Re: [SLUG] Does Berkeley error still exist?
Rick Welykochy wrote: Luke Kendall wrote: Does anyone know whether it's available for Linux? I haven't been able to find it via synaptic, rpmfind, apt-cache search, or google searches. The best I can find are the man pages. I also see that it's included in the developer tools for MacOS X. Any relation to perror() ? http://www.penguin-soft.com/penguin/man/3/perror.html?manpath=/man/man3/perror.3.inc No, none whatsoever. :-) perror() is a system call for use in programs. error is an executable for use from the command line or scripts. If you have a program that processes a source file (e.g. gcc and .c files, or groff and the Mss for an article), and the program reports lots of errors, then error can edit the source file to insert the etxt of each error just before the line that has the error. Then you edit the source file, visiting each error and fixing it up. This saves you from flipping between two files. It's one of those traditional Unix small utilities that does one thing well. Regards, luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Does Berkeley error still exist?
Rick Welykochy wrote: Luke Kendall wrote: Rick Welykochy wrote: Any relation to perror() ? http://www.penguin-soft.com/penguin/man/3/perror.html?manpath=/man/man3/perror.3.inc No, none whatsoever. :-) perror() is a system call for use in programs. error is an executable for use from the command line or scripts. If you have a program that processes a source file (e.g. gcc and .c files, or groff and the Mss for an article), and the program reports lots of errors, then error can edit the source file to insert the etxt of each error just before the line that has the error. Then you edit the source file, visiting each error and fixing it up. This saves you from flipping between two files. It's one of those traditional Unix small utilities that does one thing well. An interesting detective challenge. I began searching on google for the following: error utility error message source code which produced 240 results. Chasing up one on the SLUG list ... http://lists.slug.org.au/archives/slug/2003/05/msg00896.html * To: Sydney Linux Users Group [EMAIL PROTECTED] * Subject: [SLUG] Anyone know where to find the error utility? * From: [EMAIL PROTECTED] * Date: Sat, 24 May 2003 15:08:44 +1000 (EST) Further searching turned up this page in the CVS attic for the BSD utility: http://www.freebsd.org/cgi/cvsweb.cgi/src/usr.bin/error/Attic/ And this: http://lists.slug.org.au/archives/slug/2003/05/msg00909.html where you proposed porting the thing to Linux. And that is where I stopped searching. Good luck. Perhaps you can find the sources you need in the Attic. cheers rickw Strewth! You're right! When I followed teh attic link, and started preparing the natural place to store the source files, I found the directory already existed with all the sources in it already! I guess that means I'd better get off my behind, and have a go! Thanks, Rick. :-) luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Does Berkeley error still exist?
In Berkeley unix 4.0, many years ago, Robert Henry wrote a useful utility for developers called error, which could be used to insert the body of error messages into the source files, to make it easier to find and fix the errors. Does anyone know whether it's available for Linux? I haven't been able to find it via synaptic, rpmfind, apt-cache search, or google searches. The best I can find are the man pages. I also see that it's included in the developer tools for MacOS X. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Unix printer driver breakthrough
I think it was because of the Unix wars that Unix never succeeded in producing a single unified print driver environment. Printing was one area where Windows did it better. So this very good news (http://www.desktoplinux.com/news/NS9747520848.html) was arguably achieved because of Linux. And that announcement in turn grew out of the work at the open printing summit last year: http://www.linux-foundation.org/en/OpenPrinting/SummitLexington and of course all the hard work done in recent years that lead up to it. BTW, I know that people on the SLUG list have occasionally asked about Canon's support for Linux, so here is a presentation from the summit that may be of some historical interest: http://www.linux-foundation.org/images/9/9b/2006-10-23-Linux_Printer_Drivers_from_Canon_061022-1.pdf Toratani-san is a nice guy. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] [Printing-summit] ANNOUNCE: Gutenprint 5.0.1 Release
Just thought this might interest some people. luke Original Message Date: Sun, 17 Jun 2007 15:23:27 -0400 From: Robert L Krawitz [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: [Printing-summit] ANNOUNCE: Gutenprint 5.0.1 Release Gutenprint 5.0.1 is a stable release of Gutenprint 5.0. It contains a large number of bug fixes and enhancements. Gutenprint 5.1.3 is also released. It is a development release. It has very little functionality above that of Gutenprint 5.0.1. II) MAJOR CHANGES FROM PREVIOUS RELEASES A) NEW FEATURES AND FIXES IN GUTENPRINT 5.0.1: 1) New printers supported in this release: * Dye sublimation printers using the Olympus driver: Canon CP-10 Fujifilm FinePix NX-500 Kodak Easyshare Printer Olympus P-S100 Sony DPP-EX5 Sony UP-DR100 * Canon inkjet printers: The support for these printers is under development, and there may be issues with these printers. Please check with the mailing list ([EMAIL PROTECTED]) if you have any questions. Canon PIXMA iP2000 Canon PIXMA iP3000 Canon PIXMA iP3100 Canon PIXMA iP4100 Canon PIXMA iP4200 Canon PIXMA iP4300 Canon PIXMA iP5000 Canon PIXMA iP5200 Canon PIXMA iP6700 Canon PIXMA MP150 Canon PIXMA MP500 Canon PIXMA MP700 Canon PIXMA MP730 Canon PIXMA MP750 Canon PIXMA MP760 Canon PIXMA MP770 Canon PIXMA MP780 Canon PIXMA MP790 Canon PIXMA MP830 Canon PIXUS iP3100 Canon PIXUS iP4100 Canon i560 Canon i850 Canon i860 Canon i865 * Epson inkjet printers: Epson Picturemate Flash Epson Picturemate Pal Epson Picturemate Snap Epson Picturemate PM-200 Epson Picturemate PM-210 Epson Picturemate PM-240 Epson Picturemate PM-250 Epson Picturemate PM-280 Epson Stylus C79+ Epson Stylus C87+ Epson Stylus Photo 1400 Epson Stylus Photo 1410 Epson Stylus Photo R230 Epson Stylus Photo R240 (preliminary) Epson Stylus Photo R245 (preliminary) Epson Stylus Photo R260 Epson Stylus Photo R265 Epson Stylus Photo R270 Epson Stylus Photo R350 Epson Stylus Photo R360 Epson Stylus Photo R380 Epson Stylus Photo R390 Epson Stylus Photo RX560 Epson Stylus Photo RX580 Epson Stylus Photo RX590 Epson Stylus Photo RX640 Epson PM A820 Epson PM D870 Epson PM G850 Epson PM G4500 Epson Stylus CX5000 Epson Stylus CX5000F Epson Stylus CX6000 Epson Stylus CX7000F The driver for the Stylus Photo 1400 has two separate quality settings for 720x360 and 720 DPI. The 720x360 DPI Enhanced and 720 DPI High Quality options offer improved quality at some cost in printing speed. The small format Claria-based printers (R260, R390, RX580, and related printers) offer 720x360 and 720 DPI modes approximately equal in quality to the Enhanced and High Quality modes on the 1400 without a speed penalty. * Lexmark inkjet and compatible printers: Compaq IJ1200 Lexmark X73 * PCL laser printers (monochrome only): Lexmark Optra E220 Xerox WorkCentre M118 2) CUPS 1.2 is now supported using on-the-fly PPD file generation. In addition, the resolution names are all compliant with the PPD specification. cups-genppdupdate correctly updates on-the-fly PPD files. The PPD files now use the correct UTF-8 encoding. 3) The native CUPS driver offers a new Shrink Page If Necessary to Fit Borders option, enabling the user to choose how to fit the output to the imageable area of the page. This is useful when a printer offers a choice of imageable areas, typically normal (which has margins) and full bleed (which allows printing to the edge of the paper, or even beyond). The following options are available: * Shrink (default): the output is shrunk if necessary to fit the imageable area of the page. If a printer is capable of borderless operation but normal margins are selected, the output will be shrunk. This will print the entire page (nothing will be lost), but will not preserve the dimensions of the printout. For example, a line intended to be 10 cm long may print smaller than that. * Crop: the output is cropped if necessary to fit the imageable area of the page. The dimensions of the page will be preserved (a line intended to be 10 cm long will print out exactly 10 cm), but the edges of the output may be truncated (cropped). * Expand: the output is expanded from the normal imageable area
Editing PDF [Was: Re: [SLUG] Joining and a Question - again]
David Bowskill wrote: Dear Slug, [snip] A question - does there exist an open source pdf editor/writer for Linux - which can open and edit existing pdf files ? I am presently using Xandros for my every day activities and such a editor/writer would be useful. Searches of various sites have yielded no joy in this regard. PDF is designed not to be editable. It's meant to be a read-only format. (That's not 100% true, but close.) You're meant to keep your documents in whatever the original editable format was, and convert to PDF to publish them - this ensures they will continue to look exactly as you designed them even after they're distributed. I assume you created them using some word processing application? If so, you should really try to find something that can import *those* files. But all that said, the pstoedit package does a pretty good job of converting PDF into editable formats. (This is quite challenging when you consider the PDF could consist purely of diagrams, and that the text can be painted onto a page in random order.) On my Ubuntu system, one of the output formats for pstoedit is .svm: StarView/OpenOffice.org metafile, and others are plain ascii or .fig or .pic etc. etc. etc. HTH, luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] How to downgrade to an older package with dpkg?
