Re: remount removeable drive in Lenny - how? [solved]
On Sun, Jul 19, 2009 at 01:11:47PM -0600, Paul E Condon wrote: On 2009-07-19_18:57:40, Andrei Popescu wrote: On Sun,19.Jul.09, 08:11:21, Paul E Condon wrote: I have no objection to the status of hal as a required part of a standard desktop installation, but I do have a question as to how best to deal with a peculiar situation. I have several USB hard drives (ones with rotating machinery inside, not solid state 'disks'). From time to time I need to perform format maintenance on one of them. In order to do this, I look in /dev to see what device name has been assigned to the drive, umount it, and do whatever - e2fsck, tune2fs, etc. But when I'm finish doing maintenance, how to I remount it without pulling the USB cable, waiting a while, and reinserting the cable? Is there a console command that I can type that avoids the extra wear on the fragile little connectors and plugs? I'm looking for something that retriggers the look-up of volume label and the creation of a mount-point in /media as was there before I started mucking about. This thread seems interesting Correction: This thread *is* interesting http://www.linuxquestions.org/questions/linux-software-2/re-detectmount-usb-hard-drive-623089/ (the first hit when I googled: hal redetect devices) The solution is in the parted package. The kernel finds devices in paralell so the device order is random, udev fixes this with UUIDs and LABELs Now, hal is deemed essential to Xorg and it messes up simple cases. now we need parted to solve the problem of hal which was to solve the problem of udev, which was to solve the kernel problem. argh. Of course, all these extra packages and systems will have general bugs and the possibility of security bugs. Did somebody call this progress? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: problems during installation
On Fri, Jul 17, 2009 at 03:33:37PM +1000, w0102926 wrote: Hi I attempted to install debian GNU/Linux 5.0 several times today, but every time the process gets to select and install software, nothing appears to happen, I have left it for up to an hour with please wait 1% complete on the screen. The rest of the process happens instantly which leads to the conclusion something isn't right. Any help would be greatly appreciated We need more info. What CD are you using? Are you installing the software from the network or the CD? Personally, I always just install the base system (or the standard system) with no DTE or other stuff, from the CD. Then I boot up the system, get networking going and add software later. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: DVD download
On Sun, Jul 05, 2009 at 04:42:46PM -0400, Mark Neidorff wrote: This concerns the 5.02 amd64 DVD #1 Yesterday I downloaded the debian-502-amd64-DVD-1.iso file from debian.org's site. It was 4.4 Gb, but it failed sha1sum verification, so today I again downloaded it, but it is now 2.0 Gb and again fails sha1sum. I have now downloaded the same file-- debian-502.amd64-DVD-1.iso -- from 2 mirrors and each time the file is 2.0 Gb in size. So, I'm guessing that something has gone wrong at Debian and the errors are mirrored to the other sites? Should I even bother downloading the DVD image now or send a report to debian? If I should send a report, who should I send it to? Try using rsync to an rsync mirror. I'm on dialup and often find that after a week of getting a CD a bit at a time, that there are some errors. Rsync fixes it. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
chown question
Hello all, I have a really basic question; I really messed up my box. I was doing a reinstall on an old box after a drive failure. I restored /home but one of the UIDs were created differently so I needed to chown their directory, including all the hidden files in their ~/. Without thinking, as root I did a: r...@plot:/home/dtbrowser# chown -R dtbrowser.dtbrowser .* Unfortunatly, no everyting on the box is owned by dtbrowser. It walked up the file tree (presumably via . and ..) and changed everything. I know that I could have used find to look for all files owned by the old UID, plunked it through xargs and chowned them that way, but is there a way, as root, to chown directly the hidden files without chowning the whole box? Just for my future reference? Thanks, Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: OT: launching jobs in a combined serial parallel way
On Wed, Jun 24, 2009 at 10:05:20PM -0400, Scott Gifford wrote: Douglas A. Tutty dtu...@vianet.ca writes: On Mon, Jun 22, 2009 at 08:17:44PM -0400, Kamaraju S Kusumanchi wrote: While you may think its terribly inefficient, it isn't really. A fancy wait function is just polling anyway, you're just making it overt. Just to clarify, wait(2) and the shell wait builtin do not poll, they instruct the kernel to put the process to sleep until a child process finishes, then wake it up and return from the wait call. Because of that they are very efficient. Still, if the programs in question are doing a great deal of work, the extra work required by polling will not be very significant in comparison. Isn't that just passing the polling on to the kernel? At some level, some process has to see if a pid exists, if not, wait a period, then check again. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: is it possible to install a desktop-manager without python and perl?
On Mon, Jun 22, 2009 at 09:49:26PM -0500, Cybe R. Wizard wrote: Although there have been attempts to design one universal computer language that serves all purposes, all of them have failed to be generally accepted as filling this role. Ada does a good job. Except that since no OS is written in it, to get OS system calls, you have to use the underlying system libraries, which are usually in C. This is simple in Ada, but its still mixed-language. Of course, if you care to, even on i386 I guess you could start from scratch and build an embedded system without an underlying OS, all in Ada. However, there would be little point. On the other hand, if you're designing a new air traffic control system for a country (e.g. Canada), you write it in Ada from scratch with no underlying OS, no other libraries. Ditto if you're building a new fly-by-wire system for an airliner. The US military, before they decided to go C.O.T.S. specified Ada for everything. They wanted one language that could do everything, and they got it. The problem for the rest of us is that people were already comfortable with all the rules that they could break with C. For a commercial company that earns money fixing its own bugs, it doesn't make commercial sense to retrain everyone in Ada and retool in Ada, only to inadvertanly write software with fewer bugs (and what bugs there are, easier to fix). Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: OT: launching jobs in a combined serial parallel way
On Mon, Jun 22, 2009 at 08:17:44PM -0400, Kamaraju S Kusumanchi wrote: Currently I have a shell script that works as below. 1) launch proga, progb in the background using nohup. 2) Ask proga, progb to write a file when they finish. 3) Every five minutes check if these files are present. If they are present, launch progc. This gets me going for now. But it looks terribly inefficient. I would appreciate if someone can provide a better solution. While you may think its terribly inefficient, it isn't really. A fancy wait function is just polling anyway, you're just making it overt. You also have the ability to have proga and progb only touch the file if they complete successfully. If you merely wait until their ps disappears, you don't know if they crashed or properly completed. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Backup config files in home directory
On Sun, Jun 21, 2009 at 01:01:52AM +0800, ronggui wong wrote: I have other files and directories in the home directory, and I just want to backup all the config files, most of them are hidden files and directories. Now I use tar and manually exclude my other files and directories with --exclude argument. Is there a better way to do this? You'd have to define 'better'. A tar archive is certanly easy to restore. You can even just pull in one file out of an archive. You can look into the archive interactivly with midnight commander. Personally, I use tar for all my backups. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: any substitute for x window system?
On Mon, Jun 22, 2009 at 11:58:22AM +0800, 明覺 wrote: I'm looking for a pure c/c++ programmed desktop manager, while the xorg is depandent on perl, so i do not like it, is there any graphics system which depands only on c/c++ to replace x window system? thanks I think that you'll find that you need to start writing things from scratch yourself. Since debian requires perl (e.g for debconf), you'll be better off with NetBSD. Then, write a program in C that looks at every non-binary file to see if what language its in. Can you tolerate shell scripts? If not, you'll have to write a C-based initscript. This may be easier on BSD since it doesn't use SysVinit. Take away the ideological furvor. It would be an excellent learning experience to rewrite, from scratch, everything in NetBSD that is not C. It would be very hard with Debian since every time you update, you'll have to do it all over again. Enjoy. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Filesystem UncorrectableError for IDE disk on 2.6.26?
On Fri, Jun 19, 2009 at 09:11:07AM +, Glyn Astill wrote:
Just upgraded an etch machine to lenny, and with it I've gone from the
2.6.18-6 kernel up to 2.6.26-1 kernel.
I'm seeing the following errors under heavy disk activity:
Jun 18 18:08:05 xglyn2 kernel: [14214.048193] ide: failed opcode was:
unknown Jun 18 18:08:11 xglyn2 kernel: [14220.846157] hdc: dma_intr:
status=0x51 { DriveReady SeekComplete Error } Jun 18 18:08:11 xglyn2
kernel: [14220.846157] hdc: dma_intr: error=0x40 { UncorrectableError
}, LBAsect=26482377, sector=26482370
Previous to using 2.6.26-1 I did try 2.6.26-2 (both 686) and that gave
me tons of these errors upon boot up. I thought going back to the
2.6.26-1 version had resolved the issue but it's obviously just less
severe.
Anyone seen similar or got any ideas?
btw, the heavy disk activity is loading 80GB postgres database - I've
switched back to 2.6.18 to try this again but I suspect it'll be fine.
If you get the same errors if you boot again with the 2.6.18 kernel,
then I'd say the disk was dying. However, if you don't get the errors
with the the old kernel but do with the new, then the new kernel isn't
working with either the disk or the IDE controller. If the latter,
bummer. Since I run old hardware, I hate upgrading to a new kernel
because it likely means that more of my hardware has to be migrated away
from linux to one of the BSDs.
Doug.
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: exim4 query
On Wed, Jun 17, 2009 at 12:10:17PM +0100, Yuriy Kuznetsov wrote: I'm trying to figure out how to use exim4 for sending/receiving mails in console mode. Any recommendation for good instructions/how-tos ? I think that the only way to dirctly tell exim4 to send mail is to speak SMTP to it, since it is a Mail Transfer Agent. Most people use a Mail User Agent to talk to it, e.g. mutt. Receiving mail, for most people, is done with a mail fetcher, e.g. fetchmail, that pulls the mail from a POP or IMAP server, hands it off to your local exim4 MTA, which puts it in your mailbox. You then read the mail with your MUA, e.g. mutt. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Automatically creating user accounts from exim
On Wed, Jun 17, 2009 at 03:26:41PM +0200, David wrote: Okay, this is kind of a weird question, but it came up at work. I'm a complete exim newbie (I've never configured it before, beyond 'dpkg-reconfigure exim4-config'), but a project came up where the manager wants to use exim in a weird way. Basically, this needs to happen: 1) Exim receives a mail, from a trusted IP address 2) If the mail is to a non-existant user account, then create the system account, deliver the mail to the new account's mail file Since this mail is coming from a trusted server, why not have a script on that server first check (via ssh) if the user exists? Or, have it send the mail blindly. If the user doesn't exist, exim bounces it back. the sending script then uses ssh to create the user on the target system. 3) And always, after delivering a mail (for new or existing users): Call an external script, so that our custom logic can see the new mails immediately after they appear, and do some further handling. Are you sure that email is the best route at all for this traffic? Mail to non-existant user so that a script on a remote box can read the mail? Why not just rsync (or scp) over ssh a file containing the information? Or, have programmes at each end running with a socket between them? Or use have the target script put the output to stdout, pipe it through ssh to the receiving script taking it from stdin via a pipe from ssh? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: exim4 query
On Wed, Jun 17, 2009 at 10:52:18AM -0500, Boyd Stephen Smith Jr. wrote: In 3df35b760906170818re166e28x8be9006d7...@mail.gmail.com, Yuriy Kuznetsov wrote: Could you give some commands as examples of sending emails with exim4, please? Install bsd-mailx. echo 'Body of the message' | mailx -s 'Subject Line' 'to#1' 'to#2' OR printf 'dot-stuffed-Headers\n\ndot-stuffed-Body of message' | \ /usr/sbin/sendmail This may be what he meant, but isn't that for which he asked. To answer Yuriy's question, I don't know enough SMTP to be able to talk to exim directly to send a mail with exim4. mailx works well for scripts or short emails from the command line. For normal everyday use, install mutt. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [OT] The perfect system ...
On Fri, Jun 12, 2009 at 08:52:05PM +0100, AG wrote: If I was to plan to build the perfect system from scratch, using a motherboard bundle (with up to 8GB DDR RAM), top line graphics card, ditto sound card that would allow connection with an external amp to jive up the sound quality - what might this thing of technological beauty be? Obsolete by the time you build it? Top-line sound card isn't a sound card at all. They are usually a card connected to an external box (better sound quality, all the analog stuff outside of the noisy computer case). Two years ago, I build a box: Asus M2M-SLI MB (built-in sound), 8 SATA ports, two gigabit ethernet ports, 10 USB, yada, yada. Athlon 3800+ 1 GB ram. never had the need to expand. Asus EN7300GT silent 256 MB video (nvidia). Two Seagate SATA drives. LG DVD burner. Hooked up to my 21 Intergraph drafting monitor. --- The sound was great through the stereo. The video was great with the nVidia kernel module (etch non-free) and the dvd codecs from debian-multimedia.org. VLC was able to have the hardware resize the image from a 1024x768 box to 1600x1...@75hz with hardware smoothing. In short, it did everything I wanted it to do. I had the 3800+ because the X2 wasn't quite available yet. Now everything has muliple cores, 8GB ram is common, terabytes of storage are near and dear to many (yet backing that much data up to something reliable is an issue, as is software raid sync time). Determine what you want the box to do. If its only watching movies, most new computers with a decent video card will do that as is and isn't anything special. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: flash drive encryption between different OS's
On Sat, Jun 06, 2009 at 10:30:14AM -0700, Kelly Clowers wrote: On Sat, Jun 6, 2009 at 07:35, Douglas A. Tutty dtu...@vianet.ca wrote: On Fri, Jun 05, 2009 at 03:58:33PM -0400, Daryl Styrk wrote: I'm looking for a solution to encrypt a flash drive formated in FAT32 so it can easily be decrypted across multiple OS's with the least amount of software needed. I don't do windows. Does it have OpenSSL? I encrypt stuff with that since the BSDs have it too. Just pipe it through. e.g. stuff.tgz becomes stuff.tgz.aes Is that secure? I mean, SSL is for streams, not blocks, but I guess I don't know how that affects security if used in the opposite manner... Well, the openssl has the encrypt function for encrypting files. AES is a block cypher anyway, even in a stream SSL breaks up the stream into blocks for encryption. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Best way of restoring /etc from backups after fresh install?
