Re: bash quoting problems
On 08/08/2011 04:43 PM, Andreas Berglund wrote: Hi! I have a problem with the following sed snippet sed -i s|^\( *PATH=\)\(.*\)|\1$ADD:\2| ~/profile-test I need soft quotes in order for $ADD to expand and I also need to math against one doublequote in the regexp in for $ADD to be put in the corrct place. Does anyone know how to do this? You may want to consider putting the sed script in a file and using the -f script (or --file=script) option instead. No quoting needed. ;) -- Bob McGowan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4e41b7f4.7050...@symantec.com
Re: Wicd Wireless Woes
On 05/24/2011 02:09 PM, Wayne Topa wrote: On 05/24/2011 05:04 PM, Wayne Topa wrote: On 05/24/2011 11:33 AM, Bill wrote: Hi folks, I'm having problems connecting to my wireless access point using wicd. Wicd seems to be running ok. I can connect and disconnect from my wired network at will. Wicd also detects a variety of wireless networks in my neighborhood, including my own, but won't connect. It claims there's a bad password, but I've triple checked that. It's correct. I'm including the relevant information from the wicd log below. I hope someone can point me in the right direction. Wicd is complaining about a variety of things, any of which could be the real cause of the problem. After a re-read of your OP, I have more questions, mainly how are you starting wicd? I ask that because I have reviewed my squeeze logs where I did have problems getting wide to associate, and find no reference to wicd-cli in them. I start wicd on a console with wicd-curses and in X with wicd-gtk. My logs show no wicd-cli calls at all. Hope this helps, some WT My old eyes are seeing things. Forget the wicd-cli I read it wrong!!! wpa-cli != wicd-cli. Sorry for the noise And for replying to you directly. Iceweasel just dropped the reply-to-list button and I didn't notice it until after I hit send. G I've missed earlier emails on this subject, so this may have been said and found wanting. In Ubuntu land, there have been many references to this type of problem. The bottom line appears to be that 'wicd' and 'network-manager' get installed together (that is, when you install 'wicd', 'network-manager' is not removed), and they interfere with each other. Remove *all* packages related to 'network-manager' that are installed. This has worked for me on several different Ubuntu/Kubuntu releases. I don't know what the Ubuntu references are, that info's all on my home system. -- Bob McGowan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4dde8f15.4010...@symantec.com
Re: permissions all zero when using 'cp'
On 12/29/2010 05:56 PM, Martin Lorenz wrote: Dear Gurus, i recently noticed some errors at my mail-server and so I tried to drill it down with my limited abilities. what I found is really strange: -deleted- r...@x:/tmp# ls -altr insgesamt 20 drwxrwxrwt 2 root root 4096 29. Dez 15:06 .X11-unix drwxrwxrwt 2 root root 4096 29. Dez 15:06 .ICE-unix drwx-- 2 mlo users 4096 29. Dez 21:38 ssh-VkxmJ15962 - - -rw-rw-r-- 1 root root 0 29. Dez 21:47 test drwxr-xr-x 21 root root 4096 29. Dez 22:06 .. drwxrwxrwt 5 root root 4096 29. Dez 22:22 . notice the file test r...@x:/tmp# cp test test.bak r...@x:/tmp# cp -p test test.bak2 r...@x:/tmp# ls -altr insgesamt 36 drwxrwxrwt 2 root root 4096 29. Dez 15:06 .X11-unix drwxrwxrwt 2 root root 4096 29. Dez 15:06 .ICE-unix drwx-- 2 mlo users 4096 29. Dez 21:38 ssh-VkxmJ15962 - - -rw-rw-r-- 1 root root 0 29. Dez 21:47 test.bak2 - - -rw-rw-r-- 1 root root 0 29. Dez 21:47 test drwxr-xr-x 21 root root 4096 29. Dez 22:06 .. - - -- 1 root root 0 29. Dez 22:22 test.bak drwxrwxrwt 5 root root 4096 29. Dez 22:22 . now kindly notice test.bak and test.bak2 -deleted- I'm not sure if this is relevant, but the ls output lines for these files look odd, having extra '- ' characters at the start. Older 'ls' commands (not Gnu) might do this, if the file name contained a literal carriage return character, but the 'ls' on my system prints a question mark for non-printing/graphic characters. You could try the '-b' option and see what that prints. The Gnu ls uses backslash escapes (\r, \b, \octnum etc.). This might help in searching strace output, if you do try Bob Proulx's suggestion, since you would know the actual character to look for. Also, you ran the above as root. Since the original file is readable by all, what happens if you do the copy as a regular user? -- Bob McGowan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4d1cbebf.1050...@symantec.com
Re: Web calendar sharing in Evolution