I've been trying to get backups to dvd working on a new computer (Ubuntu 6.something), and I can't mount the DVDs I write (not on the new computer or the old one). I have dvd+rw-tools Version: 6.1-2ubuntu1 and I'd like to try version 4.7, since I know that works (on an older computer). But I can't see a way to use dpkg to downgrade a package. If I try in synaptic, the Package-Force Version menu option is greyed out. Does this mean I need to build it from source manually? Oh: I just noticed this note in the dpkg man page: --force-things | --no-force-things | --refuse-things Force or refuse (no-force and refuse mean the same thing) to do some things. things is a comma separated list of things specified below. --force-help displays a message describing them. Things marked with (*) are forced by default. Warning: These options are mostly intended to be used by experts only. Using them without fully understanding their effects may break your whole system. all: Turns on(or off) all force options. downgrade(*): Install a package, even if newer version of it is already installed. Warning: At present dpkg does not do any dependency checking on downgrades and therefore will not warn you if the downgrade breaks the dependency of some other pack‐ age. This can have serious side effects, downgrading essential system components can even make your whole sys‐ tem unusable. Use with care. So maybe building from source and not installing the tools is the best thing to do by far. What's the chance that the dvd burner isn't supported under Linux (it's a dual layer burner: LG Super multi DVD RW +R DL)? luke --- Details in case anyone is interested -- When I use dvd+rw-mediainfo on the new computer, it says: INQUIRY:[HL-DT-ST][DVDRAM GSA-H10N ][JL10] :-( no media mounted, exiting... If I try that on the old computer I get: INQUIRY:[SONY][DVD RW DW-U14A ][1.0b] GET [CURRENT] CONFIGURATION: Mounted Media: 1Ah, DVD+RW Current Write Speed: 4.0x1385=5540KB/s Write Speed #0:4.0x1385=5540KB/s etc. If I try dvd+rw-format on the new computer I get: # dvd+rw-format -force /dev/dvd * DVD±RW/-RAM format utility by [EMAIL PROTECTED], version 6.1. :-( mounted media doesn't appear to be DVD±RW or DVD-RAM If I run dvd+rw-format on the older computer I get: # dvd+rw-format -force /dev/dvd * DVD�RW format utility by [EMAIL PROTECTED], version 4.7. * 4.7GB DVD+RW media detected. * formatting 1.0- etc., i.e. it works just fine. So I thought I'd try downgrading the dvd+rw-tools to version 4.7 and see if that helps. My other thought is that maybe Linux can't drive the dvd burner properly? -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] How to downgrade to an older package with dpkg?
On 2007-03-24 18:05:27 +1100 Luke Kendall [EMAIL PROTECTED] wrote: I've been trying to get backups to dvd working on a new computer (Ubuntu 6.something), and I can't mount the DVDs I write (not on the new computer or the old one). I have dvd+rw-tools Version: 6.1-2ubuntu1 and I'd like to try version 4.7, since I know that works (on an older computer). I've solved the problem. Version 6 of the dvd+rw-tools don't support dual layer burners, I think, from this change note in growisofs.c: * 7.0: * - Blu-ray Disc support [upon release tested with Panasonic SW-5582]; * - Mac OS X 10=2 support [upon release tested on 10.4 only]; * - Linux: copy linux/raw.h definitions directly into application * code in order to secure backward compatibility; * - Linux: overcome 16MB O_DIRECT limitaton for NFS client; * - limit ring buffer size to 1/4 of RAM; * - copy volume descriptors if -C was specified with -M /dev/dvd=image * [by request from K3b]; * - -use-the-force-luke=spare[:none|min] to control blank BD pre-format * [for finer control use dvd+rw-format]; * - some units, e.g. Lite-on SOHW-1693S, seem to fire off OPC already * upon Layer Break command, therefore longer timeout is required; * - Linux: deploy BLKFLSBUF to avoid media reloads when possible; * - add unit buffer utilization indicator [by request from K3b]; When I built dvd+rw-tools-7.0 from source and tried that version, everything worked: it properly sensed when a dvd is inserted (i.e. dvd+rw-mediainfo reported correct info using the dual layer burner, and I could burn a small set of files and then mount the dvd correctly afterwards). luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Where is your /usr/local?
