 [547kB] Get:7 http://http.us.debian.org stable/main locales 2.3.6.ds1-13etch2 [3967kB] Get:8 http://http.us.debian.org stable/main nfs-common 1:1.0.10-6 +etch.1 [128kB] Get:9 http://http.us.debian.org stable/main apache2-utils 2.2.3-4 +etch1 [341kB] Get:10 http://http.us.debian.org stable/main apache2-mpm-prefork 2.2.3-4+etch1 [418kB] Get:11 http://http.us.debian.org stable/main apache2.2-common 2.2.3-4+etch1 [962kB] Get:12 http://http.us.debian.org stable/main tetex-base 3.0.dfsg. 3-5etch1 [22.7MB] Get:13 http://http.us.debian.org stable/main xlibs-data 1:7.1.0-19 [18.5kB] dpkg subsequently segfaulted, and I now get the following error when running commands like dpkg, or even 'ls': ls: relocation error: /lib/libpthread.so.0: symbol __libc_stack_end, version GLIBC_PRIVATE not defined in file ld- linux.so.2 with link time reference I'm running kenel 2.6.18, on etch. Any help would be appreciated. thank you, -lev -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: libc relocation error
[sorry for the multiple replies] On Aug 17, 2007, at 2:21 PM, Ron Johnson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 08/17/07 16:16, Lev Lvovsky wrote: replying to my initial post - I just subbed for the purposes of finding out about this issue - I'm assuming that if this was widespread, I would've heard a yeah, it's being worked on sort of reply. Given that - can anyone recommend even a place to start from as far as this problem goes? It seems that the linking to glibc has gone awry either on my side or (less likely), on the etch side, is there a glibc package I can use to fix the problem? Are you running a mixed system? I didn't think that I was, however I upgraded from sarge to etch a while back, so perhaps that's how it's left my system? I'd noticed this problem earlier when I booted into an older kernel. I mistakenly thought nothing of it given the older kernel version (2.4.x), but on second thought, that should have nothing to do with it. If I indeed am running a mixed system, how can I switch entirely to the etch release? thanks! -lev -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: libc relocation error
Ron, On Aug 17, 2007, at 2:42 PM, Ron Johnson wrote: If I indeed am running a mixed system, how can I switch entirely to the etch release? Eliminate all Sarge references from sources.list. my sources.list is as follows: deb http://security.debian.org etch/updates main contrib non-free deb http://http.us.debian.org/debian stable main contrib non-free given that dpkg is broken, is there any other way to check the list of debs that the system has installed for any sarge packages? thanks, -lev -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: libc relocation error
after further digging, I've found this thread on the debian glibc list: http://lists.debian.org/debian-glibc/2005/03/msg00145.html I'll be looking into how I can fix the problem - their suggested fix does not work. -lev On Aug 17, 2007, at 2:51 PM, Lev Lvovsky wrote: Ron, On Aug 17, 2007, at 2:42 PM, Ron Johnson wrote: If I indeed am running a mixed system, how can I switch entirely to the etch release? Eliminate all Sarge references from sources.list. my sources.list is as follows: deb http://security.debian.org etch/updates main contrib non-free deb http://http.us.debian.org/debian stable main contrib non-free given that dpkg is broken, is there any other way to check the list of debs that the system has installed for any sarge packages? thanks, -lev -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Policies (was Re: inappropriate racist and other offensive material)
I'm sure I'll get flamed to all hell with this, but... is it at all possible that these sorts of threads could be taken off this list, and put on the more appropriate debian-policy list? I'd really like to take something away from the discussions on the debian mailing lists, but subjects like these, while serving a VERY necessary purpose to the distro, seem a little bit too masturbatory, and inappropriate to a user, and devel list. thanks, -lev (where'd that asbestos suit go?) On Thu, 14 Mar 2002, Dimitri Maziuk wrote: * Duncan Findlay ([EMAIL PROTECTED]) spake thusly: ... Mr. Long, I don't know how old your daughter is, but I'm 100% sure that she will be exposed to more content that you deem inappropriate than you could possibly control. Perhaps instead of trying to protect your daughter from the inappropriateness of life, you should attempt to teach her what is right and what is wrong. Of course, I'm in no position to lecture on parenting -- I'm 15 years old. Damn right you aren't, but not because of your age. It's because you sound like an intelligent reasonable sentient life form -- something that creatures who want to limit our freedom of expression For Da Kidz(tm) are clearly not. Dima (if thine eye offends thee, pluck it the fuck out, Mr. Elongated) -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com What I'm wondering is what condiment you have 'confidence' in. I like to eat things that taste good, not ones I can vote into public office. -Dan Mick
building packages from source (apt-get source)
Hello, Could someone please point me to the documentation for building debs from source *WITH* changes to the source (via the configure script)? I'm trying to compile PostgreSQL from source, after having done an apt-get source postgresql. After the download, dpkg-source was automatically run. Going into the directory: ./configure --with-perl debian/rules build While it seems to compile, it also later complains about missing python (which is functionality that I *DON'T* want built in). I can think of no other way to keep this from happening other than editing the debian/rules file and removing whatever I don't likethis seems rather half-assed, so I'm assuming there's a better way. I've tried man'ing the dpkg-source source management utils, only to be completely baffled by the cryptic docs... help? thanks! -lev
serial console problems
I'd like to get a serial console working on my debian (unstable) box working, so that I can test various network setups on it w/o fear of breaking the connection. in my lilo.conf: boot=/dev/hda root=/dev/hda1 install=/boot/boot.b map=/boot/map vga=normal delay=20 serial=0,9600n8 image=/boot/2.2.20 label=2.2.20-console read-only append = console=ttyS0,9600 console=tty0 in my inittab (among other things): T0:23:respawn:/sbin/getty -L ttyS0 9600 vt100 I have serial support built into the kernel... When the system boots, I get all kernel messages just fine up until the swap partition is mounted, then: Adding Swap: 130748k swap-space (priority -1) f~x (pine doesn't let me print the other strange chars I get...). I later get the login prompt, but the input is slow, and often characters that I've typed are missing. I have 9600 8N1 setup on my terminal window (same speed as set by getty), and I've also seleted the ioctl.save file as recommended by the readme for serial console support. I can never login, because either the login session times out, or the password I enter is incorrect (which it isn't). help? thanks! -lev -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com Progess is the direct result of dissatisfaction. -Mark Rudholm
Re: Webmail software..