On 12/30/2010 04:34 AM, Kousik Maiti wrote: Hi list, I want to share calendar for local LAN.I try to find documentation regarding this,but don't get any good solution.Some solution is there but that is for google calendar.As I don't connect to internet that is not applicable for me. Anybody help? I can't help with what you want, but maybe you can help me ;) I've been looking for the google calendar solution documentation but have been quite unsuccessful at finding anything. Can you provide the references you found? Thanks in advanced. -- Wishing you the very best of everything, always!!! Kousik Maiti(কৌশিক মাইতি) Registered Linux User #474025 Registered Ubuntu User # 28654 Many thanks, and Happy New Year. -- Bob McGowan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4d1ce98e.40...@symantec.com
Re: JACK configuration
On 01/16/2011 11:15 AM, Robert Blair Mason Jr. wrote: Hi all, I am attempting to record audio on my laptop. To this end, i've installed Ardour/JACK. I have *finally* managed to get ALSA to work. However, I now have a problem - the sound that is wired into the Mic In jack on the front of the laptop will play back (so basically my laptop is functioning as a speaker), but I can't seem to figure out how to get JACK to connect. Can anyone help me configure JACK so that the input will go to Ardour? The connection graph shows that system:capture1 and system:capture2 are wired to go to Ardour:Audio In 1 and Ardour:Audio In 2, respectively (so it seems like this should work). However, when the Ardour is prepped for record, the volume bars that indicate input volume are relatively static, while I can hear sound being played back. After recording, the track is silent. Portions of the output of amixer that look relevant: --deleted-- I'm not sure the ALSA settings are relevant, because ... Any help would be much appreciated. Please note, this is by memory and it's been a while, I would need a little time to be able to verify my sound system setup. With that caveat: First, to confirm that you have in fact set up to record: 1. IIRC, Ardour by default only has things set up for a pass through type of operation, which matches what you've described. Again, IIRC, you will need to create a recording track and connect it between the master out and master in channels. I will verify exactly what I have set up as soon as I can and let you know the various settings. 2. If you believe things should be working because you've made connections, it may be you haven't gotten things correctly armed. In the main window, in the controls section for the track you intend to record on, there is a small round icon, which must be selected to arm the track. In the top section, there is also a round icon in a button (a bit larger than the one mentioned above ;), to the right of the playback button. You must select it next. Then, select the playback button to start processing. If everything is working, you should now see a light colored ribbon being created across the screen for the recording track only, and there should be two lines (assuming stereo input), inside the ribbon. If volumes are high enough, the lines should show amplitude variations. -- Bob McGowan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4d34bc11.8060...@symantec.com
dual boot hardware problem
I have a system I wish to set up for dual boot with WinNT and Debian. I have gotten the mini How-To for setting up the NT boot loader to access the Debian system, and have installed on the second partition of what I thought was my boot disk. But Debian thinks its on the second disk. I have a Micron PCI system with the IDE disabled, an Adpatec aha-2740 PCI (BIOS enabled) and a BusLogic FlashPoint (BIOS also enabled). I am using the BIOS to map the SCSI ID 0 drive on the adaptec to drive C for booting and the BIOS on the flashpoint to both map the second SCSI disk as drive D and to enable SCSI CD-ROM booting. Windows NT was installed first and it and the system see the Adaptec as the first controller and its drive as the boot drive. But Debian sees the BusLogic card first and the Adaptec second and so reverses the boot order. I have searched under the general topic of dual boot in the mail list archives but have not seen anything referencing this type of problem. I had tried, as an experiment, to set the Debian/Linux partition active, but on boot lilo only managed to print LI and then hung. I am planning on going ahead with the NT loader setup but am concerned that it might not work, based on this boot hang. Any thoughts or theories that might help would be appreciated. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com
RE: Formatting a file with mkfs.msdos