I like to have /home on a separate partition to the root filesystem, so that if I want to upgrade the distro then I can pretty much replace the whole root filesystem. (I usually make a new / and install to that, and change over when all is looking good.) But I just realised that unless you make a separate partition for /usr/local (or equivalent, like making it a symlink to /home/local), then if you clobbered the root filesystem you'd also lose all your local files, which are more on a par with /home. Just a thought. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Hosed Ubuntu by apt-get upgrade
On 22 Jan, Robert Collins wrote: On Sun, 2007-01-21 at 20:22 +1100, [EMAIL PROTECTED] wrote: On 21 Jan, luke wrote: Any advice would be most welcome. Dont do 'apt-get update; apt-get upgrade' - for two reasons: Firstly, a failed update is not something you want to upgrade from, so doing 'apt-get update apt-get upgrade' would be safer. Secondly, 'apt-get upgrade' does not add or remove packages, and when a kernel version changes, you may well want additional related packages. Its better to do 'apt-get dist-upgrade' for incremental upgrades within a distribution. In terms of recovering your system, it should be as simple as booting a live CD. When the prompt appears, do 'live root=/dev/md0' or something similar - I can't remember the exact syntax offhand - but that will get you booted. Thanks for the reply, Robert. (It's been a rough week at work, so I hadn't had a chance to get back to the problem until today.) I've tried your suggestion, after I worked out how to supply the boot options (hit F6), but despite changing the root=/dev/ram to root=/dev/md0, eventually I get to a screen where it asks me to choose the device to use for the root partition (instead of using /dev/md0), and the only choices are the raw partitions. I *could* choose /dev/discs/disc0/part7 (/dev/hda7, which is half of /dev/md0), but I worry that doing so would mess up the raid.. Initially I forgot to insert live and just changed root=/dev/ram to root=/dev/md0. Later I added live as the first of the boot options. It didn't seem to make much difference. Either way it ends up asking me the same question. Once you are running, use aptitude and check that all the recommended packages for the new kernel version are present, including udev. You can then use 'update-initramfs' and 'update-grub' to regenerate the initramfs which contains a copy of udev, and *that* should let you boot properly again. Sounds great, if I could only get that far. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Hosed Ubuntu by apt-get upgrade
Yesterday I did an apt-get update; apt-get upgrade on my Ubuntu 6 system. Apparently it fiddled with the kernel (I had 2.6.15-26-386 installed and it I think installed 2.6.15-26-686.) I noticed today that it suggested a reboot. I had a look at /boot/grub/menu.lst which had been modified (according to the modification time), but I couldn't pick the change. It all looked good, and I have numerous alternative kernels to fall back to. Upon attempting to reboot, I got this message: Uncompressing Linux... Ok, booting the kernel mdadm: /dev/md0 has been started with two drives. mdadm: /dev/md1 has been started with two drives. mount: Mounting /dev/hda7 on /root failed: Device or resource busy mount: Mounting /root/dev on /dev/.static/dev failed: No such file or directory mount: Mounting /sys on /root/sys failed: No such file or directory mount: Mounting /proc on /root/proc failed: No such file or directory Target filesystem doesn't have /sbin/init Obviously if I can't mount /root then it won't be able to mount /root/dev. It drops into a shell, and I saw that there's nothing really mounted (I think it's just an initrd file system). I can mount /dev/md0 (which is /) to a temporary mount point and everything's there. But I can't boot *any* of my Linux kernels. This suggests to me that it's not a problem with the kernel, it's a problem with some critical component needed by all the kernels, in the root file system. A google search leads me to http://www.linuxquestions.org/questions/showthread.php?t=446745 which suggests that in April last year udev was broken badly: Got it!!! I downloaded the testing version of udev (0.091-2) from packages.debian.org and installed it via chroot, and that did it. The system booted right up. Obviously there's a bad bug in udev version 0.092-1. I recommend everyone stay away from this one. How do I query what version of udev I have installed? Why is mount trying to mount one of the partitions that makes up /dev/md0 instead of mounting /dev/md0? /dev/md0 is /dev/hda7 and /dev/sda7, and /etc/fsab has /dev/md0 mounting on /. Any advice would be most welcome. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Google for trends
Did you know you could use Google to look at trends? E.g. Ubuntu *apparently* overtook MacOS/X over a year ago: http://www.google.com/trends?q=ubuntu%2C+osxctab=0geo=alldate=all luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Basically-solved exim mail problem
I guess I'm sending this because altjhough I've sorted the problem out, I'm such a novice when it comes to mail delivery systems that I don't fully understand what went on, so if anyone cares to enlighten me, the info would be very welcome! After a bit of a struggle, I got my wife's Linux machine sending and receiving mail using exim and fetchmail, about a week ago. But yesterday she stopped receiving any email - which was nice to receive no spam, but not so nice to not even receive test emails from herself to herself or from me to her, or from friends. Her outgoing mail seemed to be stuck, too. But not entirely: email sent from her to me (same ISP, obviously!), worked. When I ran mailq I saw lots of items which I think were mails she'd sent which had not been delivered. There were about a dozen messages e.g. like this one: 3d 1.3K 1GhcMT-0002kx-4Y [EMAIL PROTECTED] [EMAIL PROTECTED] When I tried to force a delivery (with debug turned on) via: exim -qff -d+all I'd see things like this (below), but the message still just sat there (which seemed to me to be saying that it *had* timed out on connecting to mail.usyd.edu.au, but that it had then carried on and seemed to be getting somewhere): 18:34:18 25648 Connecting to mail.usyd.edu.au [129.78.220.2]:25 ... failed 18:37:27 25648 LOG: MAIN 18:37:27 25648 mail.usyd.edu.au [129.78.220.2]: Connection timed out 18:37:27 25648 set_process_info: 25648 delivering 1GhcMT-0002kx-4Y: just tried mail.usyd.edu.au [129.78.220.2] for [EMAIL PROTECTED]: result DEFER 18:37:27 25648 ---0 Get 0x810589824retry.c 291 18:37:27 25648 added retry item for T:mail.usyd.edu.au:129.78.220.2: errno=110 333 flags=2 18:37:27 25648 ---0 Get 0x81058b024 string.c 347 18:37:27 25648 ---0 Get 0x81058c832 string.c 347 18:37:27 25648 address match: [EMAIL PROTECTED] pattern=* 18:37:27 25648 ---0 Get 0x81058e824 string.c 387 18:37:27 25648 ---0 Get 0x810590032 string.c 347 18:37:27 25648 mail.usyd.edu.au in *? yes (matched *) 18:37:27 25648 [EMAIL PROTECTED] in *? yes (matched *) 18:37:27 25648 checking status of mail.usyd.edu.au 18:37:27 25648 ---0 Get 0x810592032 string.c 347 18:37:27 25648 ---0 Get 0x810594056 string.c 347 18:37:27 25648 locking /var/spool/exim/db/retry.lockfile 18:37:27 25648 locked /var/spool/exim/db/retry.lockfile 18:37:27 25648 opened hints database /var/spool/exim/db/retry: flags=0 18:37:27 25648 dbfn_read: key=T:mail.usyd.edu.au:129.78.220.1 18:37:27 25648 ---0 Get 0x810597856 dbfn.c 291 18:37:27 25648 dbfn_read: key=T:mail.usyd.edu.au:129.78.220.1:1GhcMT-0002kx-4Y 18:37:27 25648 no message retry record 18:37:27 25648 mail.usyd.edu.au [129.78.220.1] status = usable 18:37:27 25648 ---0 Get 0x81059b040 string.c 347 18:37:27 25648 129.78.220.1 in serialize_hosts? no (option unset) 18:37:27 25648 delivering 1GhcMT-0002kx-4Y to mail.usyd.edu.au [129.78.220.1] ([EMAIL PROTECTED]) 18:37:27 25648 set_process_info: 25648 delivering 1GhcMT-0002kx-4Y to mail.usyd.edu.au [129.78.220.1] ([EMAIL PROTECTED]) 18:37:27 25648 ---0 Get 0x81059d888 expand.c 2556 18:37:27 25648 ---0 Rst 0x81059d8** expand.c 2632 16400 18:37:27 25648 ---0 Get 0x81059d8 104 string.c 856 18:37:27 25648 ---0 Rst 0x81059e8** expand.c 4397 16400 18:37:27 25648 expanding: $primary_hostname 18:37:27 25648result: cor.localdomain 18:38:00 25647 selecting on subprocess pipes An odd thing I notice is that a ping from my wife's machine (say to www.google.com.au) typically reports packet loss of up to 20%, whereas from mine, 0%. Both are plugged into a 4-port firewall router. Any suggestions about how to track the problem down? I've tried restarting exim, and the network. Ah, some more reading turned up info on using your ISP's mail server as your smart host, like so in the exim.conf file: # If you want to send all mail for non-local domains to a smart host, # you should replace the default dnslookup router with a router which # does the routing explicitly: send_to_smart_host: driver = manualroute route_list = !+local_domains mail.optushome.com.au transport = remote_smtp And restarting exim after that, all the pending mail got flushed out. So I gather I had it set up to try to connect directly to the mail server for each domain for each message recipient, and trying to talk to the smtp server directly to deliver it, and for some reason that was failing for a lot of them. Though I don't know why it was working fine for a week or two, until a few days ago. Any ideas? (Just curious.) Incidentally, is there anywhere that describes the format of the output from the mailq command? (The hours queued, message size, message ID, sender and recipients are obvious, but what do the capital D marked lines mean?) I couldn't find it in the mailq commandline or the exim on line documentation. A sanitised
[SLUG] Two grub/RAID questions?