openwebmail is nice :) -lev On Fri, 28 Dec 2001, Petre Daniel wrote: What other webmail software is up there? i know squirell and basilix.. Petre L. Daniel,System Administrator Canad Systems Pitesti Romania, http://www.cyber.ro, email:[EMAIL PROTECTED] Tel:+4048220044, +4048206200 -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com Progess is the direct result of dissatisfaction. -Mark Rudholm
Re: Email client programs (w/o html formatting - sorry)
On Wed, 26 Dec 2001, Vaughan, Curtis wrote: I am wondering what other Debian Users recommend for an email client program. First, let me state the needs of our company. We are currently very MS based at present. The goal has been set to slowly move over to Linux. Debian has been chosen, as it is the most stable that I am aware of. Of course, I don't know a hell of a lot yet. what's the goal of the change over? money? a general MS disliking? if it's a money issue, have you looked into MS's terminal server? You can use this in conjunction with linux boxes (do a search on freshmeat for rdesktop). That way all the software that you haven't foudn replacements for runs on the terminal server. I have tried out a number of email client programs, but am not satisfied by any one of them. I have tried out, KMail, Evolution, Netscape and Mozilla, oh and the email client program with StarOffice. Each one has something good about it, but none of them are as good as Outlook. One of the big problems might be the fact that I don't know how to configure them, of course. Evolution is close to being rolled out the door, though it doesn't offer the same kind of group-centric functions as Outlook, it's on its way...Mozilla is in fact differnt from Netscape - it is the next generation browser from [some] of the same authors as Netscape. It has it's own problems, and I generall prefer to use stand-alone applications vs. all-in-one. One of the big problems is that people will need to read and send mail in English and Russian. I'm trying to recall, but I don't think I could get StarOffice and KMail to permit Russian. Actually I could write in Russian in KMail provided I opened a letter that was already in Russian and composed my letter in that letter. Understand? have you tried this with Evolution? So, I'm what other programs people use and find to be very good. I haven't tried Pine, but I don't think I could get our users to go over to a non-GUI program. I use pine for all of my mail, but for regular users, there is quite a learning curve, and it doesn't offer any of the shceduling properties of Outlook. Your best bet right now is probably Evolution, though, having personally tried it, it doesn't play nice with all of Outlook/Exchanges features... again...the main question is why?. -lev
RE: Email client programs (w/o html formatting - sorry)
On Wed, 26 Dec 2001, Vaughan, Curtis wrote: One of the reasons is monetary. Why pay MS for what is already out there, and works just as well? The other reason is example. We know of another company that is totally Linux-based and has no problems. They can work not just at work, but also from home or from any computer anywhere through a web browser. Actually we are using Terminal Services. The 1 user who is on a Linux box is only on it because he is using rdesktop. I use it as well. About email clients: practically no one uses scheduling in our office, so that's not an issue. Something happened with my Evolution install that left me very dissatisfied. I think it just crashed often. The email client that we adopt certainly does not have to offer the all-in-one possibilities of Outlook. But it does have to be stable. in that case, I would go for evolution. start with a clean install, and make sure you have EVERYTHING necessary for evolution to work (use their .debs + depends). -lev
Re: Email client programs
On Wed, 26 Dec 2001, Erik Steffl wrote: imo it's very useful to have the eamil deliverystorage separate from email clients. from this point of view the ideal situation is to use IMAP, server side filtering (like sieve with cyrus) and let them use any clients they want... from a management side, letting users use whatever they like can be problematic. unification makes problem solving a lot simpler. not sure how to get there from exchange, does exchange offer imap access? yes it does. as well as pop3 (plus an MS one I believe..) -lev
Re: Email client programs
On Wed, 26 Dec 2001, Craig Dickson wrote: Brian Nelson wrote: ... but I don't think I could get our users to go over to a non-GUI program. Why is that? Because they've bought into the marketing pitch that pretty graphics == better software? That's bullshit. There's no good reason any user couldn't become more proficient with a text-mode client. What I've found in trying to get GUI-dependent people to use mutt is that they resist having to memorize a bunch of keystroke commands (no matter how easy, and despite the fact that the most common ones are listed at the top of the screen!). They'd rather point and click; they know how to do that already, and they don't see why they should go to the (minimal) bother of learning a keyboard-based UI. This isn't an intelligence problem, either; some of these people are quite bright, but they don't see learning mutt as worth their bother when good GUI alternatives exist. And Sylpheed is actually pretty good; if I didn't need a text-based interface (for accessing my mail in an ssh session across the internet, on a slow enough connection that X forwarding is out of the question), I might use it myself. But since I need a text-mode MUA, I use mutt. agreed. pretty graphics have nothing to do with better software. same thing with text-based interfaces - neither is a clencher in the argument, the underlying design is. that being said, there are people that *gasp* don't want to learn anything new in order to use a computer. some people don't delve nearly as far into it as we do, and use a computer merely as a tool. same can be said with me, and tax forms vs. an accountant. -lev
jumpstart for debian?