-Original Message- From: Steve Lamb [mailto:[EMAIL PROTECTED] Sent: Sunday, August 16, 1998 1:05 PM To: Debian-user Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Formatting a file with mkfs.msdos Does anyone have a clue as how to format a file with mkfs.msdos? Steve, I'm not sure what you are doing here, so more detail would be useful. But, let me make a few comments, which may help answer the question anyway. Anything with mkfs as part of its name generally refers to a tool used to build a filesystem. In the DOS world this is what format does. Files, on the other hand, live in the file system. So using an mkfs on a file does not quite make sense (unless you are using the word file in the VERY general sense of an item in the file system, which, in UNIX style systems, includes the names refering to disk devices). If this is the case, then you need to be looking at files found in /dev like sda1 or fd0. You will need to check your /dev directory and documentation for the correct names for hard disks on your system. Usually, fd0 will work for any floppy (autodetects the size). Also, you will want to look at the documentation for mkfs which probably calls mkfs.msdos for you. The command line may be simpler. If this is what you need and this is not enough information, let me know. I do not have the doc at hand so would have to check it out this evening. Another possibility is that you are in fact working on files (text or data). There is the concept of a DOS format text file since DOS systems use both a carriage return and a line feed at the end of lines where UNIX based systems only use a line feed. DOS files also use a ^Z (control-Z) character in the file to mark the end of the file. There are two commands to change text files back and forth, dos2unix and unix2dos (I believe these are part of the Linux OS environments in general, so should be in Debian Linux). This should NOT be done to binary (data) files since it will destroy the structure expectd by whatever probram made the file. If you do not have the two utilities I can supply you with some alternatives (assuming you have sed and tr available). --- Bob McGowan i'm: bob dot mcgowan at artecon dot com
RE: dual boot hardware problem
Many thanks to both Anthony Richardson and Nathan Norman for pointers to the right place to resolve my boot problem. An additional question would be what else do I need to worry about? Clearly, I will need to redo LILO, since the disk order will have changed. Devices listed in fstab may also need fixing. Is there anything else that needs changing? Thanks again, --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -Original Message- From: Nathan E Norman [mailto:[EMAIL PROTECTED] Sent: Monday, August 17, 1998 8:55 AM To: Bob McGowan Cc: Debian Users (E-mail) Subject: Re: dual boot hardware problem On Fri, 14 Aug 1998, Bob McGowan wrote: [ snip ] : But Debian sees the BusLogic card first and the Adaptec second and so : reverses the boot order. I have searched under the general topic of : dual boot in the mail list archives but have not seen anything : referencing this type of problem. Ah yes, this is so fun. I had the same problem with an Adaptec 2940 that Linux saw as the first adaptor, but BIOS saw as the second adaptor and booted off the NCR 53c810 instead. The fix isn't trivial, but it's not too bad. You need kernel source, of course. Once you've got your kernel source ready to go, cd to the ./drivers/scsi directory, and edit the file hosts.c. You need to edit the builtin_scsi_hosts array ... the beginning of the definition looks something like this (I've got 2.0.35 kernel source) static Scsi_Host_Template builtin_scsi_hosts[] = { #ifdef CONFIG_AMIGA #ifdef CONFIG_A3000_SCSI A3000_SCSI, ... You need to find the #ifdef statement for the Buslogic card and move it AFTER the #ifdef statement for the appropriate Adaptec card. I told you it was fun :) There may be an easier way to do this with boot parameters or something, but I never figured it out. If someone knows, please enlighten us! -- Nathan Norman MidcoNet - 410 South Phillips Avenue - Sioux Falls, SD mailto:[EMAIL PROTECTED] http://www.midco.net finger [EMAIL PROTECTED] for PGP Key: (0xA33B86E9)
RE: win98/hamm dual boot problem