I'm halfway through setting up the root partition (and /home) to be raid1 mirrors (thanks to help from Jeff Waugh and Jamie Wilkinson). I've booted up from my old installation on /dev/hda6 to set up the real /dev/hda7,/dev/sda7 to be /dev/md0 for /, and /dev/hda8,/dev/sda8 to be /dev/md2 for /home. I've set the partition types on hda7, sda7, hda8 and sda8 to fd, Linux RAID autodetect. I've also installed grub on (hd0,6) and (hd1,6) (i.e. on /dev/hda7 and /dev/sda7). My two questions are: is a raid mirror device supposed to have a partition table? If so, what should it be set to - just one partition containing the number of blocks that the real partitions making it up have? What should the start and end cylinders be set to? (The reason I ask is that the raid devices are failing the filesystem check upon boot.) And secondly, how do you tell grub to use md0 instead of hd0 or hd1? Is it as simple as telling it that device (hd0) is /dev/md0? (Though I don't see a way to do that in /boot/grub/menu.lst) How I created the raid arrays - I basically unmounted the drives I wanted to turn into pieces of the raid, used mdadm to add the (empty) partitions from sda into the raid array, telling it that the other piece was missing, then mounted the raid drive and copied the files from the (populated) partition, then unmounted that populated partition and added it to the raid and watched as the raid rebuilt itself. I could then mount /dev/md2 as /home and ls files. My first question is: how do you tell grub what to use as the root drive, given that grub seems to want a physical drive spec (e.g. hd0 or hd1)? How do you tell grub to use md0 instead of hd0 or hd1? In my case, hd0 is /dev/hda and hd1 is /dev/sda. I've only created md0 from hda7 and sda7 and md2 from hda8 and sda8. I also have this in /etc/mdadm/mdadm.conf: DEVICE /dev/hda* /dev/sda* ARRAY /dev/md0 devices=/dev/hda7,/dev/sda7 ARRAY /dev/md2 devices=/dev/hda8,/dev/sda8 (My plan is that if I can get md0 working as /, then I'll setup md1 as another raid array for the older system I installed initially, on hda6. I'll use sda6 as its mirror.) But cat /proc/mdstat though shows already md1 exists (!) and it seems to be what I thought should be md2! Personalities : [raid1] md2 : active raid1 dm-5[1] dm-4[0] 121531584 blocks [2/2] [UU] md1 : active raid1 sda8[0] hda8[1] 121531584 blocks [2/2] [UU] md0 : active raid1 sda7[0] hda7[1] 12586816 blocks [2/2] [UU] Where did md1 come from? What device is dm-5 and dm-4? I should mention that when trying to create the md0 raid array I got an mdadm error something like Invalid parameter, and an mdadm -Ds showed that md0 had the same ID as md2. I had to do an mdadm --stop --scan before I could get the mdadm --create /dev/md0 --level=1 --raid-devices=2 /dev/sda7 missing to work. After it did, an mdadm -Ds showed that md0 had a different ID to md2 (and there was no md1 at that point, either). Though just now checking that again I see that mdadm -Ds shows that there is a /dev/md1 array and that it has the same UUID as /dev/md2! Also when I try to boot up now the system drops me into maintenance mode when it attempts to check /dev/md2 (/home) and fsck reports this error: /dev/md2: the filesystem size according to the superblock is 30,382,923 blocks. Physical size of the device is 30,382,896 blocks. Either the superblock or the partition table is likely to be corrupt! If I run fdisk on /dev/hda or /dev/sda it tells me that hda8 and sda8 have 121,531,693+ blocks. If I run fdisk on /dev/md2 it says it contains no partition table, and that it has 30,382,896 cylinders. I've previously copied /home to a spare partition, so wasn't too frightened of running fsck on /dev/md2. It seemed to complete fine, reporting no errors, but it always reports the filesystem size according to the superblock is 30,382,923 blocks. Physical size of the device is 30,382,896 blocks. Either the superblock or the partition table is likely to be corrupt! Aborty? If I mount /dev/md2 /home, everything looks good - all the files seem to be there. So my two questions are: is a raid mirror device supposed to have a partition table? If so, what should it be set to - just one partition containing the number of blocks that the real partitions making it up have? What should the start and end cylinders be set to? And secondly, how do you tell grub to use md0 instead of hd0 or hd1? luke -- For reference: -- From: Jeff Waugh [EMAIL PROTECTED] Subject: Re: [SLUG] Adding RAID to Linux *after* installation? Date: Sun, 30 Jul 2006 07:03:37 -0700 To: slug@slug.org.au quote who=[EMAIL PROTECTED] mdadm --create /dev/md0 --level=1 --raid-devices=2 /dev/hda7 /dev/sda7 it
[SLUG] Ubuntu, 2.6.15 kernel, RAID and SATA and IDE
This relates to my attempt to move from my now too-long-in-the-tooth RH 7.2 system to something new. I picked Ubuntu and installed it without much trouble, highly impressed by how well it auto-detected and configured the network, sound, and graphics (to a point). It's basically just to sketch out my failure to build a kernel to meet my needs, in case it's useful to others doing Google searches, so they feel they're not alone. What I want to do is something that seems should be the norm, not something unusual: I want RAID mirroring. I have both an IDE drive and a SATA drive. (The PC bumph said it was a SATA drive system, so I bought an extra SATA drive. What it didn't say was that it was shipped with an IDE drive and was actually SATA-supporting.) I plug in the SATA drive and manage to partition it the same way I did the IDE drive (note: until you partition the drive *and reboot* the entries for the /dev/sda etc. won't appear so you can't do mkfs of any sort). After making the filesystems, preparatory to doing the mirroring, I discover the Ubuntu kernel image I'm using (2.6.15.26), doesn't appear to have md support. So I need to build a new kernel because I'll need RAID compiled into the kernel along with ext2 and ext3 and IDE drivers and SATA drivers. First thing to do is to learn is how to build a 2.6 series kernel. I discover that to avoid the usual nuisance of nvidia's proprietary graphics driver, there are Ubuntu/Debian packages for automatically performing this for you, if you install the nvidia-kernerl-common package, or something like that. Oh, and the Debian way is not to build the kernel directly, but to do the make config (gconfig or xconfig) parts and then run make-kpkg and then dpkg -i the kernel.deb image and everything should just work. So I do that and following the tips on Troubleshooters that told me all this, also did as it said and ran the mkinitrd command to make an initrd image too. Reboot - panic due to no devfs. Much searching leads to the discovery that mkinitrd and initrd are deprecated, and I should use mkinitramfs and initramfs. (Same boot loader syntax, though, namely initrd). So I do that and reboot and it looks okay up to the Okay, booting message and then the screen goes blank. That's a first. I wait a bit, then reset and decide to try again. This time, I notice that the HDD light is flashing, so it hasn't halted, and I decide to wait even longer. Eventually the failed to start X garbage-laden screen appears, and I let it show me the