On Fri, Jun 05, 2009 at 05:57:26PM +, Laurentiu Pancescu wrote: snip I guess the automatically generated uids during the new installation were different from the ones in my backed up passwd/group files. What would be the best way to restore the full system in such a case? snip BTW, can anyone suggest a better backup approach? I normally follow testing, where new versions of packages appear often, so I would like to avoid backing up everything (/usr eats up too much space with incremental backups). Backing up /etc, /home and some files in /var would be ideal, since I can get the rest by installing, but how can I best restore the full system in such a case? I usually end up doing a full reinstall when there's a new major release number (e.g. 4.* to 5.*). Here's what I do. I'm on dialup: 1. A minimal install (don't select any tasks). Create my own username during the install. 2. Install mc, links2, ppp, pppconfig, exim4, and configure 3. Set up sources.list to use the correct mirror and security.debian.org 4. Run an aptitude update 5. Install security updates. 6. Restore my backups (which were created in /var/local/backup) back to /var/local/backup from backup media. 7. untar etc.tgz.aes (via pipe through openssl) to /root/restore 8. Inspect the old passwd file for username/UIDs to recreate 9. Recreate those users, ensure that the UID's match. 10. untar home/ from backup.tgz.aes 11. using the list of packages that were installed, reinstall in related chunks (so I get some packages installed rather than waiting 5 days to get _any_ packages installed). 12. Once I get all packages reinstalled, ln -s /root/restore to /home/dtutty/restore 13. (lucky 13), Have one VT with my own user reading ~/restore/etc/* and the other VT as root, compare each file and change as necessary. It only takes about an hour. I find that this gives me the advantage of getting the updated file (e.g. newer options, comments) but with my old config needs met. I hope this helps. Doug. P.S. Please cc me when replying, I'm not subscribed to debian-user (too high volume). Done -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: flash drive encryption between different OS's
On Fri, Jun 05, 2009 at 03:58:33PM -0400, Daryl Styrk wrote: I'm looking for a solution to encrypt a flash drive formated in FAT32 so it can easily be decrypted across multiple OS's with the least amount of software needed. I don't do windows. Does it have OpenSSL? I encrypt stuff with that since the BSDs have it too. Just pipe it through. e.g. stuff.tgz becomes stuff.tgz.aes Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: lvm on a single big partition or just a single big partition?
On Thu, Jun 04, 2009 at 07:23:19PM -0500, Zhengquan Zhang wrote: On Thu, Jun 04, 2009 at 09:39:42AM -0400, Douglas A. Tutty wrote: On Wed, Jun 03, 2009 at 10:04:36PM -0500, Zhengquan Zhang wrote: On Wed, Jun 03, 2009 at 06:48:03PM -0400, Douglas A. Tutty wrote: On Wed, Jun 03, 2009 at 11:44:00AM -0500, Zhengquan Zhang wrote: I put /tmp on tmpfs, with encrypted swap (so that /tmp ends up encrypted also). Yes, /usr (4G) , /var (4-6G, depending ), and /home (encrypted) are on separate LVs. Sizes depend on what I'm doing. /usr mostly holds instaled packages so 4G is fine for my desktop system. I also have /var/tmp and /var/local as separate LVs, encrypted. I keep my backups in /var/local. KDE keeps lots of otherwise private stuff in /var/tmp. Great setup, Thanks, Doug, so do you how much space(PE) do you leave unassigned for later use? Since I only have two 36 GB drives, I don't leave any space unassigned. My current box is an HP NetServer LPr, which has two hot-swap bays. The scsi drives are connected to an HP NetRaid card which is set up to present two logical drives to the OS: one is raid1 (for the system), and the other is raid0 (for /home and anything else that is routinely backed-up and wouldn't cause a system crash if it died). If I need more space, I'll get an external scsi enclosure and move the hot-swap drives to it (the NetRaid card only has one channel) and add more drives. Right now on this box I'm not acutally using LVM since I can do most of what it can do with the NetRaid card. I used LVM on my SATA-equipped Athlon box that didn't have a raid card. FYI, here's my df: FilesystemSize Used Avail Use% Mounted on /dev/sda1 477M 166M 311M 35% / tmpfs 506M 16K 506M 1% /lib/init/rw udev 10M 124K 9.9M 2% /dev tmpfs 506M 0 506M 0% /dev/shm /dev/mapper/sdb2_crypt 26G 2.7G 24G 11% /home /dev/sda3 2.4G 1.5G 863M 64% /usr /dev/sda5 1.4G 874M 548M 62% /var /dev/mapper/sda6_crypt 16G 4.4G 12G 28% /var/local /dev/mapper/sdb1_crypt 952M 16M 937M 2% /var/tmp tmpfs 500M 68K 500M 1% /tmp Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: lvm on a single big partition or just a single big partition?
On Wed, Jun 03, 2009 at 10:04:36PM -0500, Zhengquan Zhang wrote: On Wed, Jun 03, 2009 at 06:48:03PM -0400, Douglas A. Tutty wrote: On Wed, Jun 03, 2009 at 11:44:00AM -0500, Zhengquan Zhang wrote: With one big partition, you lose the ability to: - have a separate /var (or /var/log) to keep logs from filling up / - have different mount options (e.g. noexec, nodev) on /home - have a separate /home Without LVM, you lose the ability to : - resize partitions as needed - migrate data from one disk to another, e.g. if a drive starts misbehaving but you need to keep the system live rather than reinstalling/restoring. Could you elaborate more on this? As far as migration is concerned, what is the advantage of LVM? Your system is operating. You start to get either SMART indications that the drive is dying or errors (e.g. retries, times-out, etc) in syslog. You add a drive to the system at least as big as the failing drive. You make it a physical volume for LVM, and add it to the VG of the failing drive. You then tell LVM to remove the failing drive from the LV. LVM will migrate the data, extent by extent, from the old drive to the new drive, all while the system is still active. Its all in the LVM-HOWTO (tldp, in the doc-linux-html package). Instead of a separate /boot, I often use a separate / (which contains /boot). In this way, the / partition isn't part of LVM (I make it 500 MB and usually only have under 200 MB used) and can be booted into if the need arises, with more tools available than within the initrd. Most of my boxes won't boot a live CD. So I guess for /tmp /var /usr etc you have separate LVs? or else a 500M / should be too small? I put /tmp on tmpfs, with encrypted swap (so that /tmp ends up encrypted also). Yes, /usr (4G) , /var (4-6G, depending ), and /home (encrypted) are on separate LVs. Sizes depend on what I'm doing. /usr mostly holds instaled packages so 4G is fine for my desktop system. I also have /var/tmp and /var/local as separate LVs, encrypted. I keep my backups in /var/local. KDE keeps lots of otherwise private stuff in /var/tmp. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: lvm on a single big partition or just a single big partition?
On Wed, Jun 03, 2009 at 11:44:00AM -0500, Zhengquan Zhang wrote: Though I have used lvm for some time, I have one question that I don't understand. For one harddrive I often create a /boot parition that is not lvm and create a huge partition on the rest of the harddrive for PV of lvm. Now I am thinking what is the difference between doing partition like this and just a single big partition without lvm? With one big partition, you lose the ability to: - have a separate /var (or /var/log) to keep logs from filling up / - have different mount options (e.g. noexec, nodev) on /home - have a separate /home Without LVM, you lose the ability to : - resize partitions as needed - migrate data from one disk to another, e.g. if a drive starts misbehaving but you need to keep the system live rather than reinstalling/restoring. Instead of a separate /boot, I often use a separate / (which contains /boot). In this way, the / partition isn't part of LVM (I make it 500 MB and usually only have under 200 MB used) and can be booted into if the need arises, with more tools available than within the initrd. Most of my boxes won't boot a live CD. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: replacing hard disk
On Mon, Jun 01, 2009 at 04:43:12PM -0300, Claudio wrote: I think this tool *dd*, resolve your problem. http://www.linuxweblog.com/dd-image http://www.mckeay.net/2004/10/18/using-dd-to-clone-a-hd/ Or, you can use tar to create to stdout, pipe the output to another tar process to extract from stdin. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: ed 1.33 install errors
On Mon, Jun 01, 2009 at 09:13:47PM -0400, Rob Bochan wrote: On Monday 01 June 2009 08:53:19 pm Jude DaShiell wrote: Script started on Mon Jun 1 20:50:14 2009 localhost:~# aptitude install ed ... update-alternatives: error: alternative path /bin/ed doesn't exist. dpkg: error processing ed (--configure): subprocess installed post-installation script returned error exit status 2 Errors were encountered while processing: ed E: Sub-process /usr/bin/dpkg returned an error code (1) I got the same error here. I fixed it by creating a symlink from /usr/bin/ed to /bin/ and that cleared it up. I'd imagine it's a bug in the package. Considering that the main use of ed is either for scripted editing or for editing when ncurses (or gui) isn't working, the bug is probably having it in /usr/bin instead of /bin. In Lenny, its in /bin (at least on my box). I would argue that another bug is that its not statically linked but Debian doesn't seem to worry about having basic stuff static and refer people to a live CD for system rescue. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Realtek Device ffff (rev 10) is not recognized with 8139too
On Mon, Jun 01, 2009 at 03:43:21PM +, Marcelo Luiz de Laia wrote: Here are the outputs of some commands. What does dmesg show? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: replacing hard disk
On Mon, Jun 01, 2009 at 07:39:08PM +0200, Jan Willem Stumpel wrote: So tomorrow it's off to the computer shop to get a new hard disk. Are there any tips on moving the whole system from the old disk to the new one? Or do I just have to re-install ubuntu, re-install any updates and extra programs which are installed, find and copy modified config files, mails, bookmarks, etc? Moving the / partition to a new drive is very OS-specific. You may want to ask on a ubuntu mailing list. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: What hardware to use for Debian Firewall/Gateway or server?
On Fri, May 29, 2009 at 10:18:56PM +0200, Csanyi Pal wrote: Jan Willem Stumpel jstum...@planet.nl writes: Csanyi Pal wrote: So: can one install on it say a Debian GNU/Linux Lenny? Mind that it is a headless device. Everything has to be done through ssh (or local telnet). It has no cd-rom drive, keyboard, or monitor. But it is just a Debian system (for powerpc, not for i386). Everything behaves just like your desktop Debian system. Say I'll upgrade Debian Etch on it to Debian Lenny and make a mess of the operating system somehow through ssh connection. Then what can I do? And if one can setup the first boot media to USB stick how can install the system without to seeing anything? How can they install the Debian system in the Factory of the BUBBA?? Many headless embedded-type devices have a serial port as a bios/system console. Just use that to log in. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: konqueror, fish, stalled
On Sat, May 30, 2009 at 09:19:13PM +0800, Umarzuki Mochlis wrote: Whenever I tried to copy file(s) bigger than 900 MB using fish protocol, it stalled. Any way to prevent this? Try a different protocol. Use rsync or scp (or mc with shell link) from a command line. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: hylafax
On Thu, May 28, 2009 at 10:08:37PM -0400, Tom Low-Shang wrote: Is Hylafax still the only open source fax server available? Last time I did fax, I used mgetty+sendfax. It worked just fine. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Lenny kpdf much slower than on etch
Hello all, I'm finding that kpdf is much slower on Lenny than on Etch. When I load a new doc, it takes forever to generate the first page (and the thumbnails). Is there some setting I can change somewhere? This is on my dual-P-II-450. It takes 100% of a CPU for about 10 seconds before I can start to read a 10 page document. Thanks, Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Operating system-level virtualization: how to make it?