is there any linux/debian equivalent of jumpstart as there is on Solaris? thanks, -lev -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com Progess is the direct result of dissatisfaction. -Mark Rudholm
Re: Email client programs
On Wed, 26 Dec 2001, Craig Dickson wrote: Well, the MUA should not have to worry about retrieval from POP servers. That's fetchmail's job. But certainly the MUA is the thing for interactively moving mails from one folder to another. why the insistence on fetchmail? -lev
Re: Email client programs
On Thu, 27 Dec 2001, martin f krafft wrote: why the insistence on fetchmail? it's the unix philosophy -- let one program do its job and do it well, let other programs use that... hehe, really? I guess you can count mozilla out ;) while I like the concept of modularity, I can't imagine that actually fetching the mail can be very hard and/or killer-app-able. hehe, I like me my pine :) -lev -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com Progess is the direct result of dissatisfaction. -Mark Rudholm
Re: Email client programs
On Wed, 26 Dec 2001, Craig Dickson wrote: Lev Lvovsky wrote: why the insistence on fetchmail? Without fetchmail or something like it, if your network is down when you decide to run your mail client, you can't check your mail server for new messages. Even if the network is up, you have to wait while the client connects to the server and downloads all your new mail (which may take a while). hmm... since i run my mail srver on the same box as I run pine from, it's never really been an issue...I guess though, that for a bigger setup, this would indeed be a good idea... -lev
Re: Where do you RTFM ?
linuxdoc.org for whole documentation groups.deja.com for specific questions (this list posts to usenet as muc.lists.debian.user methinks) -lev On Mon, 24 Dec 2001, Martin Emrich wrote: Hi All ! When newbies ask something, they are often asked to RTFM... But where do I find the Information ? Very often, neiter #man foo nor the contents of the /usr/doc/foo are very helpful. Where is aditional documentation usually installed ? Or do you know a good getting-documentation-howto ? (maybe even something in german.. would be nice) Thx Martin PS: Merry Christmas to all of you ! -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com Progess is the direct result of dissatisfaction. -Mark Rudholm
Re: C Manual
for C, I would recommend the Kernighan Ritchie book. -lev On Thu, 20 Dec 2001, Phil Beder wrote: Where can I find a good, complete manual for C and C++ programming languages for the gcc compiler. Other resouces for learning C also accepted. I played with C when I was a mainframe programmer and wanted to get back in touch with it again. -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com Progess is the direct result of dissatisfaction. -Mark Rudholm
Re: OT: mysql vs. postgresql
On Wed, 19 Dec 2001, Jeffrey W. Baker wrote: Oh, and don't go anywhere near mysql 4. thanks for the info up top...I'm actually in the process of choosing b/w these two right now... care to be more specific RE the warning? -lev -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com Progess is the direct result of dissatisfaction. -Mark Rudholm
Re: faxing over the internet
a small plug for the company I work for: efax.com -lev On 18 Dec 2001, Lance Hoffmeyer wrote: Anyone know if there is a way to send and receive faxes over the internet. I am in need of fax capabilities but do not have a modem on my computer, only DSL. Lance -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com she [Britney Spears] would return a year later with a self-written album co-produced by the dust brothers and aphex twin called 'the britney spears electric kool-aid lonely hearts club experience' with sitars, darkstep acid jungle beats, and post-rock idm emo-core. or maybe she'd just make a straightforward pop album, but actively encourage her fanbase to take psychadelics. -Ben Yater
aterm -n -e options
hello, i've switched from wterm in favor of aterm for the fact that aterm can shade the terminal screen when in transparency mode (wterm's lack of documentation doesn't seem to allow this). however, contrary to the man page, the -n, -T option, which gives the terminal window a title, and supposedly overrides the title from the program run with the -e command, is not in fact overriding the name of the command that's run. I need to sort my windows by name, and I run the same expect script in a lot of them, so this becomes quickly unwieldy. prolly not the right forum, but any suggetions would be appreciated... thanks! -lev
Re: Do you lose your Drafts, bookmarks,etc. when you cvs mozilla ?...........
~/.mozilla generally -lev On Mon, 17 Dec 2001, Courtney Thomas wrote: Greetings ! If you don't loose them, where are they after cvsing mozilla ? Thank you, Courtney -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com she [Britney Spears] would return a year later with a self-written album co-produced by the dust brothers and aphex twin called 'the britney spears electric kool-aid lonely hearts club experience' with sitars, darkstep acid jungle beats, and post-rock idm emo-core. or maybe she'd just make a straightforward pop album, but actively encourage her fanbase to take psychadelics. -Ben Yater
blackbox + bash aliases?
is there a way to get blackbox to expand aliases in the bash .rc files? I am opening a wterm window with the same args every time, and I'd like not to clutter up the menu files. i've tried the shopt expand_aliases option, but I still get blackbox complaining that /bin/sh can't find the alias (bbterm is the name I'm using). otherwise, if there's a way to define variables in the blackbox menu files, that would be even better! thanks! -lev -- personal site :: www.sonous.com rave site :: raves.sonous.com I'm a DJ! site :: djkgb.sonous.com she [Britney Spears] would return a year later with a self-written album co-produced by the dust brothers and aphex twin called 'the britney spears electric kool-aid lonely hearts club experience' with sitars, darkstep acid jungle beats, and post-rock idm emo-core. or maybe she'd just make a straightforward pop album, but actively encourage her fanbase to take psychadelics. -Ben Yater
process/user management
hello, a bit newbie at this kind of stuff, but what would be some good resources to look at for resource managment for processes, as well as users? managing things like CPU time, quotas, and bandwidth. this, as well as info on securing boxes for mass shell usage. I know that's a lot, but any info would be great! thanks! -lev
procmail/formail to change subject?