log files (garbage-decorated on all borders), and then try to see why. FATAL error: could not load the NVIDIA kernel module. The nvidia-glx module is installed, and can't be removed due to dependencies. Nor can it be refreshed, that I can see (I'm still a novice at Debian and dpkg). Nor can it be found if I try to modprobe it. So much for the Debian way of including the nvidia kernel modules automatically. I download the manual thing from the nvidia page and sh that to build the kernel modules, and that seems to sort things out. Then I discover that there's no SATA drive visible. No /dev/sd* So I go hunting around and dmesg has a whole bunch of errors about sata_via: unsatisfied external reference. It seems most of the SATA functions are missing. So back to make gconfig and eventually find the SATA modules buried inside SCSI if you turn on some other SCSI option that allows you to unfold the category they're buried inside, and I turn off all the drivers for all the other chipsets *not* in my PC (my best guess is that I have a VIA chipset SATA controller), because a google search has shown a post where my error messages turning up seems to be linked to having these other controllers compiled in as modules instead of turned off. Doesn't make a lot of sense to me, but I give in and build in the via choice into the kernel and turn off all the others and rebuild. This time it fails because of multiple definitions of some symbol like do_get_something_GFM in ide and sata, so I change one to have an x suffix and rebuild. The posters suffering this blame an untested patch by SuSE 10.1, but like I say, I'm using Ubuntu (6.06). Now it gets further, but fails because acpi_in_suspend is undefined. This is defined in drivers/something/acpi/sleep/main.c. From looking at the Makefile there I see that CONFIG_ACPI_SLEEP{_FS} aren't defined, and in fact I can't find a way to define them, so I manually edit the ..config file and add them. That fixes that problem, but now four other symbols are undefined. I make clean and try again. No good. acpi_in_suspend is undefined. So at this point, I've given up, severely depressed. rant All this is the same problem that's been there forever in the Linux kernel config system - one module depending on other modules, but no checking of those dependencies or even any clue that the dependencies even exist, in the make config stage - it's all spaghetti-ed through the source code. The
Re: [SLUG] A Ubuntu/Debian and a grub question
On 19 Jul, Ken Caldwell wrote: No, as I mentioned in an off list post, the bit of code in the MBR will contain a jump instruction which will jump into one of the /boot/grub directories where the stage 2 boot loader code will take over. The relevant menu.lst is the one in that directory. Ah, okay, and the grub-install uses the /boot/grub/menu.lst from the filesystem mounted as / as its data source. So if I run that from hda7, then it's that filesystem's copy of menu.lst. What you need to do is edit the first 446 bytes of hda appropriately :-) There are no doubt many ways of achieving this. If the computer has a floppy drive and you have a suitable GRUB boot floppy you can use it to boot the new installation. Once up and running edit /boot/grub/menu.lst to taste and then do sudo grub-install /dev/hda Okay Though the first link you provide (below), near the end of the page, gives the example of doing this by doing grub-install /dev/hda1. I believe your example is installing to the MBR of the drive, while the troubleshooter example assumes that the MBR on the drive is untouched, and will chain load to partition /dev/hda1? I gather in my case (root and / on /dev/hda7) I could do a similar thing by running grub and entering this: grub root (hd0,6) grub setup (hd0) This will overwrite the first 446 bytes with code pointing to /boot/grub. (depending on how you have set up your fstab, the /boot/grub on the other partition will either not be mounted at all or will show up as /some_path/boot/grub and will not be relevant) Others reading this may be interested in the following links They're great links! Though on this one: http://www.troubleshooters.com/linux/grub/index.htm These instructions seemed broken (surely the first copy is wrong?): [EMAIL PROTECTED] root]# mount /dev/fd0u1440 /mnt/test [EMAIL PROTECTED] root]# cp -p /usr/share/doc/grub-doc-0.93/menu.lst /mnt/test/boot/grub/stage1 stage2 [EMAIL PROTECTED] root]# cp -p /usr/share/doc/grub-doc-0.93/menu.lst /mnt/test/boot/grub/menu.lst.example [EMAIL PROTECTED] root]# umount /dev/fd0u1440 [EMAIL PROTECTED] root]# http://www.troubleshooters.com/linux/grub/grub.htm http://www.troubleshooters.com/linux/grub/specialboots.htm http://www.pcug.org.au/oss/wiki/pmwiki.php?n=PCUG.GRUBTroubleshooting Thanks, Ken (and James!), that all worked fine. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] A Ubuntu/Debian and a grub question
On 19 Jul, Carlo Sogono wrote: [EMAIL PROTECTED] wrote: 1) How will grub work out where to find menu.lst? How will it 'know' to use the updated /dev/hda7/boot/grub/menu.lst rather than the out of date /dev/hda6/boot/grub/menu.lst version? GRUB calls this its root device and info on where to find the root device is found in the MBR. Type grub in your console and then help root. I know. But there is /dev/hda6/boot/grub/menu.lst that says root is hda6, and /dev/hda7/boot/grub/menu.lst which says it is hda7. Which is used? And does anyone know where apt decides where root is? It doesn't get the info from the menu.lst used to boot the system, nor from fstab - it seems to be stored in a config file that was used during the install, I've discovered as a result of my copy of hda6 to hda7. James wrote: /boot/grub/menu.lst specifically shows what disks and partitions are used. you can specify root (hd0,1) bla bla or specify the root here (hd0,6)/boot/vmlinuz-bla fstab must be correct or it wont work (or worse) if / in listed as /dev/hda6 even though it is hda7 6 is used 7 is not and this explains your chaos. Yes, I know, and that's why I quickly fixed it. The best way to do this: boot knoppix fiddle your partitions fix the files eg fstab boot your system even easier have another partition/system. boot that and fix this. Well, yes, that's what I did, since I had two bootable Ubuntu installations. :-) The thing I'm trying to fix is apt messing up menu.lst because it has noted that it was originally installed to hda6 (I theorise). And I'm wondering which menu.lst grub will use to boot from. Or does it look for /boot/grub/menu.lst across all partitions and do a union of all the bootable kernels it finds? luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] A Ubuntu/Debian and a grub question
On 19 Jul, Tony Lissner wrote: The easiest way to set where grub looks for the stage* images might be to just re-install grub from the new partition. try unmounting /dev/hda6 and commenting it out in fstab, then from the new partition do: # sudo grub-install /dev/hda # sudo update-grub and it should find the new menu.lst or checkout the '--root-directory=' option install grub-doc info grub Thanks, Tony, I'll give that a try. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Is there a truly upgradable Linux distro?