On Wed, May 27, 2009 at 09:39:38AM -0500, Victor Padro wrote: On Wed, May 27, 2009 at 8:40 AM, Douglas A. Tutty dtu...@vianet.ca wrote: On Tue, May 26, 2009 at 08:46:49PM +0200, Laurent Guignard wrote: On Fri, 22 May 2009 18:02:27 +, Sylvain Le Gall wrote: On 22-05-2009, Sthu Deus sthu.d...@gmail.com wrote: AFAIK, virtualization on i386/amd64, beyond the os-specific software or testing issues, is a gimmick. It may provide one extra layer for someone to try to break out of but it also adds an extra layer to hold bugs. There is nothing like LPAR in x86/amd64 architecture. Totally different arch. Believe me I work for the eye bee m company. That was my point. Unless the hardware provides the virtualization (such as LPARs), then it doesn't accomplish much. doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: KDE is now broken (Fwd: Heads-up: KDE4 hitting testing tonight (UTC) )
On Wed, May 27, 2009 at 11:18:54AM -0500, lee wrote: On Tue, May 26, 2009 at 10:46:47AM -0500, Boyd Stephen Smith Jr. wrote: In 20090526142918.gc5...@cat.rubenette.is-a-geek.com, lee wrote: On Mon, May 25, 2009 at 01:17:16PM -0500, Boyd Stephen Smith Jr. wrote: Use the old software. It might not run on the latest release of Debian, but it should run on whatever version you had before. Older releases are maintained in the archive, and you can archive whatever you need yourself if you don't want to depend on the Debian infrastructure. How do you get it to run on contemporary hardware? Run it on the hardware you were running it on before. We are talking about accessing the data for the purpose of migration; you should still have the hardware (and software) you are migrating from. 1.) When I'm changing hardware, I'm usually replacing board, CPU and RAM. I take that out of the case and put the new stuff in. That means I can't run the software on the old hardware anymore, not without changing the stuff out again. It would really suck if I had to do that. If your backup/archive hardware won't connect to the new computer, then you'll need to migrate the data to new archive hardware first, then migrate the computer to new hardware. 2.) I'm not so much talking about migration as about keeping data readable. Keep it on your disks or put it aside on some removable storage medium, then after 15 or 30 years, try to read it. Having used a mysql database to store the data doesn't make it easier to read it after 15 or 30 years. As far as I know, the only digital media that is designed to last that long on the shelf without data loss is tape. Since tape technology moves apace, you should probably archive a couple of tape drives along with the data. Find your local LUG and ask around. I can virtually guarantee that there someone with a storage unit full of old hardware they are keeping for some reason. Even better if you have a local FreeGeek. Who would keep all the old hardware? And for what? And it's nothing you could rely on. Actually, I need old hardware. Newer hardware gives my wife headaches. It varies, though. Usually, she can tolerate my NetServer LPr dual P-II-450, but right now its a problem. I'll get my 486 out of storage and put NetBSD on it and see how it is for her. Of course, its ISA and won't boot if there's a drive bigger than 1.2 GB on either IDE controller. What I really need is an old 100 MHz or slower SMP server with scsi. Or, at least, an ISA scsi card. How do you maintain 15 or 30 year old hardware? Carefully. Memory is still available for my 486. The biggest problem for me is hard drives; they die and aren't made small anymore. Scsi fixes that (since there are no bios issues with scsi). And who guarantees that 30 year old hardware you kept in storage will still work when you need it? You do. No, I don't. I have no way to do that. I didn't manufacture it. I can only assume that it might work or not after 30 years. If what you're saying is practical for you, go ahead and keep your pile of hardware over 30 years or longer and try to put something together to read your data when you need to. That isn't practical for me. Choose hardware in the first place that allows upgrade. E.g. scsi drives instead of IDE. Also, do you really need the data to sit on a shelf for 30 years, or can it be cycled to new media every 5 years? There's something to be said for tarring to a raw disk partition (so there's no filesystem to be corrupted), and putting the same data to three different drives. Then using some data comparision utility (there's a deb available, I forget the name) to choose the correct block for every block of the data. This is far more reliable given three partially corrupted data sets than e.g. raid where if a certain number of blocks fail, the whole disk is marked bad. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: KDE is now broken (Fwd: Heads-up: KDE4 hitting testing tonight (UTC) )
On Wed, May 27, 2009 at 01:51:25PM -0500, lee wrote: On Wed, May 27, 2009 at 09:21:15PM +0300, Dotan Cohen wrote: If the issue is saving $20 on a case, then you are just $20 short of having a working solution. Sounds good to me. Where do you get a good case for $20? Shop around a bit and you'll find that useable ones start at about $250, and good ones cost more --- if you can find one at all. Buy an old computer and use that case. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: KDE is now broken (Fwd: Heads-up: KDE4 hitting testing tonight (UTC) )
On Thu, May 28, 2009 at 10:08:00AM -0500, lee wrote: On Thu, May 28, 2009 at 09:42:57AM -0400, Douglas A. Tutty wrote: As far as I know, the only digital media that is designed to last that long on the shelf without data loss is tape. Since tape technology moves apace, you should probably archive a couple of tape drives along with the data. Well, I used to have some tape drives when the disks were smaller, but I haven't looked into them for years. I always bought them used, and they were still rather expensive. Can you get a reliable tape drive, incl. some tapes, that stores at least 1TB per tape, for max. $200 now? No. You can get used LTO-3 drives for about $250 on ebay. Tapes, of whatever capacity, end up costing about $50 each (I suppose unless you buy in great volume). You have to do the math once you determine your data set size to know whether its cheaper to have a lower-capacity drive with more tapes or a higher-capacity drive with fewer tapes. I just went through the math for my 12 GB data-set size and compared tape to USB sticks (not for archive, but for off-site backup). The break-even point was 72GB: less than that and USB sticks (16 GB) were cheaper; more than that and a used LTO-3 tape changer was cheaper. Who would keep all the old hardware? And for what? And it's nothing you could rely on. Actually, I need old hardware. Newer hardware gives my wife headaches. Why is that? It limits her to very slow hardware which could be a problem sooner or later because the old hardware isn't up to the task anymore. It depends on what she wants to do, of course ... Whether it limits her (actually me) or not is not the issue. If it gives her a headache then I can't use it at all, which is very limiting. Other than web-browsing (e.g. firefox/iceweasel), I can do everything on my 486 if I have enough drives. controller. What I really need is an old 100 MHz or slower SMP server with scsi. Or, at least, an ISA scsi card. You can still get that --- but for how long? How do you maintain 15 or 30 year old hardware? Carefully. Memory is still available for my 486. The biggest problem for me is hard drives; they die and aren't made small anymore. Scsi fixes that (since there are no bios issues with scsi). Yeah, SCSI is great, but it's not affordable anymore. What's the price for a 1TB SCSI disk now? Like $1500? And I'd need two because I don't put data on disks that aren't at least RAID1. I've seen too many disks failing for that. Look at the sweet-spot price point for used scsi drives, then get a used hardware raid card. My HP NetServer LPr (dual P-II-450) came with 1 GB ram, two 36 GB scsi drives, and a HP NetRaid 1si card for $65 (and all cables, terminators, etc). You can get the raid cards cheap if used. Get a 14-bay external scsi hot-swap enclosure to hook up to it (another $50) and load it up with the scsi drives (of whatever size). Of course, if you have to pay for power, I guess at some point its cheaper to buy two 1TB drives. And look at the cables and terminators. You end up paying about $100 just to connect a few SCSI disks. I still have the controller and disks, but the cables got lost when moving. They'd be nice to have, though rather loud, but I didn't want to spend all the money on cables. You can get 1TB SATA disks for the price of the SCSI cables and the terminators ... And forgo the reliability. You really comparing SATA to SCSI? SAS to SCSI sure; SATA to IDE sure. Check ebay for the cables and terminators. There are lots of computer recyclers that only sell through ebay. In your case, you could have the computer for your wife boot over the network and run it without any disks. Put the sever for that at some place where it doesn't bother her. That would be about 500 feet away. First, I'd have to buy a lot that's big enough, then build a data centre 500 feet away... Also, do you really need the data to sit on a shelf for 30 years, or can it be cycled to new media every 5 years? I keep it on the disks. I don't have a solution for making backups anymore. I'm going to need new disks soon, and I'm also going to need a second set for backups (still better than none) --- but I don't have the money for that atm. There's something to be said for tarring to a raw disk partition (so there's no filesystem to be corrupted), and putting the same data to three different drives. Then using some data comparision utility (there's a deb available, I forget the name) to choose the correct block for every block of the data. This is far more reliable given three partially corrupted data sets than e.g. raid where if a certain number of blocks fail, the whole disk is marked bad. You want to buy 8 sets of 3 disks each for your dayly and monthly backups and an SATA controller that can do hotplug? That's about $2500 --- maybe you can get a tape drive for that kind of money
Re: Operating system-level virtualization: how to make it?
On Tue, May 26, 2009 at 08:46:49PM +0200, Laurent Guignard wrote: On Fri, 22 May 2009 18:02:27 +, Sylvain Le Gall wrote: On 22-05-2009, Sthu Deus sthu.d...@gmail.com wrote: How I can organize a Operating system-level virtualization on a server for every service I would isolate? Use a chroot (standard) or a vserver (search for vserver in debian archives there is a kernel version and two packages for userland tools). vserver is more flexible and allow you to assign IP address et al. Beyond the question, what is the interest to virtualize services. I understand the need to virtualize different machine for OS specific server software, tests and so on. Is there anywhere to find when virtualization is the best way to solve a problem and when it isn't ? Unless something has changed, to be really secure, virtualization has to be fully supported in the hardware of the CPU so that there are no CPU instructions that can be issued from within the virtual machine to break out of it. i386/amd64 don't meet that criteria. I don't know what other vendors have, but e.g. IBM's Power architecture does, and provides logical partitions (LPARs) at the firmware level which appear to the OS as a real piece of hardware. AFAIK, virtualization on i386/amd64, beyond the os-specific software or testing issues, is a gimmick. It may provide one extra layer for someone to try to break out of but it also adds an extra layer to hold bugs. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Luks encrypted partition gets identified as ntfs
On Wed, May 27, 2009 at 06:52:20AM +0100, Aron wrote: About 4 years of research I have in there so far got that gut feeling it's going up in smokes. I surely hope you have backups, either not encrypted or encrypted with something else (I use openssl). Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Lenny. Do I need to check the system after improper shutdown?
On Mon, May 25, 2009 at 11:33:31PM +0400, Mark Goldshtein wrote: Hello, list! Do I need to clean up something or check hard drive consistency after system's hang up during recovering from Suspend-to-RAM state? An improper system shutdown by 'power' key was forcibly applied and during a boot process were reports about journal transactions replayed. That was it. The journal replayed with no problems, so everything is consistant. However, you may have lost data, but fsck can't help that; its job is to make the filesystem consistant. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: What hardware to use for Debian Firewall/Gateway or server?
On Thu, May 21, 2009 at 09:31:14PM +0200, Csanyi Pal wrote: I have at my home a small network: firewall/gateway: Pentium II Class PC box with 64 MB RAM, 5,1 GB HDD server : Pentium IV Class PC box with 2 GB RAM, 60 GB HDD desktop : Pentium IV Class PC box with 2 GB RAM, 2 * 320 GB HDD On all these PC boxes run Debian GNU/Linux: firewall/gateway: Etch Server : Etch desktop : Lenny The firewall has a buggy hardware and can't to install on it Lenny so I decide to buy a new hardware for firewall/gateway. Put the drive in another computer, install to that drive, then move the drive back? I think about that that I could to use the server box as a firewall/gateway and the new PC box for the server.. What is the recommended new hardware for firewall/gateway or for a web, mail, file printer server at a small home network? Well, since the PII worked just fine, I think you'll find that any computer on which you can install Lenny will work for you. I used to use a 486 with 32 MB ram but Etch couldn't install on it. It runs OpenBSD very well. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
how capture segfault trace for bug report; lenny megaraid driver?
Hello all, I use an HP NetRaid 1si raid card in my HP NetServer LPr. It worked fine on Etch and was able to retreive status info from /proc/megaraid, e.g. # cat /proc/megaraid/hba0/raiddrives-0-9 With Lenny, there are a couple of problems: 1. Rescue mode doesn't see the drives even though they show up in dmesg. 2. When I cat any info in the /proc/megaraid/ tree, the first time, I get a segfault. A second time just hangs. The process is immortal (it doesn't even get killed by init on shutdown). I'd like to get this fixed, or at least submit a bug report. However, how do I capture the segfault info (I don't have a serial console at this time, but I suppose that's a possibility). I've tried: 1. ulimit -c 10; cd /var/tmp/; cat /proc/megaraid no core file anywhere that I can find 2. script to megaraiderr file; cat /proc/megaraid 21 The script just gives the segfault warning, but all the messages are lost. Nothing is logged to syslog. Questions: 1. How do I capture the info? 2. To whome do I submit a bug report? Thanks, Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: sudo vs. su (was Re: new to list, new to debian, new to linux)
On Fri, May 22, 2009 at 05:50:28PM -0500, dwain wrote: On Fri, May 22, 2009 at 2:16 PM, George nutn...@comcast.net wrote: I was a little disappointed being called out on my suggestions in my original post. Obviously the person isn’t a sys admin and from my understanding the whole purpose of sudo is so the user only has root privileges for that given command instead of during the entire terminal session. I personally see nothing wrong with what I suggested other then using visudo to edit the sudoers file instead of vim. i too was disappointed. i tried su and authentication failed. how do i update my system without being able to log in as root? now i don't mind a spirited discussion on the pros and cons of sudo vs. su, but my original question still has not been answered; and with this new development i am really at a loss. you are saying that sudo and su are not available from a user console, then how do i fix this so i can become root when i need to? Root login from a secure serial console has been described as the ultimate command line of last resort. I always have a serial console set up in inittab (and in grub too for that matter). I guess if you can't log in as root (or otherwise get root), you'll need to boot a live CD such as grml and fix whatever is preventing you from getting root. A last resort would be: 1. physically disconnect the box from the network. 2. boot a live CD 3. edit the password file to allow root login without a password 4. reboot into the system and log in as root 5. passwd as root and give yourself a root password. 6. shutdown 7. reconnect the box to the network. 8. carry on as normal. If you want to use sudo for most things, but have root login available, put the root password in a card in an envelope in a locked location. You'll know if someone needed the root password by the tear in the envelope. Unless it was removed, the root login should be recorded in syslog as well. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Convert HTML to PDF from CLI?