I'm using procmail to process a number of mailing lists. Having just subscribed to several Debian lists, and having them all be delivered to a Debian mailbox, I'd like to make it obvious which specific debian list they are all from (and I don't want to make extra mailboxes for each list). Is it possible to put something like [user], and [security] in front of the appropriate messages? Could this account for replies also, where it'd put Re: [user] instead of [user] Re:? dunno if this is a tall order. thank you! -lev -- personal site :: www.sonous.com Persons of leisurely moral growth often confuse giving with taking. -Larry Wall
Samba .deb defaults
Hello, I'm in the process of experimenting with Samba and Netatalk in combination, and would like to have Samba compiled with the interoperability between these two turned on. How do I know what options the Samba .deb file has been compiled with (or all .deb files for that matter), and how, after I'd DL'ed the source (and the subsequent Debian patches), do I change the options. Also, uninstalling Samba 2.07-2 (?) totally deleted /etc/samba ... has this been addressed at all? thank you! -lev -- personal site :: www.sonous.com Persons of leisurely moral growth often confuse giving with taking. -Larry Wall
services I don't need - turn them off?!?!?!
Hello, for the sake of knowing how to do such thigns, I'd like to turn off some services that I feel comprimise my server security, and aren't being used. with portscan, this is what I get from a non-local host. 9 - discard 13 - daytime *21 - ftp *22 - ssh *23 - telnet *25 - smtp 37 - time 53 - domain 79 - finger *80 - www *110 - pop-3 111 - sunrpc 113 - auth 512 - exec 513 - login 514 - shell 515 - printer the ones with the *'s are the things I KNOW I need...I'm sure there are others that I need. the most annoying one is the sunrpc...i've tried renaming portmap, and restarting inetd, but it's still there. thanks for any help! -lev
restarting daemons
Hello, could someone please let me know how to restart all of the apache www daemons that are running so that it'll read the changes made in the httpd.conf file? 'apache restart' doesn't do anything is there a generally accepted way of doing this for all 'net daemons? also, i'm sruprised nobody knows how to shutoffthe sunrpc daemon (or at least that's what I believe it is when I portscan my machine)...any clues as to where I could find info on this? thank you!! -lev //sig: //Social graces are the packet headers of everyday life
portscan: sunrpc etc...
hello, upon runnign portscan on my computer from a nonlocal host, I get the following: 9 - discard 13 - daytime 21 - ftp 22 - ssh 23 - telnet 25 - smtp 37 - time 79 - finger 80 - www 110 - pop-3 111 - sunrpc 113 - auth as I use pretty much all of the standard daemons, I don't know what 'discard', 'daytime', 'time', 'sunrpc', and 'auth' areif anyone can suggest ways for turning these off, I'd appreciate it (inetd.conf doesn't do have anything that I've not checked as far as I know). I especially need help with sunrpc, as it's nowhere to be found in rc*... also, can somone give me a straight answer on how to restart initd? I've heard 'kill -HUP inetd', but it gives me an error of not being able to find it...I just kill it by pid, nad restart it manually (type 'inetd'). thanks -lev //sig: //Social graces are the packet headers of everyday life
SGI Indy500 linux
if anyone has experience with this setup, please let me know. Our school had this system donated to it, and it's currently running IRIX (me -no experience). I'm going to be using the modified Red Hat Hard Hatdistro put together by ppl @ SGI, but am curious if anyone has any useful tidbits of info. thanks! -lev //sig: //Social graces are the packet headers of everyday life
SGI indy 500 linux?
Hello, I work at my school radio station as a member of the web department. Currently, we stream audio over the 'net using an NT computer (I serve the html from a Debian box sitting in my room)...we recently had an SGI Indy500 donated to the department, and would really like to use it as the streaming server if at all possible. Needless to say, I'd like to load linux on it (currently it's equipped with IRIX). does anyone have any experience with this? SGI has made available a Hard Hat distro of a modified Red Had 5.1...am I correct in assuming that whatever software is available for linux can simply be compiled on the machine to run on it? thanks! -lev
gtk-config problem...
Hello, I decided to install a gtk-based IRC client (to get help in #linux ironically enough), and I encountered some problems with the configure script. I've installed all of the gtk and glib libs, and also the dev files, so that gtk-config exists (though I can't find glib-config), and I'm sure they're more recent than the minimum req'd version (1.0.4)...but whenever I run the script, I get: configure: error: Cannot find correct version of GTK - is gtk-config in path? and I've also checked, and 'ldconfig'ed the /etc/ld.so.conf file, and made sure that it's got all of the proper paths in it (typing 'gtk-config' @ the CL gives me the proper prompt. anyone have any suggestions? thanks!! -lev
Davicom 9102 NIC w/ 2.1.12 kernel
Hello, I recently purchased a MB that has pretty much everything integrated onto it, including the NIC and sound card. The manufacturer of the NIC was kind enough to actually write up a linux driver which must be compiled and added to the kernel as a module. Prior to my switching (or attempting to) to kernel 2.2.x, I was running generic slink, and the module compiled just fine. trying it now gives me the error message that the module was compiled for 2.0.36, and that it won't work...is there any way around this? am i stuck with 2.0.36 unless I get a new NIC? the main reason why I went with the new kernel was for [assumed] better sound card support...I don't know if the SC is sound blaster compatible, and I didn't see it mentioned as a supported model. am I SOL here too? [and the reason why I am having problems with gtk, is 'cus I wanted and irc app to ask questions in #linuxwhat a tangled web] thanks, -lev
Re: TeK WYSWYG editor?