AFAIK, no Linux distro is considered quite safe to upgrade from one release to the next (e.g. from SuSE 9.2 to SuSE 10.0, or FC 4 to FC 5). Wise people still routinely advise Install the new system on a spare partition, and switch over when it's properly installed and configured. The problem with this is that if you've tweaked things so that sendmail is running nicely, and you have all the RealPlayer and Flash 7 and innumerable video codecs installed, and your soundcard working well and the DVD burner (and TV card?) etc. etc. all working well - then you have to do all this work afresh on the new system, and that can take days. So: does anyone know of a Linux distro that is so easily managed and so well structured, that not only can you easily update all your packages (via apt or yum or whatever), but you can even upgrade the whole distro, 99.99% reliably? (And no, I don't really want to install BSD which can do this, I believe, because AFAIK Linux still has far greater hardware support and much faster development.) I suppose a halfway decent approach might be to mirror your old working system onto a spare partition, and *then* try running the upgrade on *that*. If it doesn't work, then you're no worse off, having only spent an hour or so installing/upgrading. I must be one of the few people on the planet still running RH 7.2. (I do it because I begrudge spending the days or weeks getting all the extra packages installed that I like.) But it's now too old, and really should be replaced. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] The mystery of the missing memory!
If I run top and hit M to get a listing sorted by memory usage, I see that 98% of memory is used up, as follows: 20% buffers 20% cached 23% non 0.0% processes 1% free Where is the missing 36%? Ah! I think to myself I have 52 processes other processes using '0.0%' memory. What if I guess they're each really using 0.04% memory Rats, that only explains another 2.6%. Any ideas? I'm stumped. $ top --version top (procps version 2.0.7) This is on an old RH 7.2 system using kernel 2.4.28. top interesting bits: 11:13pm up 74 days, 56 min, 1 user, load average: 0.10, 0.04, 0.01 84 processes: 77 sleeping, 2 running, 2 zombie, 3 stopped CPU states: 0.0% user, 0.3% system, 0.0% nice, 99.6% idle Mem: 255716K av, 251356K used,4360K free, 0K shrd, 53528K buff Swap: 265032K av, 237716K used, 27316K free 52428K cached PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND 1273 root 18 0 255M 10M 732 R 0.0 4.3 1012m X 2451 luke 18 0 13932 8864 2388 S 0.0 3.4 7:21 postilion.exec 1061 xfs9 0 16780 7492 336 S 0.0 2.9 0:49 xfs 17278 stella 9 0 6440 6440 4092 S 0.0 2.5 0:00 postilion.exec 15573 stella 9 0 6432 6432 4088 S 0.0 2.5 0:17 postilion.exec 15571 root 9 0 2236 2184 1480 S 0.0 0.8 0:07 sshd 5937 luke 9 0 12192 2096 1160 S 0.0 0.8 0:00 postilion.exec 745 ntp9 0 1924 1924 1732 S 0.0 0.7 0:00 ntpd 16392 root 9 0 1644 1644 1256 S 0.0 0.6 0:00 fetchmail 17160 stella 9 0 1416 1412 400 S 0.0 0.5 0:00 ispell 1290 luke 9 0 1868 1344 684 S 0.0 0.5 2:06 wmaker 18057 luke 9 0 2948 1164 428 S 0.0 0.4 105:16 smbd 17321 luke 18 0 968 968 756 R 0.1 0.3 0:00 top 1309 luke 9 0 1108 944 644 S 0.0 0.3 0:02 bash 5918 luke 9 0 1356 808 368 S 0.0 0.3 0:00 ispell 1307 luke 9 0 936 724 428 S 0.0 0.2 0:01 bash 1295 luke 9 0 1948 648 296 S 0.0 0.2 0:07 wterm 1296 luke 9 0 1956 640 352 S 0.0 0.2 0:09 wterm 625 rpcuser9 0 664 568 564 S 0.0 0.2 0:00 rpc.statd 1084 root 9 0 880 564 428 S 0.0 0.2 0:04 nmbd 572 root 9 0 608 552 512 S 0.0 0.2 0:03 syslogd 15901 root 9 0 508 508 436 S 0.0 0.1 0:00 portsentry 577 root 9 0 1352 496 452 S 0.0 0.1 0:00 klogd 15897 root 9 0 492 492 420 S 0.0 0.1 0:01 portsentry 597 rpc9 0 572 488 484 S 0.0 0.1 0:00 portmap 800 root 9 0 452 440 252 S 0.0 0.1 0:13 smartd 1297 luke 9 0 584 428 328 S 0.0 0.1 0:00 xcb 947 root 8 0 848 424 308 S 0.0 0.1 0:06 sendmail 10423 luke 9 0 668 420 300 S 0.0 0.1 0:00 wterm 818 root 9 0 528 400 336 S 0.0 0.1 0:04 sshd 1299 luke 9 0 468 364 260 S 0.0 0.1 45:07 wmmail 1079 root 9 0 688 296 208 S 0.0 0.1 0:00 smbd 874 lp 8 0 360 252 208 S 0.0 0.0 0:00 lpd 10424 luke 8 0 592 240 236 S 0.0 0.0 0:00 bash 1298 luke 9 0 296 228 176 S 0.0 0.0 0:10 wmclock 852 root 9 0 328 140 136 S 0.0 0.0 0:00 xinetd 10616 luke 9 0 220 13292 S 0.0 0.0 0:03 asmixer 985 root 9 0 172 11676 S 0.0 0.0 0:00 crond 903 root 9 0 200 112 108 S 0.0 0.0 0:00 rpc.rquotad 533 root 9 0 168 108 104 S 0.0 0.0 0:00 mingetty 1142 root 8 0 136 8060 S 0.0 0.0 0:00 rhnsd 1 root 8 0 120 7252 S 0.0 0.0 0:06 init 1121 daemon 9 0 104 4432 S 0.0 0.0 0:00 atd 1087 root 9 0 456 3632 S 0.0 0.0 0:00 nmbd 1764 root 9 0 156 2824 S 0.0 0.0 0:00 rpc.mountd 1201 root 9 0 268 2016 S 0.0 0.0 0:00 login 1203 root 9 076 1612 S 0.0 0.0 0:00 mingetty 1204 root 9 076 1612 S 0.0 0.0 0:00 mingetty 1205 root 9 076 1612 S 0.0 0.0 0:00 mingetty [... the remaining 35 0.0%-memory using process omitted] I also notice that if I start 5 galeon browser windows, I get 6 galeon processes each using 34.4% memory (another 200% memory used). The man page for top says %MEM The task's share of the physical memory. None of these numbers add up, for me! Even if I assume the man page is wrong, and it should say %MEM The task's share of the physical memory, except when the process that's listed is sharing memory with others. Puzzled, luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs:
[SLUG] Linux music appliances?