On Tue, May 19, 2009 at 01:15:25PM +0300, Dotan Cohen wrote: 2009/5/19 Dotan Cohen dotanco...@gmail.com: I don't know if can handle UTF-8, but I see nobody mention htmldoc, which is in debian etch, I suppose must be in lenny, so another option to try. I did not know about htmldoc, that is a great program! It does not support UTF8 as packaged for Debian, however, according to wikipedia version 1.9 does support UTF8 and I will try to install that now. Scratch that, even with basic UTF8 support, Arabic and Hebrew are not supported: http://www.htmldoc.org/articles.php?L28 I'm just doing an aptitude search (~dhtml ~dpdf !doc), trying to see if there's something not considered before in this thread. I've never used OpenOffice, but I see unoconv. Its supposed to be able to convert from any file OO can import to any file it can export. Can OO import html and export pdf, with your UTF requirements? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: LVM how to
On Wed, May 20, 2009 at 07:17:48AM +1000, gianni wrote: how can I resize the LVM default partition layout of debian? the root is to small around 400mb... I looked around the web but it look like I need to do that from a rescue cd, which one should I use? any good link for a easy how to :) No rescue CD required. Install (if you can) the doc-linux-HOWTO package, or go to tldp.org and read the LVM HOWTO. It has examples of how to do most things. Then read the man pages for each command. Basically, you have to shrink the file system size (resize), then shrink (resize) the LV its on. Then grow (resize) the target LV, then grown (resize) the target filesystem. Do a backup first, please. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Hardware diagnostics
On Tue, May 19, 2009 at 11:37:51PM -0400, Scott Gifford wrote: I have a Debian Etch installation that's beoming increasingly unstable. It periodically freezes up, with nothing in the logs until it is rebooted. I suspect a hardware problem, and would like to identify it or rule it out before doing an upgrade to Lenny. Can anybody recommend a good hardware diagnostic or burn-in program? I have used memtest86 and will try that, but ideally I'd like to stress test more of the system than just the memory. Something that can run on Debian Etch while the machine is live is ideal, or something that can be run from a boot CD. Free is preferred (of course), but any suggestions are welcome. Also, if anybody has a suggestion of what might fix an Etch system that's freezing up periodically with nothing in the logs, those suggestions are welcome too. :-) Apart from diags for specific hardware (e.g. my HP NetServer LPr diagnostic disk), I use GRML (grml.org) 0.9 CD. I run bad blocks (or the appropriate fs checker with badblocks read/write/verify check) on all filesystems. grml also has memtest+ as a boot option (it can't run properly with an OS running as well). Have the kernel do verbose logging. Consider remote logging; if your hard drive freezes, there's no way for the log to be written. Any serious drive errors should be sent to the console unless you've told the kernel to not send messages to the console; I guess you won't see them if you are in X at the time; consider a serial console to another box (or a real VT). Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Convert HTML to PDF from CLI?
On Sun, May 17, 2009 at 06:37:45PM +, Andrew Malcolmson wrote: On Mon, May 11, 2009 at 12:30 PM, Dotan Cohen dotanco...@gmail.com wrote: I need to convert an HTML document to PDF from the CLI. Currently, I ... Any other ideas? Is there a konqueror- or KDE way to do this? Am I missing something obvious? Thanks! Konqueror wouldn't be from the command line. If you are willing to use konqueror, then just 'print' it to a pdf file. It works great. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: failed to upgrade to next kernel package
On Sun, May 17, 2009 at 07:04:11PM -0400, Patrick Wiseman wrote: [Replying to debian-user, including OP's reply to me, which was presumably intended for debian-user.] On Sun, May 17, 2009 at 6:32 PM, gianni giovanni.favor...@gmail.com wrote: Hi Patrick this is the result from df -h Filesystem Size Used Avail Use% Mounted on /dev/mapper/machina-root 322M 272M 34M 90% / tmpfs 1.5G 0 1.5G 0% /lib/init/rw udev 10M 128K 9.9M 2% /dev tmpfs 1.5G 0 1.5G 0% /dev/shm /dev/sda1 228M 25M 191M 12% /boot /dev/mapper/machina-home 136G 99G 32G 76% /home /dev/mapper/machina-usr 4.6G 3.1G 1.4G 70% /usr /dev/mapper/machina-var 2.8G 1.3G 1.4G 49% /var tmpfs 1.5G 20K 1.5G 1% /tmp the system is only 1 week old... I used the default option with the LVM, what should I delete? You have a very small root partition (322M) apparently, which is almost full, and which is where /lib resides (as you have no separate mapping for it). I'm not sure what to suggest at this point (which is why these conversations should stay on the list; others may have all kinds of partition magic they can suggest, perhaps to expand the root partition while preserving your others, etc.). Do you have a kernel that you are not using to boot which you could remove, e.g. a -1 if you are trying to install a -2 version kernel? I have a 477 MB / partition, of which 166 MB is used and I have both 2.6.26-1-686 and 2.6.26-2-686 kernels installed. You should probably see what is taking up so much space in / since you have separate /home, /var, /usr, with /tmp on tmpfs. FYI, my /boot only has 14 MB in it (two kernels, plus the grub stuff). If there is nothing extraneous in / (including stuff in /root that shouldn't be there), since this is LVM, can't you resize the partitions? Assuming that you don't have any free space in the machina VG, take 500 MB from machine-home and add it to machina-root, taking care to do whatever filesystem resizing is necessary (depending on what filesystem type you are using). Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Safe change of uid
On Mon, May 11, 2009 at 05:27:14PM -0500, Elmer E. Dow wrote: Douglas A. Tutty wrote: On Mon, May 11, 2009 at 08:54:47AM -0500, Elmer E. Dow wrote: Please cc me as I am not currently subscribing to the list. Since you only just created the user, I'd just go ahead and delete it (use: # cd /var/tmp # deluser --remove-all-files --backup then use adduser to create the new user To be safe, I'd then examine the backup tarball to ensure that nothing was removed accidentally, before deleting the tarball. From what I've read, it's like deluser in that it only deletes the user stuff in the home directory, so I still need to edit /etc/passwd, /etc/group, /etc/shadow, etc. and delete the user's group. Or is there a command to take care of those, too? If there is, I haven't found it yet. Or will adduser overwrite the previous info when I add the new user of the same name? I've read of permission problems, boot problems, etc. caused by changing uid so I'm a bit paranoid. If you use deluser, you won't have to edit other files. That's what deluser is for. If you find any files in the backup that the user will need, you can untar the tarball, chown them to the new user, and give them back. I tend to use Midnight Commander (mc) for that type of thing. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Safe change of uid
On Mon, May 11, 2009 at 08:54:47AM -0500, Elmer E. Dow wrote: Just finished a fresh Lenny install and added an account for my daughter -- and kuser assigned it uid 500 instead of 1001, which I must correct. After looking at man pages and archives, I see that kuser in the past has done well creating accounts but not modifying them. Is that still true with the version used in Lenny? Is usermod a better option for dealing with this situation or would deleting and recreating the account -- either using kuser or userdel -- be the simplest and best method? Please cc me as I am not currently subscribing to the list. Since you only just created the user, I'd just go ahead and delete it (use: # cd /var/tmp # deluser --remove-all-files --backup then use adduser to create the new user To be safe, I'd then examine the backup tarball to ensure that nothing was removed accidentally, before deleting the tarball. I've never used (or heard of) kuser to know why it created uid 500. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: filter utility to send cron e-mail only if unexpected output
On Wed, May 06, 2009 at 11:27:55AM -0400, Barclay, Daniel wrote: Does Debian have any utility to address the following situation? I have some scripts that I run both manually and as cron jobs. The scripts generate stdout/stderr output reporting what they're doing. I want to see the output when I run the scripts manually. However, when the scripts are run by cron, normally I don't want cron to e-mail me that bulky output for each run--I want the output (the full output) only when the output is different than expected (e.g., if something has gone wrong or has changed, which I want to notice). Why not rewrite the scripts to take a command-line option such as either -q (quiet) or -v (verbose), and have it print out the required detail depending on the option? Then have that option supplied by cron when its run (or by you when you run it manually). Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: tar up a symbolic linked directory
On Sun, May 03, 2009 at 11:23:33PM -, Cameron Hutchison wrote:
Douglas A. Tutty dtu...@vianet.ca writes:
On Sat, May 02, 2009 at 09:04:38PM +, T o n g wrote:
I want to tar up a symbolic linked directory as if it is a real
directory. Is there any easy way to do it?
Let me explain with an example (that you can try):
mkdir d1
touch d1/{a,b,c}
ln -s c d1/d
ln -s d1 d2
I want that the result tar file looks like this:
tar -tvzf d2.tgz
drwxrwx--x tong/tong 0 2009-05-01 09:38 d2/
-rw-rw tong/tong 0 2009-05-01 09:37 d2/a
-rw-rw tong/tong 0 2009-05-01 09:37 d2/b
-rw-rw tong/tong 0 2009-05-01 09:37 d2/c
lrwxrwxrwx tong/tong 0 2009-05-01 09:38 d2/d - c
Any easy way to do it?
add -h to the tar parameters. It dereferences the symbolic lyinks.
However, then you won't get the d2/d - c reference.
I can't see that it would be possible to dereference the top-level
symlink but no others. No commands that I know of support selective
symlink dereferencing, except find(1) with -H.
That lead me to try:
$ find -H d2 | cpio -o -L -H ustar d2.tar
This comes close, storing d2/d as a link, but as a hardlink, not a
symlink.
Before we get too far down the garden path, would you review the why
of this?
It may be simpler to do it in a couple of steps. Tar up those you want
with the dereferencing, telling tar to ignore those you want without
dereferencing, then do a second tar run without dereferencing, telling
tar to ignore those you did with dereferencing, appending it to the
first tarball.
Doug.
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: BSD handbook - was Re: debiantutorials.org seeks input and new blood
On Sun, May 03, 2009 at 07:01:39AM -0500, Neal Hogan wrote: On Sat, May 2, 2009 at 1:19 PM, Bret Busby b...@busby.net wrote: On Sat, 2 May 2009, Neal Hogan wrote: On Sat, May 2, 2009 at 9:30 AM, Douglas A. Tutty dtu...@vianet.ca wrote: On Sat, May 02, 2009 at 06:27:44AM -0500, Neal Hogan wrote: So, if BSD is more complicated than using package management like RPM in Red Hat and .deb in Debian/Ubuntu, then it is probably too complicated for me. It's not . . . http://www.openbsd.org/ports.html The ports system works very easily, very similar to apt-get. However, right now, they don't have security updates for ports in -stable. If you run -current and want to update a port, AFAIK, you have to upgrade to the next snapshot for the whole system. For me, that's a lot of bandwidth on dialup. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Backing Up CMOS Settings Under Linux
On Sun, May 03, 2009 at 05:32:28PM -0400, Michael Pobega wrote: On Sun, May 03, 2009 at 03:19:50PM -0500, Martin McCormick wrote: I first thought it was my imagination, but I have had two Dell Dimension computers change their boot drive order. I don't know when it happens because they change to boot the hard drive just after trying the floppy such that the CDROM is last. This makes it hard to boot from any CDROM until the CMOS gets changed back. As a computer user who is blind, this is annoying because one must look at the screen to set things back as there is no network interface or serial port or much of anything else up when in BIOS setup mode. Perhaps you should try editing your /etc/fstab and mounting the devices by their UUIDs, rather than their /dev/ names. I'd go into more detail, but I have to head out; a quick Google search for mounting by UUID should suffice. AIUI, the problem is how to boot from CDROM instead of the hard drive. Can grub, booted by the bios, then boot a CDROM, or a floppy? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: tar up a symbolic linked directory
On Sat, May 02, 2009 at 09:04:38PM +, T o n g wrote:
Hi,
I want to tar up a symbolic linked directory as if it is a real
directory. Is there any easy way to do it?
Let me explain with an example (that you can try):
mkdir d1
touch d1/{a,b,c}
ln -s c d1/d
ln -s d1 d2
I want that the result tar file looks like this:
tar -tvzf d2.tgz
drwxrwx--x tong/tong 0 2009-05-01 09:38 d2/
-rw-rw tong/tong 0 2009-05-01 09:37 d2/a
-rw-rw tong/tong 0 2009-05-01 09:37 d2/b
-rw-rw tong/tong 0 2009-05-01 09:37 d2/c
lrwxrwxrwx tong/tong 0 2009-05-01 09:38 d2/d - c
Any easy way to do it?
add -h to the tar parameters. It dereferences the symbolic lyinks.
However, then you won't get the d2/d - c reference.
Doug.