if you mean Tex, Lyx is a great program that I used to make a whole scientific report (granted it took longer than if I knew the language, but I didn't have time to learn it). Debian.org has the packages... hth -lev On Mon, 30 Aug 1999, Guilherme Soares Zahn wrote: Hi there, do anyone know a good WYSWYG TeK editor for Linux (much like Scientific Word for Windows)? In case you don't... Is there any program to convert from any of the 'regular' word processors to TeK (even if I'd have to do some 'fine tuning' on the TeK file later, it would be *a lot* easier then learning TeK from scratch, wouldn't it? ;-) []'s Guilherme Zahn -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
X window resolution extremely low
hello, I installed slink on my workstation here @ work, and I'm having problems with the resolution of X...it's horribly low. I'd say 300x240. I've run xf86setup multiple times (with different setups) to no avail. I think that hte problem may lie in the fact that it's a non-standard vid card, that's embedded on the MB. the motherboard is a SiS Slot1/Socket 370 model, and the vid card unit is a SiS620 (if I'm correct)...with, what I believe is 8MB RAM (anyway for me to check this?) I'd like to get it up to at least 8x6 @ 32bit, if not 1024x768 (16 or 32bit would be fine) any help would be appreciated! -lev
LILO problems with new HD
Hello, I've recently discovered problems with the drive that had the /boot and /home partitions of my Debian box, so I decided to go out and get a new HD in transferring the files, I simply copied /boot/* and /home/* to the new drive, and ran lilo on it. It gave me the message that lilo was installed, and I tried rebooting. Lilo being's to load, but only goes so far as LIL- until it freezes...in the lilo docs, this says that there is some sort of corruption of the map files, or a problem with the map installer. Seeing as these copied files were from a corrupted (suspected at least) disk, I figure that's the problem. Can anyone help me in getting fresh copies of the files that LILO requires for installation? I don't know if they are custom to my setup (untouched slink kernel), or if they are generic files thatI could simply pick up somewhere off of the CD's. thanks in advance!! -lev
Re: LILO problems with new HD
I've recently discovered problems with the drive that had the /boot and /home partitions of my Debian box, so I decided to go out and get a new HD in transferring the files, I simply copied /boot/* and /home/* to the new drive, and ran lilo on it. It gave me the message that lilo was installed, Is that all you did? What drive partition did you put /boot on? The info you supplied is insufficent for me to help you with. My crystal ball is cracked, sorry. hehe, sorry, I post under the assumption that ppl's balls are in working order ;-D. /boot is hda1...there are three partitions on the drive, none of them extended. this is what I do @ the console: ** silver168:~# lilo -v LILO version 21, Copyright 1992-1998 Werner Almesberger Reading boot sector from /dev/hda Merging with /boot/boot.b Boot image: /boot/linux Added Linux * /boot/boot.0300 exists - no backup copy made. Writing boot sector. ** I'm a little bit confused as to the difference between this command, and 'lilo -v -b /dev/hda1', which gives me the following: ** silver168:~# lilo -v -b /dev/hda1 LILO version 21, Copyright 1992-1998 Werner Almesberger Ignoring entry 'boot' Reading boot sector from /dev/hda1 Merging with /boot/boot.b Boot image: /boot/linux Added Linux * /boot/boot.0301 exists - no backup copy made. Writing boot sector. ** again, when I reboot, and take out the floppy (the way I've been booting for a long time now seeing as my other HD gave me this same problem), I get a LIL-, which the docs say is a problem with the map file. any help would be appreciated :) -lev
Re: LILO problems with new HD
Ok I guess I have to spend more time to ask specific questions.. sigh You said I simply copied /boot/* and /home/* to the new drive, and ran lilo on it. You didn't say if you updated the /etc/fstab file. yes, /boot, /home, and the swap partition are updated to reside on the new drive...they are workng properly as I type. I wouldn't think that this would effect lilo botting, but either way, everything's fine here. You didn't say if you updated /etc/lilo.config to look for the new /boot on a different drive. well, since the new drive is essentially a copy of the old drive (just bigger), I didn't change anything here. In the process of changing things over, I mounted the new /boot, and /home in temporary directories, copies the dirs and files to them, and then unmounted the old /boot, and /home, and mounted the new ones in their place. I ran lilo after this, assuming that everything would just adjust to the changes. A suspicion is that the map file doesn't properly represent the new, bigger drive, but perhaps I'm wrong on this, as the map installer is part of 'lilo', and maybe it just updates the file after looking @ the new geometry. the following is my lilo.conf: ** boot=/dev/hda root=/dev/sda1 install=/boot/boot.b map=/boot/map vga=normal delay=20 image=/boot/linux label=Linux read-only ** You didn't say if you changed/removed the old lilo.conf entry that pointed to the OLD location of /boot. again, the new drive has the same dir/partition structure (right down the the numbering, so essentially I didn't even change anything in /etc/fstab You didn't say if you now have 2 drives ( I assume you do as you only copied /boot and /home to the New one) or if there is now only one drive. yes, I have one SCSI drive, and one IDE drive...the scsi drive CANNOT boot, as it's hooked up to a non-bios scsi card (which requires drivers to initialize). I have written oncee in an effort to help. I have written a second time to try to help YOU to help yourself. well, hopefully this answers both of our questions... thanks -lev
secure ftpd?