On Wed 19 Aug 1998, Michael Stenner [EMAIL PROTECTED] wrote: Following instructions I found on this list some time ago, I installed hamm and win98 separately: each have their own drive - only one was hooked up at a time. Than, I put both drives in: Linux as hda and win as hdb. THis was nothing new for linux, so I started up linux and ran lilo with the following lilo.conf: If I read this correctly, you have done the following: 1. Installed Linux on a drive that is seen by the BIOS as drive C. 2. Removed this drive and replaced it with another. 3. This new drive, also seen by the BIOS as drive C, has Win98 installed. 4. You set this second drive to be the second (changing its address) and replaced the Linux drive so it is the first. 5. This means the Linux drive is seen as C and the Win98 as D. If this is true, then it may be that your Win98 is failing because the addressing is wrong for where it is installed. It expects to be on the first drive but isn't. You could try re-installing the Win98 to the second drive (maybe you want to nuke the existing Win98 partition so it doesn't think you are trying to do an upgrade) and let it handle the MBR on the first disk (note I am not too experienced with the Win95/98 way of doing things - this would work for WinNT no problem). Then redo the lilo.conf (you should have a Linux boot disk before this, just to be sure you can get back to your Linux installation). There is a How-To for setting up a dual boot for WinNT/Linux (using the NT boot loader) which may be helpful in your situation. I do not have the URL for it but I found it in a search of the mailing list archives (look for 'dual boot'). The NT details may or may not be useful but the over all discussion is very good. Hope this helps ;-) --- Bob McGowan i'm: bob dot mcgowan at artecon dot com
RE: freeing space on /usr?
Mike, I have used this technique on a few Solaris systems when the original partitioning proved to be inadequately sized and there were no other simple options available. The only issue I can think of that might be a problem is symbolic links within the moved directory structure which could end up not pointing to the right thing. The type of problem I am thinking of can be seen if you use a standard Bourne shell (NOT ksh or bash, for sure). A symlink to a directory can be cd'd to, but a pwd command will show the actual path to the current directory, not the symlink path, which can be confusing to users. It could also be confusing to programs accessing a symlink of the relative type ( -- ../something ) since the thing pointed to could be in the original location (I don't know how clear this is...;-) $ ln -s /tmp /usr/tmp/somethingnew $ cd /usr/tmp/somethingnew $ pwd /tmp #ksh and bash would show /usr/tmp/somethingnew I have not tried this sort of thing myself on Linux nor have I tested the relative path issue in any environment. So the bottom line is that some symlink things may break but in general the procedure will work. As to appropriate things, nothing other than trying to avoid directory structures with lots of symlinks comes to mind. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -Original Message- From: Mike Miller [mailto:[EMAIL PROTECTED] Sent: Friday, August 21, 1998 2:22 PM To: debian-user@lists.debian.org Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: freeing space on /usr? I'd like to make a little more room on my /usr partition. Is it safe to move /usr/doc to somewhere else and make a symbolic link back? Is there something more appropriate that can be moved to make space? Mike -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Install problem - kernel problems with Adaptec 2740 EISA cards
I downloaded the Official CD images 2 or 3 months ago and did all the checks and balances before burning a CD. I also checked the CD image file against the burned image (dd and cmp under a UNIX system) and found no errors. I tried to install from the CD using a DOS boot and running the install.bat file to start Linux (the one in the /boot directory). Initially, the kernel would find the 2740 cards, download sequencer code, reset the bus (3 times, once for each card), do a few other things (qlogicisp probe and eata-dma probe) and panic with the message Encounterd spurious interrupt. I checked the EISA config for the cards and found they were set to level trigger on the interrupts, changed this to edge trigger. The kernel then reports 3 spurious interrupts (1 per card, I presume), then aborts some scsi command due to timeout, resets the scsi bus, then enter an endless loop timeing out and resetting. The abort message is aborting command due to timeout: pid0, scsi0, channel0, id0, lun0 Test Unit Ready 00 00 00 00 00. I have also created the boot floppy (rescue disk) from the CD and tried to boot from it with the same results. I have just downloaded from the ftp site the rescue disk image, on the off chance that there have been changes (but this looks unlikely - the CD image is 1.3.1 and the directory and time stamps look like mine). I have also checked the FAQ and scanned the currently open bugs lists to no avail. Any thoughts/suggestions would be most appreciated. --- Bob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: Problem with my shell