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: debian and ubuntu - answer from user not pretending to be guru
On Sun, May 03, 2009 at 07:29:07AM -0500, John Hasler wrote: Bret Busby wrote: Before I try it, please advise whether, in removing the sudo facility for users, the package management (both adding/removing packages, and, downloading and installing updates, and using synaptic) will work by entering only the root password. The package management software just needs root privileges. It doesn't care how it got them. Nobody is suggesting anything exotic here. Sudo is intended to be configured by the system administrator. That's you. However, does the package management software (as aptitude does) store user preferences in the home directory? If, for example, you always run aptitude as yourself then give it the root password when prompted, it stores your preferences in your home directory. If you later run aptitude as root, those prefernces won't be active. Also, vis-versa. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: BSD handbook - was Re: debiantutorials.org seeks input and new blood
On Sat, May 02, 2009 at 06:27:44AM -0500, Neal Hogan wrote: FYI - While many of the fBSD folks will tout there ports/package system, I found it to be a pain (especially the upgrade), as did many others. There has recently been some chatter on their general mailing list to overhaul how they handle packages. Again, I found oBSD's package handling system to be superior. Last I looked (last week), OBSD doesn't have security updates (patches) for their packages; they only provide patches for the base release. If you want to run -current, then the packages get security patches. Since I'm on dialup, that would mean a lot of bandwidth time; basically, every time firefox or some third-party app required a security fix, I'd have to download the source for _everything_ and recompile _everything_. I wish I had time to work out a system that would run on base OpenBSD yet compile debs with OpenBSD's souped-up compiler. Then one would have the security of OpenBSD with good package security (Debian's security team with OpenBSD's compiler, with good responsivness). All the BSD's have a system to audit your installed packages for ones listed in a database as being insecure but the follow-on of patches to fix them is missing. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB PCI card to buy: [SOLVED] Startech PCI625USB2I
On Thu, Apr 09, 2009 at 10:01:40AM -0400, Douglas A. Tutty wrote: On Tue, Apr 07, 2009 at 10:52:41PM +0300, Dotan Cohen wrote: It's been a _long_ time since I've had a problem with a PCI USB card. If nobody pipes up with a negative, then I'd suggest that you give it a shot. The odds are in your favor. By the way, you may want to write to Belkin and ask them if the card will work with Linux. At the very least, you are expressing interest in having Belkin develop for Linux. They will never write works with Linux on the box if nobody is asking for it. Here is their address: http://www.belkin.com/contactus/index.asp Here's the reply I received from Startech: Both chipsets (Nvidia and NEC) are natively supported in the Linux = kernel since 2.4.x, but we do not directly support these cards in Linux, = nor have we tested with Debian, so compatibility cannot be guaranteed. Sincerely, Ray Aoki Technical Support Team StarTech.com Tel: 519 455-9675 Tel: 800 265-1844 Fax: 519 455-9425 The Startech PCI625USB2I 6-port (4 external, 2 internal) with the nVidia chipset works just fine, using the normal usbcore module, handled automatically by udev in Etch. Just thought I'd close out this thread with a solution for the archive. Thanks for your help. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: debiantutorials.org seeks input and new blood
On Sun, Apr 26, 2009 at 01:10:15PM -0400, JoeHill wrote: Dotan Cohen wrote: Maybe you forgot how great of an OS Win98 was at the time. This has to be a joke. Win 98 wasn't even an operating system. It was an application that ran on top of DOS for pete's sake. That was a different world than today, and even now seeing how responsive Win98 is on old hardware (it flies on 64 MiB RAM) it makes me wonder why Debian is sometimes sluggish on 512 MiB machines with 1 GHz procesors. What?? I seriously hope I'm misunderstanding you here. You're wondering why a modern, fully functional OS needs more resources than a flaky featurless GUI that was still running on top of 16 bit DOS code? They call it progress. 95% of what I do with my computer is the same as what I did on my 486. Progress means that I now need a computer a thousand times more powerful with five-hundred times more drive space to do exactly the same thing. I would be very happy with Debian Woody (or even Potato) with security updates only. Then again, modern monitors won't plug into my 486 since it doesn't have holes for all the pins on the D-sub. Luckily, I have a few spare monitors. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: debiantutorials.org seeks input and new blood
On Sun, Apr 26, 2009 at 11:27:24AM +0300, Dotan Cohen wrote: Document it all you want. But don't expect Joe Toothbrush to read it all. If one _wants_ go through pages upon pages of docs to create something new, that's great and the more the merrier. But if one _must_ go through the docs to use the product, then by virtue or Occam's razor the OS with the least documentation is the easiest to use. I remember when I got OS/2. It came with lots of books on the OS, to which I added a full set of RedBooks, and I had to learn REXX, with its own set of books and RedBooks. I spent a week floating out in a canoe (while the rest of the family had a reunion) reading. That was for my 386, back in the days when computers cost about what a car did, and ram went for $1,000 / MB. I forget how much AutoCad cost. I'm one to read the 1000 page book cover-to-cover. That way, I'll rememeber a significant amount and know exactly where to look when I need something I don't remember. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Installing on Compaq Armada 1500c - boot
On Sat, Apr 25, 2009 at 07:51:06PM +0100, Nuno Magalh??es wrote: I have this old laptop laying around. Currently is has no CD drive and i'd hate to rely on floppies. It does have a working PCMCIA eth card and a minimal OpenBSD that sees my LAN. I can't access its BIOS and i doubt it has one... these laptops used to have the bios in a hdd partition and this one got a new drive a while ago. So, i'm kinda stranded on that but i doubt it supported PXE (my initial idea) anyway. The floppies i have laying around for the bios utilities are kinda corrupted and won't help much - creating a new bios partition would require unalocated diskspace anyway. Any suggestions for giving this laptop a bios would be appreciated. So then i decided maybe the bootloader could support PXE or some other form of booting, but i'm still searching for BSD stuff on that matter... I'm assuming that you can't boot a USB stick. The OpenBSD install floppy can give you a shell running OpenBSD in memory. From within that, you can probably get your network going. You should then be able to repartition the hard drive (using fdisk) to clear out the OpenBSD install. Then install debootstrap and run it. It should pull in a basic Debian system and install it on the drive. See the installation manual for the use of debootstrap (installing from another flavour of UNIX). You may want to dd a copy of grub-disk (its a debian package) that gives you a grub menu (editable) for booting an OS. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: HP proliant ML115 G5 on debian lenny
On Sat, Apr 25, 2009 at 03:09:28PM -0500, Zhengquan Zhang wrote: On Sat, Apr 25, 2009 at 10:03:21AM -0400, Douglas A. Tutty wrote: I don't keep a backup server in the safety deposit box :), I keep the backup media. In this case, big USB stick (hard drives don't fit and This opens another option of using USB drive for very important data. But the capacity of USB sticks are small. Back to my question of what is the backup set size. Right now, I'm using 16 GB USB sticks, but larger are available (32, 64 MB). Before I got the sticks, I did a cost comparision between new USB sticks and used LTO-1 tape drive ($250) and tapes. The break-even point assuming 2 sets (one at home, the other in the bank) was 72 GB. However, USB sticks are a lot more robust than tapes. Tapes should be transported to the bank in a padded carrier, and don't drop them when you transfer to the safety deposit box. Apparently, a number of data centers in the World Trade Centre had their off-site backup in the other tower. Keep that in mind when you choose the different building. That is true. They two building should be of different heights. :) Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: debiantutorials.org seeks input and new blood
On Mon, Apr 27, 2009 at 09:57:40AM -0400, H.S. wrote: Douglas A. Tutty wrote: SNIP I'm one to read the 1000 page book cover-to-cover. That way, I'll rememeber a significant amount and know exactly where to look when I need something I don't remember. Now a days google is a *huge* help in this. Not if the book isn't on-line for google to index. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: debiantutorials.org seeks input and new blood
On Mon, Apr 27, 2009 at 10:10:13AM -0400, Miles Fidelman wrote: H.S. wrote: Douglas A. Tutty wrote: On Sun, Apr 26, 2009 at 11:27:24AM +0300, Dotan Cohen wrote: Now a days google is a *huge* help in this. There's still something awfully useful and compelling about a serious reference manual, all in one place, with a comprehensive table-of-contents, detailed index, and embedded references. Especially when the problem is that the computer won't boot, or can't get on the internet to run google... Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: HP proliant ML115 G5 on debian lenny
On Fri, Apr 24, 2009 at 02:54:38PM -0500, Zhengquan Zhang wrote: On Fri, Apr 24, 2009 at 03:19:01PM -0400, Douglas A. Tutty wrote: Well, debian has different requirements re licensing of modules. Your guess may be wrong if HP has provided a propriatary module for the kernel that e.g. suse has included in its kernel but debian can't include. For some things (e.g. the nVidia driver), you can still get an install done and add a module later; for the boot drive that becomes a bit of a problem :) I understand what a boot drive is now. But I still do not understand why 'for the boot drive that becomes a bit of a problem'. particularly I don't understand what 'that' refers to. Does that mean I will fail to install lenny on it or adding module on boot drive is more of a problem? If the kernel needs a module to find the drive which will be the boot drive, then the installer kernel would need the module before the installer could find the disk on which to install. I don't know if the installer program has a means of adding a module then rescanning for hard drives. 3. Copy the data from the backup server to some remote location either with removable media or a second backup server at a remote location. A lot of this depends on the size of the backup set and your options of remote location. I keep a backup in the bank's safety deposit box. safety deposit box? are you serious:) so how to vent the heat? I don't keep a backup server in the safety deposit box :), I keep the backup media. In this case, big USB stick (hard drives don't fit and sticks are cheaper and more rugged than LTO-1 tapes [and no separate drive required]). I will put the backup machine in a different building. Apparently, a number of data centers in the World Trade Centre had their off-site backup in the other tower. Keep that in mind when you choose the different building. Another consideration: Imagine that all you have is the backup media. What info will you need to use that media to recreate the entire system? Will you need any printed documentation to access the data on the media? Will you need any special hardware (e.g. a tape drive of a particular model) to access the media? Is the data encrypted? What happens if the media is lost or stolen? Here, media could be your disk array; they still fit in a backpack. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: HP proliant ML115 G5 on debian lenny
On Thu, Apr 23, 2009 at 10:18:23AM -0500, Zhengquan Zhang wrote: Dear debian community, We plan to buy an HP proliant ML115 G5 for server backup. The CPUs would be amd opteron 64bit. And there is embedded sata raid controller. I will use raid1 on two 1T harddrives. I would like to consult debian community to see if anyone have some experience with installing lenny on it. I searched debian wiki and other sites, but did not find specific info on this particular type of HP server. And on HP website, the debian compatible list does not include this type. I am especially concerned with the raid controller part. Will that work with lenny? The only __definitive__ way to know would be to take the netinst CD to the box, boot it up and check dmesg (and the installer screens) and see if it sees the drives. Note that embedded sata raid controllers are generally fake raid. You're better off just using them as normal disk interfaces (don't try to configure as raid) and using software raid from the installer CD. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: HP proliant ML115 G5 on debian lenny
On Fri, Apr 24, 2009 at 09:41:12AM -0500, Zhengquan Zhang wrote: On Fri, Apr 24, 2009 at 10:34:21AM -0400, Douglas A. Tutty wrote: The only __definitive__ way to know would be to take the netinst CD to the box, boot it up and check dmesg (and the installer screens) and see if it sees the drives. Note that embedded sata raid controllers are generally fake raid. You're better off just using them as normal disk interfaces (don't try to configure as raid) and using software raid from the installer CD. This is exactly what I want to know. I 'guess' if it officially supports redhat and suse, debian will have no problem seeing the hardware. But I am not sure about the so called embedded raid, and could you explain a little bit why it is 'fake'? Well, debian has different requirements re licensing of modules. Your guess may be wrong if HP has provided a propriatary module for the kernel that e.g. suse has included in its kernel but debian can't include. For some things (e.g. the nVidia driver), you can still get an install done and add a module later; for the boot drive that becomes a bit of a problem :) So-called 'fake' raid is, as I understand it, hardware that allows you to configure the raid in the bios, but the actual raid happens in windows software rather than in the hardware. Also, is there any other penalty or downside for using software raid? As I know, for RAID 1, the performance is not affected much. There is very little performance difference for software raid. Think about two scenarios: 1. hardware raid application tells the OS to tell the logical drive (device presented to the OS by the hardware raid card) to store some data. OS waits while the actual disks store the data after the hardware has sent the data to both disks. 2. software raid application tells the OS to tell the md (device presented to the rest of the OS kernel by the software raid portion of the OS kernel) to store some data. The OS waits while the md driver sends the data to both disks. For there to be any observable performance hit, the wait while the data is presented to each disk would have to be considerable; the wait while the data goes to the disks is the same. And the server will merely be used for backup. A couple of issues then. 1. Performance may or may not be an issue, depending on how many other computers will be using the server for data backup at the same time. 2. With hardware raid, unless the raid card can save the configuration to each disk in the array, if something happens to the card (which could happen if a drive fails and takes down the controller), then the whole array could be caput if you put in a new controller card. 3. With software raid, the configuration is on the disk itself. Pop those disks in a new box and they should work (assuming that the new box's hardware can be booted by the old box's initrd). 4. Hardware raid comes into its own with exotic raid types (e.g. raid50 or raid60), with hot spares, hot swap, auto rebuild, etc. 5. There has been some talk recently here on the increased liklyhood of raid failue after a single drive failure. Apparently, the time it takes for a replacment second drive to rebuild makes the liklihood of the other drive failing before the rebuild is complete of some concern with very large drive sizes. In this case, having three active raid1 drives with a hot spare (4 drives total) is one way to mitigate this risk. You may need to do lots of research depending on: 1. The size of your backup set 2. The importance of the data 3. The number of locations of the backup data. Good luck. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Boot Debian to a command line and prevent X from starting on boot
On Fri, Apr 24, 2009 at 02:05:00PM -0400, debian debian wrote: On Fri, Apr 24, 2009 at 11:25 AM, Robert Menes viewtiful.icc...@gmail.com wrote: Hi folks, I have a Debian lenny-stable install that I need a little memory refresher with. I need to boot and go straight to a terminal, and not start X on bootup. I know there was a way of doing so, but I forgot (I'm a little rusty and need to not be rusty). Can someone tell me how to just boot and land on the command line, bypassing X? Thanks! You could always boot into single user mode (add single or s to the kernel boot options) Never having used an X login manager (only using startx), I don't know if single works. Therefore, if you find that you still get X, then add init=/bin/sh to the kernel command line. This will mean that init does not start, and therefore no initscripts run. You will have to mount your / drive (or manually run each initscript in order with the start parameter) before you can edit any files. When done, you can exec init to get going again. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: HP proliant ML115 G5 on debian lenny
On Fri, Apr 24, 2009 at 01:41:45PM -0500, Zhengquan Zhang wrote: On Fri, Apr 24, 2009 at 02:10:41PM -0400, Douglas A. Tutty wrote: Thank you very much for your reply! Well, debian has different requirements re licensing of modules. Your guess may be wrong if HP has provided a propriatary module for the kernel that e.g. suse has included in its kernel but debian can't include. For some things (e.g. the nVidia driver), you can still get an install done and add a module later; for the boot drive that becomes a bit of a problem :) My plan is to use the 250G harddrive as system drive and use 2x1T dardrives to do RAID 1 backup for several other servers. So what do you mean by 'for the boot drive'? The boot drive is whatever drive the system boots with. In the model you're describing, it would be the 250 GB hard drive. Note, that for a simple backup server, any drive over 1 GB (say 6 GB) will hold all the system you need. For extra redundancy, you may want to experiment and try installing a system onto a 6 GB partition on one of the drives. I bet you'll find it more than big enough. You could then reinstall, but put a 6 GB (or 10 or 12, whatever) partition at the beginning of each drive, in a raid1 fashion. In this way, if any drive fails, you'll still be able to boot the system. So-called 'fake' raid is, as I understand it, hardware that allows you to configure the raid in the bios, but the actual raid happens in windows software rather than in the hardware. Could you please explain 'in windows software' a little bit? Does that means the processing for raid is done in CPU rather that in the RAID controller? Yes, with fake raid, the hardware can't do any raid processing. It is done by the main CPU just as in normal software raid. And the server will merely be used for backup. A couple of issues then. 1. Performance may or may not be an issue, depending on how many other computers will be using the server for data backup at the same time. three linux servers(1 development server 1 mail and web server and 1 misc server) will be backing up their data to the back up server I purchase. Will they be backing-up at once or one-at-a-time? If one-at-a-time, then unless those servers are using raid striping, the throughput of the servers' hard disk will be similar to the throughput of the backup server's hard disk. However, if all three boxes will be spitting data to the backup server as fast as their hard drive (and network) can move the data, then the backup server will need to be of higher performance if it to avoid being a bottleneck. 2. With hardware raid, unless the raid card can save the configuration to each disk in the array, if something happens to the card (which could happen if a drive fails and takes down the controller), then the whole array could be caput if you put in a new controller card. So this must be the downside of using hardware RAID? Unless the card can safe the config to the actual disks. Another is that the raid cards have their own cache. If you don't have a UPS, then you'll want to set up the cache so that they don't tell the OS that the data is on disk until it really is on disk (not just in its cache), unless you get a raid card with a battery backup for its cache. 3. With software raid, the configuration is on the disk itself. Pop those disks in a new box and they should work (assuming that the new box's hardware can be booted by the old box's initrd). This is great! 4. Hardware raid comes into its own with exotic raid types (e.g. raid50 or raid60), with hot spares, hot swap, auto rebuild, etc. I will only use raid 1, that is because this is simple and effective as it appears. However, re performance (above), if you have three boxes streaming data to the backup server, you may want raid10 (which you can do with software raid) or raid 50, or raid60 (which can handle multiple drive failures). Its always a trade-off. 2. The importance of the data mailing list archives, web data, database, svn repos.. home directories.. They are very important and I can not afford to lose any of them. 3. The number of locations of the backup data. What do you mean by this? Is it ok to put all of the backup in one backup server? That means a single point of failure. What happens if the power supply fries the whole backup server? Think of it as several steps: 1. Backup the data, somewhere on the same box (optional) 2. Copy the backup data to the backup server. 3. Copy the data from the backup server to some remote location either with removable media or a second backup server at a remote location. A lot of this depends on the size of the backup set and your options of remote location. I keep a backup in the bank's safety deposit box. -- To UNSUBSCRIBE, email to debian
Re: debian with raid1+cryptsetup+lvm on notebook?