This is probably a dumb question ... I'm looking for a small, quiet, Linux-based music appliance. The ideal would be a small fanless PC with CD drive and hard drive, little 2-line LCD screen and remote, with stereo audio outputs to load CDs in that then get converted into MP3s on an internal hard drive. So you have all your music in a handy system you can connect to your stereo system. The MP3beamer sounded interesting but no contact info, a dead and archived support forum at Linspire, make me think it's dead. Searches on http://linuxdevices.com/ seem to keep turning up dead/defunct business and old products. (Or things like the perfect PhatBox, which is designed as a car stereo and mainly onsold via Volvo, BMW, etc.) The only other things I've found are expensive high-end systems like the Sonos (above $1000 US). Surely someone is putting together a small noiseless PC loaded up with software to make a music appliance? Surely? I could build one myself given time, but I'd rather take the quick way out if there was one. luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Maxi-multi boot layout?
On 10 Oct, Jon Teh wrote: Waa waa waa. [Insert raised eyebrow] People would be more likely to be willing to lend assistance to you, if it wasn't for all the bitching and complaining about supposed problems in a tool you have used for the first time, and have made some errors in the configuration of. I wasn't asking for assistance, Jon. Sorry, perhaps I could have made that clearer. I have used grub off and on, and in fact about a year or two ago asked about it here, to see if my impressions were correct, or whether the problems I had with it were intrinsic. I got many fruitful replies from slug, but everyone who responded agreed that my key problem with it was real (that it doesn't check the configuration you give it, unlike lilo, which does). You may call that bashing grub, but I think it's a fair comment. I also chose to state that in a reasonably forthright way, thinking that if I were wrong, a staunch grub defender might enjoy pointing out my error. (And I would indeed be grateful to learn of a reliability check mode.) Everyone makes mistakes, especially if it's the first time they've used something. Looking at ways to resolve the problem is much more likely to reap positive results than bashing the tool. As I said, I wasn't asking for someone to fix my (self-induced, if grub-assisted) grub problem. Nor is it my first use of grub. But without knowing all the ins and outs of the cryptic grub command syntax, and hidden syntax (such as Tab-completion), it's generally easier to boot up off a rescue media and fix the problem outside grub. Lilo and Grub each have their own pluses and minuses, so while it may take a bit of getting used to Grub's syntax, it tends to be a bit better at actually making operating systems, especially non-standard ones, boot. Grub also has a nice console that you can try syntax on at boot time. Yes, but the learning curve is steep. There are useful hidden features, such as the null command to effectively do a kind of ls, but these facilities are buried away. I do agree it's more powerful and functional than lilo, though. To a power user, they'd be able to work wonders and recover from problems with insouciant ease. You are correct in saying that Grub does not require the BIOS to read drives, well, at least in the way I think you mean it. Thanks for that info, it means it's worthwhile persevering. Also, what is so hard with the Debian text installer? I'm not sure which screen it is you're having a biff with, but I have a feeling that it's not a very critical screen in any case. It presented me with a text console for configuring all the kernel modules as part of the installation process! To go through each one would have required a couple of hours. There was no suggestion about just accepting a default option - maybe I should have? But Ubuntu and Kubuntu are debian based, so I'm not sure there's much value in trying a vanilla Debian install. If they don't care about usability in the installation process, the attitude probably carries through to other areas. They still use dselect, for example, don't they? And the text install mode seemed a bit behind the times too. Traditionally, Unix has been weak in valuing attention to usability, so perhaps that just means they're more traditional Unix people. :-) luke -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Maxi-multi boot layout?
On 10 Oct, Grant Parnell wrote: Interesting approach. In short I think you'll find GRUB more flexible in the long run because you can fixup your mistakes at boot time thanks to the command shell. However I have a lilo solution for you too. I think you're right, once I become adept at grub. 1) Your boot partition needs to be first, it should be within the first 512 cylinders to be sure of working with old BIOSes. Both lilo and GRUB need to be loaded by the BIOS. Lilo uses a map of sectors to load the kernel and initrd files, GRUB however understands filesystems and can reach beyond the first 512 cylinders (once the stages are loaded) - ie it can load kernel and initrd from the outer limits or other partitions. 1st partition is a 200MB primary, currently pretty empty. 2) When installing distro's don't have them make /boot a separate partition. Have them make a boot floppy to get you up and running the first time as a precaution. That way they'll write their own version of /boot. Good tip, thanks. 3) Manually format populate the real /boot partition and mount as say /masterboot then put in copies of /etc/fstab and /etc/lilo.conf into /masterboot/etc/ and /boot/grub/ into /masterboot/grub/ Uh huh. 4) Use lilo -r /masterboot to update lilo with the new configuration. This may require copying bits of lilo and libraries it needs because of the chroot operation. Alternately if using GRUB just append bits to the /masterboot/grub/menu.lst (alias grub.conf) file. Okay, ldd will be my friend there I suppose. Or with grub, you mean I should build up a single menu.lst file that describes all the bootable images off all the partitions, if I'm understanding you. (Sounds fine.) 5) In the distro under test, symlink /etc/lilo.conf to /masterboot/lilo.conf if you're confident it won't stuff it up. Similarly link /boot/grub/menu.lst and /boot/grub/grub.conf to /masterboot/grub/menu.lst and /masterboot/grub/grub.conf Will that help? It's looking like lilo can't boot images off the big, 2nd drive. 6) I don't know if you can get away with symlinking /etc/fstab though, probably not because the system won't be able to find it when starting up Correct. That would be doomed to failure! :-) - however you could add to your /etc/rc.d/rc.local in each distro to copy /masterboot/etc/fstab to 'update' it, if you're confident that won't mess up the distro under test. Also possibly in /etc/rc6.d/ and /etc/rc0.d appropriate links to copy a changed /etc/fstab back to /masterboot/etc/fstab before the system unmounts /masterboot. Um, that sounds scary. I was planning to wear the pain of (manually) making the /etc/fstab in each distro, as similar as I could and just have to change one line. It's a little painful, but a safe option I think. As a sideline... here's how I install GRUB onto alternate boot media. mke2fs /dev/hdb1 mount /dev/hdb1 /masterboot mkdir /masterboot/grub cp -a /usr/share/grub/i386-redhat/* /masterboot/grub/ cp /boot/grub/grub.conf /masterboot/grub/ edit the new grub.conf file as desired check that menu.lst is symlinked to it sync grub grub find /grub/grub.conf (hd0,0) (hd1,0) {ours} grub root (hd1,1) Filesystem type is ext2fs, partition type 0x83 grub setup (hd1) Checking if /boot/grub/stage1 exists... no Checking if /grub/stage1 exists... yes Checking if /grub/stage2 exists... yes Checking if /grub/e2fs_stage1_5 exists... yes Running embed /grub/e2fs_stage1_5 (hd1)... 16 sectors are embedded. succeeded Running install /grub/stage1 (hd1) (hd1)1+16 p (hd1,0)/grub/stage2 /grub/grub.conf... succeeded Done. Thanks. That looks very similar to the recipe I followed (if I had copied in the grub files to go with it). Though looking just now (a bit silly, after midnight), I see that the grub directory has grub files dated from May 1st 2002 (from the mandrake 9.1 install, I think), that is probably causing my problem. Um, no. Ah well, it's late and I think I'll try again on Wednesday. Naturally, if you intend to boot this hard drive you'd need to a) make it the master IDE device or b) adjust the BIOS to boot off it (some can, some can't, some mess with your device numbers - ie swap hd0 with hd1 but /dev/hda and /dev/hdb are still correct). This procedure also works with other removable media such as USB keys, compact flash and external drives. Uh oh. The big 2nd drive is the slave drive, currently. Do you mean physically swap hd0 with hd1? So that what was /dev/hda becomes /dev/hdb, and vice versa? I thought it was more a function of lilo or grub as to whether they can boot from the drive or not: but where lilo needs the bios to read the kernel image, etc., grub is smart enough to know how to do that itself (as well as understand a few filesystems)? Have I got that wrong? Thanks, Grant, you've given me lots to think about, anyway. luke -- SLUG - Sydney Linux