hello, a freind recently introduced me to the concept of ssh, and I really like what it allows one to do (or not do rather :). I'm curious though, is there such a thing as a secure ftpd where the data is encrypted to keep people from seeing what you're transferring (sorry, just a little paranoid)? I'm generally new to secure communication, and perhaps this is a function of ssh. Would a special client program be necessary for this? are they widely available (for various platforms)? any help or URLS o nthis subject would be great!! thanks, -lev
LaTeX
Hello, I need to use LaTeX for a project in one of my classes, and was wondering whether anyone could recommnd a particular type that does mathematics-related things (integrals etc..), i'm quite a newbie to LaTeX (understatement), so a GUI-based version would be cool. thanks for any help! -lev
Re: downloads...burst, then nothing
Thanks for the reply :) the NIC in the linux machien was set to half duplex. However, the setting on the Windows machine is Auto, and it doesn't work when I manually set it to half fuplex. The card is an Intel Etherxpress 10/100. downloading form the web/ftp from a computer not on my network works just fine. I'm assuming the duplex situation is the problem? thanks! -lev At Sunday 06:16 PM 5/16/99 , you wrote: Lev Lvovsky wrote: I'm having a problem with downloads from my machine (slink). On the network I have at home, I can upload at the regular high speeds, but downloading form the machine (to a Win98 machine), whether it be by ftp, or by http gives me a little bit of the file, and then slows down to a total crawl. Obviously if it owrks up, I'm assuming all ofthe hardware is fineis there some sort of buffer for this sort of thing that might be set too low? Make sure that the network interface on both machines are set up for half duplex operation (unless you are using a switch). Having one or both interface at full duplex might cause problems like these. Also, check if you are getting an abnormal amount of collision on the network during download (ifconfig eth0 on the Linux box).
downloads...burst, then nothing
Hello, I'm having a problem with downloads from my machine (slink). On the network I have at home, I can upload at the regular high speeds, but downloading form the machine (to a Win98 machine), whether it be by ftp, or by http gives me a little bit of the file, and then slows down to a total crawl. Obviously if it owrks up, I'm assuming all ofthe hardware is fineis there some sort of buffer for this sort of thing that might be set too low? thanks!!! -lev
DNS is dead....
Hey all, first off, thanks for the previous replies to my question RE DNS. now a bigger problem... as far as I can tell, my DNS server is dead to all outside connections. the domain that I'm hosting is klaradio.com, using nslookup, I can find various subdomains, and entries, byt when I set the server to anythign but local host, nothing can find it. pinging the IP works fine, and the doamin name given to me by my ISP (silver168.mminternet.com) works fine too. I've tried 'ndc reload', and 'ndc restart', but neither do anything in being able to access it from the outside... also, I'm slightly suspicious that this might have played a role in not having the lookup of www.klaradio.com appear quickly (usually 10-30 seconds!!)...even if I were on a modem it'd prolly be faster than that. and help appreciated!!! -lev
DNS time to live lookup delays...
Hello, I recently set it up so that my linux box runs the klaradio.com domain. I am running my own DNS server off the machine, and I have a few questions: one of the addresses that I'd like to put under a subdomain of the site (stream.klaradio.com) has it's IP changed somewhat frequently (once a month), a guru here at work mention the TTL option in the DNS settings for the site as being updated more frequently than normal for this so that the cache files of other servers aren't outdated...since then the guru has left (*snifff*), and I'm really curiou how to do thismore specifically, would this change the TTL times of the whole site? I assume that it would, as I 've only seen one entry for TTL for a whole zone. Also, I'm running this over an ADSL line 384/384kbps that usually sees around 20Kb/s during peak hours...is it normal for there to be a delay of around 10 seconds for a lookup to be performed by a non-local machine (heh, go ahead, give it a try :). thanks !!! -lev
spam debian-user?
hey all, I noticed thast this list gets posted to usenet...what do ppl on the list do to avoid gettting their e-mail addresses getting picked up by spam bots? does the system get rid of this possiblity somehow? -lev
setting up DNS
Hello, I'm going to be registering a domain soon, and I'm having some problems...after having gone to www.internic.net, and filling out the forms there, I got some error messges back: ERROR: Name Server SILVER168.MMINTERNET.COM/209.241.149.168 is not registered with InterNIC. You cannot register this host through domain registration process because... You are not authorized to register hosts in MMINTERNET.COM domain First register this host through the host registration process or choose a different name-server. my ADSL provider is mminternet.com, and my address is silver168.mminternet.com (the IP is listed above). am I missing something here? more importantly, if I'm not mistaken, I will have to run a DNS server...I've read through some of the DNS-HOWTO, but I need to first install a DNS server...I've looked for one on the CD's (from www.cheapbytes.com), but was only able to find dnsutils. Am I looking in the right place? are there any URLs what describe in-depth what I'm trying to do? thanks for any help!! -lev
Re: setting up DNS
At Sunday 05:03 PM 4/11/99 , dyer wrote: Lev Lvovsky wrote: my ADSL provider is mminternet.com, and my address is silver168.mminternet.com (the IP is listed above). You've answered your own question here mminternet.com _is_ the domain, silver168 is only a machine on the domain. You don't register each machine. hrm, maybe there's still something that i'm missing...let's say I want to register www.linuxnewbie.com (just an example), and have it point to my machine (silver168). The ADSL provider provides me with a static IP, so I figured this'd be possible (plus a friend of mine with the same setup did it)... you don't _need_ to be a dns server. ( You can install bind) BTW, I just pinged your box, and got a reply. YOur'e good to go. :-) hehe, I should hope you'd be able to ping the machine, it's my mail server ;). Cool, I just found bind (was using DNS as a search string). thanks :) -lev
gnome-session missing?