On Tue, Apr 21, 2009 at 12:39:38PM +0200, Peter Jordan wrote: Hello, since my ThinkPad T400 has two 250GB HD, i considered to install debian testing with raid1+cryptsetup+lvm on it. Has anyone experience with that kind of setup? Any significant reasons against my plan? Sounds like a good idea. I think that the installer has that out-of-the-box as one of the guided-partitioning options. If not, you can certainly do it manually. This came up not that long ago. It was suggested that having / encrypted can prevent someone trojaning executables on / (e.g. /bin/ls). However, since you need an unencrypted /boot, then someone could trojan the kernel or the initrd itself (perhaps to email the attacker the password you enter to decrypt the filesystem), who knows? I suppose that you could have /boot on a USB stick so that without the stick, the laptop won't boot and there won't be any unencrypted data on the laptop. There's good LUKS documentation: read it. I'm sure that this has been (and is being) looked at by people with a particular interest in laptop security. Just don't assume that raid1+crypsetup+lvm will make your laptop absoulutly secure. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: File won't be printed, all others will!
On Tue, Apr 21, 2009 at 03:15:30PM +0100, Rodolfo Medina wrote: In my sister's home directory there is a pdf file that won't respond to the `lp' command. All others pdf files in the same directory behave all right, and the permissions are the same. The only difference is the creation date, which is today wheras the other files are older. The thing looks mysterious to me. Can anybody suggest any explanation/remedy? 1. Are you sure that it is a .pdf file? run file on it. 2. Did she put that file there? 3. If you are unsure of the file, use a sacrificial user (i.e. create a new user named e.g. goat), have her email that file to goat, then log in as goat, run mutt, save the file to its home (or ~/uldl) directory, then start x, and try to open the file with a pdf viewer (xpdf, kpdf, whatever). 4. If the file opens with a pdf viewer, what happens if you try to print from the viewer? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Software raid OK?
On Mon, Apr 20, 2009 at 03:29:00PM +0300, Michael Iatrou wrote: When the date was Monday 20 April 2009, BAGI Akos wrote: Hi List! I installed a software raid, level1 with 3 disks, one of them is a spare. I have 2 partitions: md0 is for / and is made of sda1,sdb1, sdc1 md1 is for swap and made of sda2,sdb2, sdc2 There is no particularly good reason to have the swap on RAID. You should define three independed swap partitions; if disk fails, kernel will use the other available. If swap fails, what happens if something important to the running of the system (not just a user app) is swapped-out? I've seen advice on this list many times that to avoid a crash, if other system stuff is on raid, that swap should be as well. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: clean up the laptop body/keyboard
On Sat, Apr 18, 2009 at 01:33:38AM -0300, Hashimoto wrote: What kind of product do you suggest to use to clean up my laptop body and keyboard ? To be explicit: other than the screen. I've always started with air, then a damp cloth. The owner's manuals for plastic electronic things generally say not to use anything else. My next step has been to figure out what the dirt is and from that determine what cleaner to use. The problem is that key markings are often chemically similar to the dirt. Another problem is that cleaners that don't require rinsing often rely on alcohol to dry; alcohol may be bad for the plastic. However, not rinsing a cleaner that requires it leaves the surface sticky and prone to getting dirty. What kind of dirt is on the laptop body and keyboard? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how can i turn /dev/null into an MTA?
On Tue, Apr 14, 2009 at 03:08:41PM +0200, Dirk wrote: Douglas A. Tutty wrote: On Tue, Apr 14, 2009 at 10:26:17AM +0200, Dirk wrote: nah.. instead of configuring a package i don't want to install in the first place i just run a cronjob that de-installs the MTA every 30 minutes using dpkg --force-all --purge exim4 ..so i can run updates and the cronjob makes sure it turns out the way i want it.. running dpkg every 30 minutes uses far more resources than running exim. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: how can i turn /dev/null into an MTA?
On Tue, Apr 14, 2009 at 10:26:17AM +0200, Dirk wrote: (see subject) i don't want an MTA running on a system... but many programs require it as dependency to spam me with their stuff (which should belong into just a log file (IMO))... Unix without an MTA??? Why not install exim, then look at the filters section of the docs. Exim can do the same kinds of filters as procmail for each user. Don't have an alias for root's mail, put a filter in /root/.forward (or wherever its supposed to go; its been a while), and then have that filter put the mail wherever you want. If you really don't want the mail, the filter can delete it without it being read. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: SD-NECU2-5E1I - Re: Linux (Debian) support of USB card
I thought I'd send this along for the archive, FYI. So far I haven't found _any_ that say that they _are_ compatible with Linux, either they know that they aren't (such as Syba) or they believe that support is in the chipset but that they haven't tested them. I guess I'll go with startech since at least its possible. Doug. On Mon, Apr 13, 2009 at 10:19:17AM -0700, michael wrote: Dear Customer, Thanks for your inquiry. For those 3 PCI USB 2.0 cards, they are not compatible with Linux. Unfortunately, we do not have a PCI USB 2.0 card which it is compatible with Linux. Please let me know whether the answer is helpful for you. Best Regards, Michael, mich...@us.syba.com Syba Support Team - Original Message - From: Douglas A. Tutty dtu...@vianet.ca To: support...@syba.com Sent: Saturday, April 11, 2009 7:04 AM Subject: Linux (Debian) support of USB card Hello, In my small town, the only USB-PCI cards available off-the-shelf are Belkin and Belkin says that they won't work with Linux. I have to special order a card with no possibility of returns if it doesn't work, so I need to know from the manufacturer that it will work. I'm interested in your PCI USB 2.0 6x port (5+1) SD-NECU2-SE11 with NEC chipset 720101, or the SD-81012053 (I couldn't tell the difference on the web site); or the PCI USB 2.0 5x port SD-V2-5U with NEC chipset uPD720101. Can you tell me if either will work with Debian Linux (Etch) running 2.6.18 kernel (or Lenny with whatever 2.6 kernel it uses)? Thank you. Doug Tutty dtu...@vianet.ca Orillia Ontario Canada. No virus found in this incoming message. Checked by AVG - www.avg.com Version: 8.0.238 / Virus Database: 270.11.54/2056 - Release Date: 04/13/09 05:51:00 -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: related to mail servers, mta and mda
On Sun, Apr 12, 2009 at 08:53:11PM -0400, H.S. wrote: This is a beginner's (no experience with setting up mail servers) query about MTA's and MUA's. I am trying to see if I can setup an mta or a related application on my Debian machine which is being run as a router for my home lan such that it can send email to an external email address. It does not need to receive any public email at all. I have already tried heirloom and I can send email to my gmail account if I put my gmail log in info in my mail's conf file. Now I am wondering if I can allow outgoing email (need to have port 25 open?) with no need nor requirement to receive any in coming email from the WAN without having to use a particular email's log in info. The idea is that email from that machine (and perhaps from lan machines) may be sent to any valid email address with reply-to address changed to a fixed email address. I'd suggest that you use exim and use your isp's mail server as a smarthost. When you install exim, the debconf questions will give you this choice. You'll need the hostname of your isp's smarthost. Local mail will be delivered locally, non-local mail will be sent to the smarthost (with address rewriting so that it appears to the receiver as coming from your public mail name (e.g. h...@example.com, rather than h...@myhome). If you want to receive public mail, you'd use something like fetchmail. Unless you're running a firewall, you'll already have port 25 (all ports) open. Installing an MTA will simply provide a server listening to port 25. However, with the standard debconf smarthost, I don't think it actually will be listening on your public IP, only on your localnet. I'd further suggest that you install the doc-linux-howto (something like that) package, probably in html format. You'll find lots of info, including mail admin howto. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Lenny overheating, preventing installation
On Sat, Apr 11, 2009 at 02:26:30PM +0200, Aleksa ??u??uli?? wrote: The laptop is less than a year old and still in warranty. It has never been used in dusty or dirty places. And this overheating only happens with Debian (installing OpenSuSE or Mandriva or Ubuntu or Fedora works a breeze). The only other instance it does happen is when my current OpenSuSE system freezes (stops responding), ramping up the CPU to 100% usage: if I don't switch to a virtual terminal and reboot within, say, 10 minutes, the laptop will shut itself off from overheating. Hence my assumption that the machine simply is not DESIGNED to work at full throttle (100% CPU usage) for any length of time. But I may be wrong, of course. As a sidenote: I've found a thread on internet a while ago stating that you may risk overheating and even frying a laptop if you try installing Windows98 as a virtual machine, since Windows98 does not support the CPU idle instruction. I assume something vaguely similar may be going on here. Modern laptops with fairly powerful CPUs apparently rely on certain subsystems of the OS to effectively prevent overheating. If some of those subsystems don't work as expected, overheating will occur. I find it hard to believe there aren't more laptop users with this sort of problems... So they install a powerful CPU for the marketers to get you to drool over, then don't provide the necessary cooling so that when you actualluy _use_ the CPU power, the unit fails. I'd call it a design flaw and return the unit. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Lenny overheating, preventing installation
On Sat, Apr 11, 2009 at 03:46:31PM +0200, Klistvud wrote: Dne sobota 11 april 2009 ob 15:22:38 je Douglas A. Tutty napisal(a): On Sat, Apr 11, 2009 at 02:26:30PM +0200, Aleksa ??u??uli?? wrote: I agree ... to a point. Namely, I've never managed to overheat the unit by just _using_ the CPU. It exclusively happens: 1) during a Debian Lenny installation 2) during a certain type of system lock-up which forces both the CPU frequency and its usage to go to 100%. I'd compare (if I may) the situation with a car having, say, a range of 0 to 7000 RPM, of which only 2000 to 5000 is actually the working range. Now, forcing the car in a very low gear and running it at a constant 7000 RPM, how many minutes until the engine overheats? And, more importantly: how _stupid_ should one be to actually try doing this at home? It's, in my opinion, what's happening here: some runaway process or OS flaw simply ramps up the CPU to a regime that wasn't intended to be used for a prolonged time in the first place. In normal usage, leaving a CPU running at 100% usage is a rare occurence (I'm not talking about CPU frequency here, I'm talking about CPU usage - 100% meaning no idle cycles whatsoever over several minutes or even hours!). What about a big compile? Retouching a movie? For normal usage (other than using iceweasel), I have difficulty getting my dual PII-450 to less than 90% idle, but there are things that will bog it down, and no it doesn't overheat. Then again, its a 2U server, not a laptop. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB PCI card to buy
On Fri, Apr 10, 2009 at 05:37:28PM +0300, Dotan Cohen wrote: Here's the reply I received from Startech. Hi Doug, Both chipsets (Nvidia and NEC) are natively supported in the Linux = kernel since 2.4.x, but we do not directly support these cards in Linux, = nor have we tested with Debian, so compatibility cannot be guaranteed. I would recommend letting them know that you will not buy until they support Linux. But I need it so I will buy it. From their perspective, they know that the chipsets are supported in the linux kernel, but with all the different distributions and versions (compared to just testing for windows), it would be hard for them to test them all. Whether or not you actually do buy the hardware is another thing, but ask the _vendor_ about Linux compatibility issue returns beforehand. Well, its special order, no returns for non-defective products. They define defective as they can't plug it into their computer and get it to work. Defective products are sent away to the manufacture for their warranty. The vendor has not warranty of their own. That's life. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Advice on raid/lvm