[SLUG] Red Hat spell breakages
Just FYI. If you only care about American spell checking, you'll have no problems. Otherwise you may be interested to know that spell was replaced by a shell script that called ispell, around the time of RH 5 or 6. But RH's spell script was dead stupid. I think it accepted one option, instead of spell's 5 or 6. I wrote a replacement script that did right thing, calling ispell. I reported this bug to RH, including the fix, which they ignored. Later, ispell was replaced by a shell script that calls aspell. RH's ispell script is badly broken. Also, aspell appears not to support a dictionary option (if you can trust the silly ispell script). Nor does aspell come with a man page. I'm not sure if there's any point in reporting it again to RH. Just thought I'd let slug people know of the problem. FWIW, the fixed spell script *if* you still have a real ispell executable, is attached. When I fix the aspell script at home, I'll post that if anyone's interested. luke #!/bin/sh # # Make ispell work a bit more like spell. # # Author: Luke Kendall # for arg do case x$arg in x-b) # British spelling SPARGS=$SPARGS -d english ;; x-i) # Make deroff ignore .so and .nx commands. SPARGS=$SPARGS -n ;; x-l) # Follow the chains of all included files. How? ;; x-v) # Print all words not literally in the spelling list SPARGS=$SPARGS -m ;; x-x) # Print every plausible stem, one per line, with = preceding SPARGS=$SPARGS -m ;; x+*) # local spelling word file, 1 per line, sorted pdict=`expr x$arg : x\+\(.*\)` SPARGS=$SPARGS -p '$pdict' ;; *) FILES=$FILES '$arg' ;; esac done eval cat $FILES | ispell -l $SPARGS | sort -u
Re: [SLUG] Have I destroyed my system?
On 12 Feb, Grant Parnell wrote: You might have some short term luck with upgrading to RH7.2 but basically you've screwed it up. It's possibly recoverable but the effort to do so is going to be more than starting afresh. Yep. I spent 3 hours trying to repair it, and failing, yet it took only 30 mins to install RH 7.2 - plus about 4 hours last night reinstating 95% of the local mods. Some surprising omissions from the RH install: no Window Maker; no rxvt. I'd start backing up the data you need to somewhere. If you have /home on it's own partition, put all your stuff in there and do a fresh install with manual partitioning. You can then select which partitions to format and simply skip the /home partition. Yep, that's what I did. I'm *so* glad /usr/local was on a separate partition, too. I'd recommend that to everyone. I'd suggest writing down what's in /etc/fstab (or output of mount) , backing up all of /etc and stuff you want in /home. I took a copy of all /etc (since I hadn't cvs committed the current status before the destruction, since I'm still tweaking my cvs-etc scripts). Basically, though, it was pretty painless. Tonight: tweak terminfo/termcap, and recompile a more modern kernel, chucking in CD-RW support. I think I'll be fully back in action, then. Regards, luke -- SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/ More Info: http://lists.slug.org.au/listinfo/slug
[SLUG] Have I destroyed my system?
Re: [SLUG] How to install about 150 rpms? On 11 Feb, Tom Massey replied to: Well, I feel like taking a gamble and seeing whether this breaks my entire system or not, so I'll try it Um, that may not be a great idea. This assumes that the packages you have in your temporary directory fulfil all the requirements that all the other packages in that directory depend on. This may not be the case. Generally it's safer in this situation to do an rpm -Uvh *.rpm and try and fulfil the dependencies that rpm complains about before you start playing with --force or --nodeps. Searching on http://rpmfind.net can often find rpms that satisfy these dependencies. As a generality, with an rpm based system, if you find yourself using --nodeps or --force, you've probably done something wrong. Those options should really only be used, for example, by people who've installed the needed dependencies from source and so skipped adding them to the rpm database. There are reasons that rpm's have dependencies, there are reasons that the rpm program has --nodeps and --force flags. Installing potentially unstable software is not one of them. Well, actually it is. But only if you're willing to poke at the instability until it becomes at least as stable as a badly cooked blancmange. Looks like Tom was 100% right. I now have a bunch of programs that are running, but almost any other program will dump core. Including things like su or login. ldconfig reported that /usr/bin/libglib.so didn't exist, for example, so I think I'm screwed. Although I'm logged in, and even have a root shell, I suspect things are too broken to recover from easily. Will I have to re-install? I'm running RH 7.1; perhaps a way to recover would be to get a RH 7.2 CD and upgrade? I also have a Tom's root boot disc, and also one of those Linuxcare recovery CD-ettes (v 1.2). But I'm just hanging back from shutting down until I get some advice. Pity I hadn't hung back from doing the rpm forcing. Live and learn. Any advice? luke -- SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/ More Info: http://lists.slug.org.au/listinfo/slug
[SLUG] Re: just like the old days ...
On 14 Jan, Peter Allworth wrote: Based on their walk-through strategy, I can see a need for a Linux screensaver that looks just like the default NT screensaver or login screen! :) Ah, so Linux becomes like the French Resistance, hiding itself from the Microsoft scouts looking for businesses to target with their new anti- Linux warfare strategies. Brilliant! Actually, I remember a talk at SLUG, where afterward the presenter mentioned that he and another engineer were secretly using Linux on their machines instead of the mandated NT. They did a screen grab of the NT desktop and used it as the screensaver image, and as the background on a 2nd desktop they could flip too, in case a manager walked by. Then all they had to do was explain how they could keep working when all the Windows machines on the network went down. luke John Rosauer wrote: http://www.theregister.co.uk/content/4/23518.html ... when IBM, DEC, etc were doing this against Unix -- SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/ More Info: http://lists.slug.org.au/listinfo/slug
Re: [SLUG] Question for a friend
On 2 Nov, Jon Biddell wrote: I've had this problem when trying to dual-boot a WinME system (word of warning - DON'T DO IT ! It involves swearing, blood and LOTSA agro !!). In fact, the Linux partitions that Partition Magic (at least up to 5.0) create are non-comparable with Linux's fdisk / fips as well. Solution ? Apart from the obvious, I have not looked into it much further. He only installed Linux because he's been having enormous trouble with the 40Gb HDD under Windows ME - so much so that he thought the drive was faulty. But after an exhaustive scan of the disc, it reported no errors. So he installed Linux over the top of ME and it was rock solid. So it suggests that all the hardware is fine. Then, when he went to put Windows back, he couldn't! The Linux install prevented the Windows install from working. I know the solution, though: run Linux. :-) (Actually, I hear that Windows can cope if you use Linux's fdisk to delete the partitions first.) luke -- SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/ More Info: http://lists.slug.org.au/listinfo/slug