Hello again :) I used the apt method of installation to get Gnome from debian.org (under /~jim)...everything seemd to go OK (except for some wierd libgtk, and libglib dependencies which I stil haven't figured out, but they were only impeding the window-coordinate dealy from getting installed), and enlightenment was added, as well as pretty much everything from gnome (it seems) except for gnome-sessionI can't find it anywhere (and, therefore I can't run gnome)...has anyone else had this problem? -lev
loading Xwin32 getting an Xterm
Hello, I've decided to start using Win9x for only multimedia applications (Sound Forge etc..), and as such, would like to start up an Xwindow session ffrom my linux box on the windows box as soon as the windows box boots up (read:I'm too lazy to telnet in, and set the DISPLAY variable). Is there any way to do this? I understand that is could possible be unsecure, but I'm the only person using this computer. thanks :) -lev
need to scan drive thoroughly
Hello, well, if any of you are wondering, I was able to get debian installed on the system by low level formatting the IDE drive from BIOS (LILO kept giving crc errors upon uncompressing Linux). Now I have another problem...anything that I download to the IDE drive (where the /home and /boot partitioins are), I get crc errors upon uncompressing them. using 'fsck' on the various partitions returns no errors, but I think this might be where Im getting my crc errors when I'm decompressingthe linux image also... Sooo...is there any utility (or switch in fsck) that will do a thorough scan of the drive? Also, I have my swap partition on this drive (strangely enough I'm not getting any wierd errors while running linux)...do I have to unmount this to scan the drive? If so, how can I do this safely? thanks -lev
debian setup headaches
Hello, OK, in trying to iron out that problem that I was having with not being able to boot from my IDE disk instead of my (non-bootable) SCSI disk, I've run into a problem in even getting Debian running at all. I've started from scratch, repartitioned, and reformatted through the setup process, and made a boot floppy ( did I mention that the setup still doen't let me boot from the IDE drive?). Problem is, and I've done this like 5 times already (from scratch), I try booting on the floppy, and I get crc errors. This is using different floppoes every time (and they check out when scanned). In case you're curious, my RAM timings are set to their defaults in BIOS, and there's nothign trciky about any of the other settings in BIOS. To be perfectly honest, I thought Debian would be a thoughtful, and competent upgrade from Red Hat, but for something as simple, and probably very common as my setup, I cannot get it to work. Just for kicks, I tried RH 5.1 to see if there was anything totally messed up in my setup. Nope, setup just fine, w/o a hitch. Hopefully someone can steer me in the right direction on this, or cheapbytes.com is getting another order, from me, and it ain't gonna be for Debian. advTHANKSance -lev
Re: What DO you lose with Linux ??? (fwd)
[Lev, who's recovering after a debilitating and unexplained MB crash replies] Somone mentioned that Winamp was available as a linux port...is this true? if so, what's the URL? In answering your question, Winamp is well written, well-supported, and there's no maturing nicely excuse :). -levi On Mon, 29 Mar 1999, Michael Beattie wrote: Mac, and that's still arguable)...and to top it off, Winamp, 'nuff said. Excuse me? Who wants _Win_amp on a linux box? X11amp works perfectly :) Even though it is still maturing nicely, x11amp has everything I need. (Except the File Info box - which is coming :) )
crc errors with LILO
OK, I don't know if anyone caught my last post RE this ,but here's a quick brief: I have a SCSI and IDE system...the scsi drive holds root, but the scsi adapter doesn't have the ability to load w/o first having drivers loaded up. So what I thought would work (as it did with my RH install), was putting /boot on the IDE drive, and everything else on the scsi drive. Debian however didn't let me do this, and insisted that the SCSI drive boot first. THe ony way I can get this system to boot now, is with a floppy. So now, I used lilo, and tried installing it on the IDE drive: silver168:~# lilo -b /dev/hda5 Ignoring entry 'boot' Added Linux * now, when I try to boot, it says loading linux, and then when it gets to the decompression part, it gives me a crc error, and halts. I've checked the disk, and copied over the kernel that's on the floppy (which is good), and I still get the error. Anyone know how to fix this? I don't like to boot off of a floppy... thanks!! -lev
not booting into Xwindows
can anyone tell me how to keep my computer from booting into Xwindows? This is a rather annoying automation... thanks :) -lev
XFree86 dselect questions
Hello, I'm new to the Debian distro, and relatively new to Linux in general (was using Red Hat until its security issues began to bother me). Go easy on me :) several questions: 1. I can only run 'xinit' or 'startx' as root, running it from my regular acct says that I don't have permissions to use it. Any suggestions? I'm using xf86config to set it up 2. I'm really confused when it comes to the packaging system for debian. Although RPM was a big reason for *leaving* Red Hat, I'm kind of put off by 'dselect', and even more by 'dpkg'. I bought the CD's from cheapbytes.com, it's a 4 CD set, 2 binary, and 2 source...perhaps I'm doing soemthing wrong, but was something as simple as Pine left out of the distro? Anywho, basically, I guess dselect is the easier way to do things, but I can't even find Netscape in there. Is there a page onthe web or perhaps a HOWTO that describes these two things? 3. In RedHat, the 'su' command allowed and '-l' switch, which would take the path settings of user to be su'ed to (ususally root in my case)...any way to do that with the debian 'su'? thanks so much!! -lev
Re: XFree86 dselect questions
On Sat, 27 Mar 1999, Daniel J. Brosemer wrote: On Sat, 27 Mar 1999, Lev Lvovsky wrote: I'm new to the Debian distro, and relatively new to Linux in general (was 3. In RedHat, the 'su' command allowed and '-l' switch, which would take the path settings of user to be su'ed to (ususally root in my case)...any way to do that with the debian 'su'? su - [user] [user] is optional if su'ing to root. HTH. -Dano right, I usually use 'su' w/o the user specified, as I normally wanna get root priveleges...problem is, my PATH's aren't set. With RH, you could do an 'su -l', and get the PATH of the user you were switching to. Is there a switch for this in the debian 'su'? thanks :) -lev
Re: X-window client?