On Thu, Apr 09, 2009 at 10:00:40AM +0300, Tapani Tarvainen wrote: On Wed, Apr 08, 2009 at 06:02:26PM -0300, Henrique de Moraes Holschuh (h...@debian.org) wrote: On Wed, 08 Apr 2009, Miles Fidelman wrote: I suggest that a small (1GB-4GB) partition for simple md-raid1 be used for / instead. That won't give you any headaches, including on disaster recovery scenarios. If you're going to have separate partitions (e.g. /usr, /var, /home), then I would not call a 1-4 GB / small. I have a 500 MB / of which only 117 MB is used. I would respectfully disagree. There are significant advantages in putting / in LVM, it is a well-supported, standard configuration, and avoiding it only gives false sense of security: in a disaster situation you need to know basics of mdadm and lvm anyway, if you use them. Yes, leaving / out of LVM does give you a more complete environment to work with when system crashes in a way that LVM (the volume group containing /) is inaccessible. It doesn't help much though unless you also leave /usr out, and I've lost count on how often I've enlarged /usr and been grateful it was under LVM. What is in /usr that you'd need (ok, other than man pages)? All the essential tools for managing software raid and lvm are, however, available even without / - indeed they're in initrd, and if you can't use them, you're out of luck anyway. You don't have whatever notes you've left yourself in /root On the other hand, having / in LVM means: * you can enlarge / when necessary; You should never have to enlarge a 500 MB / * you can encrypt / if desired; Why would you need / encrypted (if swap, /tmp, /home, and parts of /var are encrypted)? * you can use other RAID configurations besides RAID1 with /; True, but for 500 MB is that helpful? If you have more than 2 disks, just put a 500 MB partition on each and have more than 2 components to the raid1 array. * you don't have to create separate volumes for each of /usr, /var and /tmp (although you probably should anyway); * it's the standard configuration, offered as automatic default installation option, and many people are using it so finding someone to help when needed shouldn't be hard. I've never used the automatic default; It always wastes resources on my boxes. As for the rest of your points, well, both software raid and lvm do increase complexity and require learning some new tricks, but they're well worth the trouble if you manage any system more complex than a simple workstation, IMHO. Figure out what all documentation, man pages (in text format), notes, etc that you would want and put them in /root/doc. Any scripts that you find helpful for rebuilding arrays you could put in /root/bin. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Advice on raid/lvm
On Wed, Apr 08, 2009 at 06:34:09PM -0500, Mark Allums wrote: Not really answering your question directly, but may I suggest, if cost is not *absolutely* critical, that you consider RAID 10? If it is a server, then certainly you will want to get away from a three-drive RAID 5. A RAID 10 is a good compromise between redundancy, speed, and cost. It just takes four drives instead of three (or two.) Is there an advantage of software raid10 over multiple raid1 arrays joined with LVM? Capacity can be dynamically added with pairs of disks. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Trouble burning netinstall CD : SOLVED -- it's trouble with k3b's check.
On Wed, Apr 08, 2009 at 10:49:56PM +, Hendrik Boom wrote: Anyone know what I should do with three Lenny install disks? :-) Put one somewhere safe, with your off-site backup. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: dialing phone numbers
On Thu, Apr 09, 2009 at 10:05:37AM +0200, Pol wrote: Any hints about dialing numbers by your laptop keyboard to a phone connected through modem? minicom? echo? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB PCI card to buy
On Tue, Apr 07, 2009 at 10:52:41PM +0300, Dotan Cohen wrote: I need to add USB to my HP NetServer LPr. Do they all just work now? It's been a _long_ time since I've had a problem with a PCI USB card. If nobody pipes up with a negative, then I'd suggest that you give it a shot. The odds are in your favor. By the way, you may want to write to Belkin and ask them if the card will work with Linux. At the very least, you are expressing interest in having Belkin develop for Linux. They will never write works with Linux on the box if nobody is asking for it. Here is their address: http://www.belkin.com/contactus/index.asp Here's the reply I received: Dear=20Sir=20Or=20Madam,brbrYour=20recent=20request=20for=20support=20= has=20been=20answered.=20Please=20click=20a=20href=3D'http://www.belkin.c= om/contact/milan/?id=3D0C580C0C0E0C5C5E570A0C0D5B090E5C305F5E'here/a=20= to=20view=20your=20response=20from=20our=20support=20team.brbrAOL=20us= ers=20please=20copy=20and=20paste=20the=20link=20below=20into=20your=20bro= wser.brbrhttp://www.belkin.com/contact/milan/?id=3D0C580C0C0E0C5C5E570= A0C0D5B090E5C305F5EbrbrMany=20thanksbrbrBelkin=20Support When I clicked on it, I got the message that they are not supported in Linux. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB PCI card to buy
On Tue, Apr 07, 2009 at 10:52:41PM +0300, Dotan Cohen wrote: I need to add USB to my HP NetServer LPr. Do they all just work now? It's been a _long_ time since I've had a problem with a PCI USB card. If nobody pipes up with a negative, then I'd suggest that you give it a shot. The odds are in your favor. By the way, you may want to write to Belkin and ask them if the card will work with Linux. At the very least, you are expressing interest in having Belkin develop for Linux. They will never write works with Linux on the box if nobody is asking for it. Here is their address: http://www.belkin.com/contactus/index.asp Here's the reply I received from Startech. Hi Doug, Both chipsets (Nvidia and NEC) are natively supported in the Linux = kernel since 2.4.x, but we do not directly support these cards in Linux, = nor have we tested with Debian, so compatibility cannot be guaranteed. Sincerely, =20 Ray Aoki Technical Support Team StarTech.com Tel: 519 455-9675 Tel: 800 265-1844 Fax: 519 455-9425 Startech has a 4+2 port with nVidia chipset and a 3+1 port with NEC. Any preference? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Advice on raid/lvm
On Thu, Apr 09, 2009 at 04:43:17PM +0200, martin f krafft wrote: also sprach Douglas A. Tutty dtu...@vianet.ca [2009.04.09.1532 +0200]: On the other hand, having / in LVM means: * you can enlarge / when necessary; You should never have to enlarge a 500 MB / I bet you'll be wrong in 10 years. What load of gunk will be dumped into / to take it bigger than 500 MB? If ever / becomes bigger than 500M, then booting my old boxes will again require a separate /boot (so that they can boot lower than the 504 MB limit). * you can encrypt / if desired; Why would you need / encrypted (if swap, /tmp, /home, and parts of /var are encrypted)? Because it contains e.g. /bin/ls and you don't want that to be trojaned. Obviously, an integrity checker can also help. How does encrypting / prevent trojaning a binary? I suppose it prevents an attacker gaining root when the box is turned off and not physically secured, but I don't know. Does encrypting root counteract the age-old wisdom that physical acess to the hardware will allow root compromise? An integrity checker would only help if its being run from a known-secure box, not the box with the questionable /bin/ls. Encryption is great to protect secret content, while the box is powered-off. It doesn't help while the box is powered-on (since the filesystems will be decrypted). Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Decrypt, Edit and Encrypt a File
On Tue, Apr 07, 2009 at 06:17:56PM -0400, H.S. wrote: Douglas A. Tutty wrote: On Tue, Apr 07, 2009 at 05:46:31PM -0400, H.S. wrote: Douglas A. Tutty wrote: Where does it hold the decrypted data? Does it stay in RAM, does it get swapped, does it go to a scratch file? This might help: http://www.easypg.org/ yea, it looks like it can leak info. I am just starting to play around with the two tools (gnupg.vim and easypg). You appear to already know about vim and have pointed out a possible problem with easypg. Does this problem not exist in the vim plugin? I don't use the vim plugin. I currently have encrypted swap, with /tmp on tmpfs (so /tmp is encrypted), and have /home encrypted too. On boxes where /home isn't encrypted, I still have swap and /tmp encrypted. To decrypt a file without leaking info (other than that ram itself isn't encrypted, but that's another kettle of fish), I pipe the file through openssl and point the output to somewhere in $TMP. Example: In /etc/pam.d/common-session, I have: session optional pam_tmpdir.so which gives a $TMP of /tmp/user/1000 $ cat /proc/swaps: FilenameTypeSizeUsedPriority /dev/mapper/sda2_crypt partition 979956 0 -1 $ mount | grep -i /tmp /dev/mapper/sdb1_crypt on /var/tmp type ext3 (rw,data=journal) tmpfs on /tmp type tmpfs (rw,size=500m) Encryption/decryption is done with OpenSSL's enc To encrypt: openssl aes-256-cbc -a -e -salt -in file -out file.aes To decrypt: openssl aes-256-cbc -a -d -salt -in file.aes -out file All I need to do to keep unencrypted data off an unencrypted /home (or USB stick, whatever), is to ensure that the output of decrypt only goes to /tmp (or additionally in my case, /var/tmp). I can then edit the file, do whatever I want (as long as the software doesn't itself leak info), then re-encrypt the file putting the encrypted file whereever I like. There are a limited number of place where a piece of software can leak info: 1. to /tmp:Its encrypted 2. to /var/tmp:Its encrypted 3. to somewhere on ~/ I have /home encrypted 4. left in swapits encrypted 5. in a core dump after a crash: use ulimit (see man bash) to limit core dump size to 0 Of course, the uncrypted data is in memory, so anything that can read any memory segment can read your data. I hope that this helps in your considerations. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: 2.6 kernel issues
Steven, Please don't top post. I've tried to reorganize this in the correct order, but the quoting wasn't consistant. -Original Message- From: Boyd Stephen Smith Jr. [mailto:b...@iguanasuicide.net] Sent: Wednesday, 8 April 2009 2:17 p.m. To: debian-user@lists.debian.org Subject: Re: 2.6 kernel issues Steven Jones wrote: Is there any alternative apt-get sources that allow me not to upgrade my kernel in stable? The new 2.6 kernel I am forced to use wont boot on my scsi raid card, it panics Does any 2.6 kernel work for you? If so, simply continue using the Etch kernel package until you can get one that works for you into a Lenny update (by filing a bug and helping the maintainers and developers). The 2.6 Etch kernel will run any software allowed into Lenny (to allow Etch - Lenny upgrades during Etch's continued security support.) If no 2.6 kernel works for you, you will not be able to use Lenny. On Wed, Apr 08, 2009 at 02:32:47PM +1200, Steven Jones wrote: No, no 2.6 kernels work and no 2.4 kernel past 2.4.27 works as a bug was introduced into the megaraid driver about thenits never been solved. The reply I have is no developer has the hardware and they say its too old to bother with. I have an HP NetRaid 1si hardware raid card, running Etch, and it boots just fine. I haven't upgraded to Lenny yet (I'm waiting for a full backup, but need more media). The Lenny installer in rescue mode doesn't see the drives although in install mode it sees them. If at some point you find that Linux stops working for you, for one OpenBSD still supports the megaraid. You may find that the other BSDs do as well. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: creating a compact binary
On Tue, Apr 07, 2009 at 11:13:53PM -0400, Mag Gam wrote: On Tue, Apr 7, 2009 at 10:39 AM, Ron Johnson ron.l.john...@cox.net wrote: On 2009-04-06 18:59, Mag Gam wrote: I was wondering if its possible to compile src code (for example rsync) to create 1 large binary. I want to do this to easily distribute rsync. To what kind of platform? You can't just require that certain libraries exist? Besides, static binaries are *huge*. Definitely not compact. Do you really mean stand-alone? I don't really care about the size. But I really want the entire rsync distribution to be in 1 file. What is the difference between static binary and standalone? Isn't there a tool that takes a dynamically linked binary and creates a statically linked binary without recompiling from source? I still can't imagine needing to create a statically linked binary to distribute it. Is this for unix? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Decrypt, Edit and Encrypt a File
On Wed, Apr 08, 2009 at 11:26:20AM -0400, H.S. wrote: Douglas A. Tutty wrote: On Tue, Apr 07, 2009 at 06:17:56PM -0400, H.S. wrote: Douglas A. Tutty wrote: On Tue, Apr 07, 2009 at 05:46:31PM -0400, H.S. wrote: Douglas A. Tutty wrote: $ cat /proc/swaps: FilenameTypeSizeUsedPriority /dev/mapper/sda2_crypt partition 979956 0 -1 $ mount | grep -i /tmp /dev/mapper/sdb1_crypt on /var/tmp type ext3 (rw,data=journal) tmpfs on /tmp type tmpfs (rw,size=500m) There are a limited number of place where a piece of software can leak info: 1. to /tmp:Its encrypted 2. to /var/tmp:Its encrypted 3. to somewhere on ~/ I have /home encrypted 4. left in swapits encrypted 5. in a core dump after a crash: use ulimit (see man bash) to limit core dump size to 0 Of course, the uncrypted data is in memory, so anything that can read any memory segment can read your data. I am curious, what is the performance cost to have your ~/ and /tmp and /swap encrypted? What kind of machine are you using? I see no performance cost and I've done this on a PII-233 on old IDE disks and now on a PII-450 with two SCSI disks. Think of it this way: Data has to go out to disk. That takes a lot of time compared to moving data in the memory. It doesn't take that much time to start the encryption process (which can continue while the data is streaming to the disk). I'm sure its more complex than this, but that's the idea. If I want to have the fastest streaming possible, I'll have a non-encrypted partition in a convenient spot, say /var/local/cache but I've found that I haven't had a problem without it. I suppose it may be important if you're doing streaming video or something, but you'd want a separate raid array for that anyway. Also, since you have to use mount, which requires root privileges, the above method is not possible for a normal (non-root) user, is it? These partitions are all mounted on boot. I set up the encrypted partitions during the install (Etch). Here's fstab: # /etc/fstab: static file system information. # # file system mount point type options dump pass proc/proc proc defaults0 0 #/dev/sda1 / ext3 defaults,data=journal,errors=remount-ro 0 1 LABEL=root / ext3 defaults,data=journal,errors=remount-ro 0 1 /dev/mapper/sdb2_crypt /home ext3 defaults,data=journal 0 2 #/dev/sda3 /usrext3 defaults,data=journal 0 2 LABEL=usr /usrext3 defaults,data=journal 0 2 #/dev/sda5 /varext3 defaults,data=journal 0 2 LABEL=var /varext3 defaults,data=journal 0 2 /dev/mapper/sda6_crypt /var/local ext3 defaults,data=journal 0 2 /dev/mapper/sdb1_crypt /var/tmpext3 defaults,data=journal 0 2 /dev/mapper/sda2_crypt noneswap sw 0 0 tmpfs /tmptmpfs size=500m 0 0 /dev/hda/media/cdrom0 udf,iso9660 user,noauto 0 0 /dev/fd0/media/floppy0 auto rw,user,noauto 0 0 Thus, its totally transparent to the normal user. I have to enter the pass phrases at boot; I've been too lazy to set it up to only require one, so I have to do it three times (one for each encrypted partition) (swap doesn't need a passphrase). LUKS is flexible enough that you could use a USB key if you like. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Debian's glacial movement--a rant
On Mon, Apr 06, 2009 at 01:06:00PM -0700, Freddy Freeloader wrote: Michael Biebl wrote: Freddy Freeloader wrote: I'm experiencing a bug in Gnucash that appeared a couple of days ago on my system that makes Gnucash completely unusable for me. I turned in a bug report on Friday, checked on it yesterday, and by today the bug had been blocked from being displayed. It could be found by searching Debian's bug tracker, but only if you know the bug id number. If you just search for bugs in Gnucash the bug does not appear to exist. The bug was closed, and blocked, because it's been fixed upstream in version 2.2.9 which was released by Gnucash in February of this year. Could you discuss how you're experiencing a new bug (appeared a couple of days ago) if you're not using gnucach in a new way? I'm sure there are bugs in every piece of software I use. There are a couple of which I am aware (if I care to think on it) but I have work-arounds or I wouldn't have accepted the software for use and would have chosen a different solution. I've never run into the appeared a couple of days ago situation where the problem was a new bug; the problem has always been in a different system (or is upstream of the keyboard :)) Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
USB PCI card to buy
I need to add USB to my HP NetServer LPr. Do they all just work now? I'm looking at either the Belkin F5U220v1 5-port (4+1) which I think has an NEC chipset, or one of the Startech's: PCI625USB21 6-port (4+2) with nvidia chipset PCI330USB2 4-port (3+1) with NEC D7201026c chipset PCI 220USB 3-port (2+1) with VIA VT6202 chipset. in sort of this order of preference. Any suggestions? Things to avoid? I don't need the internal ports since this is a 2U server with no front bays (2 hot-swap scsi, a floppy, a slim CD). Thanks, Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB PCI card to buy
On Tue, Apr 07, 2009 at 10:52:41PM +0300, Dotan Cohen wrote: I need to add USB to my HP NetServer LPr. Do they all just work now? It's been a _long_ time since I've had a problem with a PCI USB card. If nobody pipes up with a negative, then I'd suggest that you give it a shot. The odds are in your favor. By the way, you may want to write to Belkin and ask them if the card will work with Linux. At the very least, you are expressing interest in having Belkin develop for Linux. They will never write works with Linux on the box if nobody is asking for it. Here is their address: http://www.belkin.com/contactus/index.asp Done, waiting for reply. Other than any compatibility issues, is one chipset better than another? Do some cards work better? The main purpose of adding USB is for using USB sticks, and syncing my palm. I may want to use a CD burner at some point (right now, if I want to burn a CD, I have to get my old box out of storage; the NetServer doesn't have any free bays). Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: USB PCI card to buy
On Tue, Apr 07, 2009 at 11:00:57PM +0200, Johannes Wiedersich wrote: Dotan Cohen wrote: By the way, you may want to write to Belkin and ask them if the card will work with Linux. They will never write works with Linux on the box if nobody is asking for it. +1 ;-D ??? please translate. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Decrypt, Edit and Encrypt a File
On Tue, Apr 07, 2009 at 11:16:37PM +0200, Samuel B?chler wrote: Thanks a lot to everyone! Just as a short description: I installed easypg. In the console environment I type `emacs keys.pgp'. This starts emacs and prompts for passphrase of `keys.pgp'. After entering the passphrase you can edit your file with the passwords. When you are finished with altering the password-file type `ctrl+x ctrl+s' and close emacs. Where does it hold the decrypted data? Does it stay in RAM, does it get swapped, does it go to a scratch file? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Decrypt, Edit and Encrypt a File
On Tue, Apr 07, 2009 at 05:46:31PM -0400, H.S. wrote: Douglas A. Tutty wrote: Where does it hold the decrypted data? Does it stay in RAM, does it get swapped, does it go to a scratch file? This might help: http://www.easypg.org/ yea, it looks like it can leak info. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Debian won't boot
On Tue, Mar 31, 2009 at 06:59:43PM -0700, Vwaju wrote: For months, I have been booting Debian 3.1 every day and experimenting with networking tools. Today I didn't do much except read man pages, and I'm not aware of doing anything to change any configuration, but when I rebooted my computer it stalled out at INIT: Entering runlevel 2 After it hangs here, it starts printing out messages of the form: Out of Memory: Killed process 2253 (sysklogd) If I scroll back on the monitor, there are a number of messages like this: modprobe: FATAL: Error inserting pciehp modprobe: FATAL: Error inserting shpchp modprobe: FATAL: Error inserting hw_random modprobe: FATAL: Error inserting rtc What is going on here? Is there any way to boot in this circumstance? Since it hangs after entering runlevel 2, you may try booting single (runlevel 1). If it still hangs, try booting with init=/bin/sh which will not run any initscripts. You can then run them in order with a 'start' paratmeter. This way, you'll figure out exactly where the problem is. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Decrypt, Edit and Encrypt a File
On Sun, Apr 05, 2009 at 05:36:08PM +0200, Samuel B?chler wrote: I store logins and passwords of some dozen of Web-Services in an encrypted file. I used to use kgpg to read and update this file. Some weeks ago I found on debian-security [1] the following script: #!/bin/sh gpg keys.gpg /dev/null emacs keys rm keys.gpg gpg -r user-ID -e keys rm keys What do you guys think is this approach reasonably secure? I like the script above because it is rather simple. You may want to set up your system with encrypted swap, then put /tmp on tmpfs (so that /tmp is encrypted). Then use the libpam-tmpfile (I think that's what it's called) so that every usere automatically gets their own directory in /tmp as $TMP. I use openssl to encrypt and decrypt files, and if I'm not on a system with encrypted /home, I keep the decrypted files in $TMP. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Several users need to work in the same directory
On Sat, Mar 28, 2009 at 05:37:07PM +0100, Rico Secada wrote: I know that the user can chmod the files he just created manually and then set the execution bit, but in our case the developers sometimes upload several new files using sftp, and I am wondering if they really have to logon again using ssh and then manually set the execution bit on every single newly uploaded file? sftp doesn't do well with permission bits. What about using rsync instead? The file is still transferred with ssh but all the attributes are preserved. Whould sudo help? Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Question about get-selections
On Thu, Mar 26, 2009 at 04:26:12PM -0600, Paul E Condon wrote: But I think I would like to have a record of what packages were actually installed. So I'm thinking of writing a script, to be run nightly, that puts a fresh copy of my selections in /etc/apt, e.g. # dpkg --get-selections /etc/apt/selections Many of those packages will have been installed automatically by your package manager. If you use aptitude, you only need to record the packages which you manually installed: aptitude search `~i!~M' You can then install them, then aptitude will automatically install what is needed by them. I wouldn't put my backup info in /etc/apt. Most of what is in /etc/ (just like the rest of the file system) is managed by packages which are managed by apt. You may find that what you put in /etc/apt gets changed (I don't know, but I wouldn't risk it). I keep my backups in /var/local/backup. Your remote script can run a local script to create the backups, then rsync what you want to the remote (or any other) box. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: [SECURITY] [DSA 1751-1] New xulrunner packages fix several vulnerabilities
On Mon, Mar 23, 2009 at 07:57:11PM +0200, Andrei Popescu wrote: On Sun,22.Mar.09, 18:35:21, wrote: Hello all, I'm running Etch, and use Iceweasel. I'm concerned about this security advisory. It says that the Etch release notes said that the Mozilla products would have to be stopped prior to the end of the Etch support period. I don't see this. Here it is: http://www.debian.org/releases/oldstable/i386/release-notes/ch-information.en.html#s-mozilla-security That, again, is just like in Lenny, where they say that, at some point in the future, security support may be dropped. They still do the security support for Lenny, but they didn't announce dropping it for Etch. For how long have I been running a (knowingly) insecure Iceweasel? I'm glad I use a different user for it. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: lenny reinstall
On Tue, Mar 24, 2009 at 09:38:23AM -0600, postid wrote: I need to reinstall lenny using netinstall. (I had some problems with the network, etc. during the first attempt.) I want to keep the partitioning I currently have set up. I especially don't want to risk messing up hda1 since I promised the user that I'd give them the option of booting either system. To prevent having network problems mess up an install, I always just install a minimal system (don't choose anything from tasksel). Then run aptitude and install what I want. If you want to use tasksel, just install it later. hda1 win2000 hda2 / hda5 /swap hda6 /home Will the partitioner give me the option of reusing the current scheme and formatting over the old data on hda2 and hda6 or do I need to delete the contents of hda2 and hda6 using a live distro (or the installer command line?) before I start the install? I seem to recall reading somewhere that Debian won't install over itself. Any pitfalls to watch for here or am I just paranoid? I don't want to lose the contents of hda1. I'm planning on using the graphical installer and manual partitioning. It should present the existing partitions as they are. You'll have to select each one and tell the installer what to do with each, e.g. hda2 use as ext3, mount point / hda5 use as swap (not /swap) hda6 use as ext3, mount point /home You'll have to tell it to reformat the partitions. Don't tell it to do anything with hda1. It should leave it alone. Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: SOLVED Re: Lenny won't install on an old Pentium that used to run Etch. Try 2
On Sun, Mar 22, 2009 at 05:11:43AM -0700, Thorny wrote: On Sat, 21 Mar 2009 16:50:32 -0400, Stefan Monnier posted: Discussing this has inspired me to put another line on my hobby list, I will eventually drag out an old P1 100MHz I have and try loading Lenny on it. Or, maybe I shouldn't thank you for that, it's not like I don't already have enough projects. chuckle FWIW, I find that running Lenny on a 64MB machine is bearable but slowish, and Etch was already too slow on a 32MB machine. So make sure you have enough RAM. For very simple uses, it works with less RAM, but as soon as you need to apt-get install or apt-get upgrade the memory limit becomes severe. It will be only a hobby system, slow would be acceptable. Thanks for sharing your experience. However, consider that a lot of the slowness is due to heavy swap usage. On an old system, with old drives, it may cause concern for the old IDE drives. My old hobby systems run OpenBSD: much faster (even when not swaping). Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
accessing www.mls.ca map search from linux
Hi all, We're looking for a new house and here in Canada, the way to do this (other than getting a realestate agent to do it), is to go to www.mls.ca and do a search. MLS has recently changed how you choose the geographic area: it used to be by clicking on a simple map to choose the realestate board in which to search. Now, its a MS-based map. Unfortunatly, at least one realestate company has gone to the same search method on their own web site. I get the map OK, but I don't get any results. The help page says that only IE on MS is supported (it doesn't mention Linux, but does say that MacOS is not supported). I've tried it with (Etch) iceweasel and konqueror. Could someone who's running Lenny see if it works from Lenny? FYI, it will only show less than 500 results. If you zoom in the map, it will tell you how many results there are. When this gets to less than 500, it should show them to you. If anyone has any ideas, I'm all ears. Thanks, Doug. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