At work (tech support of UCLA's ISP), we run a linux server for certain things, and many of the em-ployees like to run Xwindows on the windows machines...without a doubt, Xwin32 is the best Xwindows client for win32, but you have to pay a big amount of money for it. Or you could get the demo version that lasts for two hours. MI/X lacks a whole lot of font support, and the display is easily messed up/does not work with certain programs (even xterm). HTH -lev There is MI/X I believe, I've never used it, but it's worth a try... check out http://www.microimages.com/www/html/freestuf/mix/ .adam
Re: XFree86 dselect questions
ahh, cool! I also messed around and found that the '--rcfile' option worked. thanks for the help! -lev On Sun, 28 Mar 1999, Gregory T. Norris wrote: You can use the the - option, as in su -, to get a similar effect. This causes the your new session to use through the new user's startup files (.bash_profile, etc). This will give you any environment settings (including PATH), as well as aliases, that they would normally have set. If what you're wanting is to inherit only the PATH, then I'm not aware of a similar option...
Re: What DO you lose with Linux ???
OK, my personal take on this: I've been using a linux box for a couple of weeks after the PS in my windows macihne died (well, using *only* linux). I used that time to learn a whole lot of stuff about linux, and also see just how many apps I could replace from my windows box. I'd honestly like to use the windows CPU for the linux box (overclocked celeron), but the fact it, Linux is lacking in a few VERY important things: 1. A killer-app X-windows e-mail client. Unless someone knows about something I don't, both 'Balsa' and 'M' don't cut it when compared to Eudora. As bloated as the latter may be, it still does thigns that no other mail client does...things I need done. And no, I'm against using Netscape's mailer as it's nont intuitive, and I don't like suite apps. Pine is out of the question also, as I quickly get tired of arrow keys for navigation. 2. You mentioned music composition...I dabble in it, and after doing more of the same searching, linux is NOWHERE near windows (or Mac for that matter) in this arena. First off, there's very little support for specialized sound cards (not just your regular sound-blaster compatibles, but ones with DSP and other goodies). Second, there are no sound manipulation apps like Sound Forge...add to it plug-ins, and you quickly realize that it's going to be YEARS until linux can even think of catching up to Win/Mac in this department (it took Win a long time to catch up to Mac, and that's still arguable)...and to top it off, Winamp, 'nuff said. for just about everything else, I'd say that there are viable alternatives, but needless to say, in a world where the biggest use for the 'net is e-mail, and 'multi-media' is a buzz-word, these two categories are huge. the question is tho, other than electricity consumption, what's so wrong with running more than one computer? my two cents. -lev On Sun, 28 Mar 1999 [EMAIL PROTECTED] wrote: [snip] However, can I ask people what they would use for music composition, accounting and personal finance? I'm aware of good programs for creating musical scores which can also generate MIDI output, but I'd hardly call them top-flight composition tools; and it does seem that the accounting/finance area is thinly served. Now that vmware is out ( http://www.vmware.com ) people who want to
LILO non-BIOS scsi card
hello, OK, in my attempt to get my system back to normal after ditching Red Hat, I'd like to iron out some problems with LILO. First an explanation of my system: IDE hard disk being used as /boot, and /home SCSI hard disk being used as /root SCSI host adapter that is ISA, and does not work without drivers (software)...in other words, I cannot boot up on a SCSI drive. In Red Hat, I had no problem making it so that the IDE disk had all the stuff necessary to load up the drivers for the SCSI drive, and after that, everythign would go as normal. However, in Debian, no matter what I did in the setup, I couldn't get it to tel me that the IDE drive would be the first to boot up. It would always default to the SCSI disk. So now I'm stuck using te floppy whenever I need to boot (not really a big issue as I keep it on all the time). can anyone give me some advice on how to fix this? thanks!!! -lev BTW, the problem with the Xwindows thing starting only as root had to do with an odd thing in the /etc/X11/Xserver file...it seems that it had a whole bunch of garbage in it after the initial '/usr/bin/X11/XF86_SVGA' line...it looked like text from xf86config program...anywho, I deleted all of it, and put 'Console' below the first line, and it worked fine.
Re: What DO you lose with Linux ???
As far as physical space goes, an actual computer case can get pretty small...My curiosity was piqued with Linux especially becasue I could simply run X over my network here @ home, no need for a video card even. Hell, if I weren't runnig a scsi system, a bare-bones install would only require a motherboard, cpu (celeron), and an HD...maybe a cdrom..hell, even that and the HD could be done away with if you run it through NFS (am I stretching it here? ;). trust me, I know what it's like to be 'po (UCLA, need I say more)I thikn though, that for me learning how to use *nix is far more profitable in the long run (read: it's good on a resume). -lev On Sun, 28 Mar 1999, Kent West wrote: Lev Lvovsky wrote: the question is tho, other than electricity consumption, what's so wrong with running more than one computer? 1) Up-front cash outlay for the second computer (important to poor folks like myself) 2) Physical desk space. Other than that, I'm all in favor of having two or three computers running at the same time. Just my 1 cent (I'm too poor to pay attention, much less give away 2 cents) :)