-Original Message- From: Jeff Shilt [mailto:[EMAIL PROTECTED] Sent: Thursday, May 07, 1998 2:44 PM To: [EMAIL PROTECTED] Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Re: Problem with my shell On Mon, May 04, 1998 at 02:08:36AM -0400, Jeff Shilt wrote: I wrote a postinst shell script for a package and when I went to test it I get bash: ././postinst: no such file or directory I changed my PATH environent variable around a bunch, even took out . ---deleted--- I did remove the remove the spaces, that wasn't it. However, I did solve it by using mcedit from mc. Apparently it had something to do with how ae was saving it. I've stopped using ae! Thanks, Jef Interesting. I am also using an environment (Cygwin developed) to run unix utils in an M$Win environment. One of its features is distinguishing between (in bash) text and binary mode (a DOSism). When I set things up to run binary, scripts I had written earlier would not work, and with the same error message, come to think of it. The reason is/was the presence of a carriage return character, which in text mode was being stripped but in binary was becoming an invisible part of the string, which of course could not be found. Could be ae was writting in DOS Text Mode?? --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: how to zmodem via telnet?
-Original Message- From: shaul [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 13, 1998 12:09 PM To: Vesa Kaihlavirta Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Re: how to zmodem via telnet? ++ ++ My ISP has lynx, that contains a `use zmodem to download to the local ++ terminal'-option. ++ ++ I'd like to telnet to my ISP, download a file via lynx, and zmodem it to my ++ local terminal. Problem is, when I choose that `use zmodem...', I only get ++ the zmodem handshaking line (with *B and lots of zeros). ++ ++ So, how do I launch rz? ++ ++ Since I always get confused about when to use rz and when to use sz, I ususaly try both of them. ++ Maybee you should try sz file_name also ? No need to use sz, the original poster wants to receive the file from the remote (the system s/he is telnet'd to) to their local system, so 'rz' is correct, to 'receive' the file. But I am a bit confused, here. Generally, using telnet implies a TCP/IP network connection, so an ftp to the same remote system to retrieve the file would also work and would probably be easier. Generally, to run the receive zmodem program would require you to 'suspend' the telnet (usually, telnet uses a control-] to do this) and then run the rz command from a subshell, pointed at the same 'device' the telnet is using. This can work OK if instead of telnet you were using a standard serial connection with something like 'cu', where you could reach the serial port as /dev/tty??, but I don't know how you would make it work over a TCP/IP connection. Many modern terminal emulation programs can also be configured to detect the 'send Zmodem' initialization string (the *B... stuff) and automatically invoke the receive zmodem procedure. For telnent to do this means a change in the code to support the Zmodem protocol (which has been done, in at least one case, for a telnet/rlogin program for M$ Windows systems - I don't remember the actual package offhand, if you need it let me know and I will search my archives for it). I hope you find this info useful, though perhaps not exactly helpful, in this case;-) --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: V.90 modem recomendation
I too have some questions re. Linux and modems. Presumably, the concern about winmodems has to do with modem cards and not external modems? Though it is obvious an external modem should use the latest UART (16550?)? And if I already have an internal modem, will it NOT work with Linux if it is a winmodem or are there workarounds? I know some of this is probably documented and I just haven't found it yet, so pointers to reading material are also appreciated. TIA, --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -Original Message- From: Hamish Moffatt [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 13, 1998 4:23 PM To: K. Claussen; Ian Keith Setford Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Re: V.90 modem recomendation On Wed, May 13, 1998 at 01:05:08PM -0400, K. Claussen wrote: Personally, depending on how much money you want to drop on it, I would highly recommend the USRobotics (or is it 3Com now? Not sure) Courier On the other side of the fence, I have been modeming for over eight years and have always found the cheaper Rockwell option to work just fine. I would recommend saving your money! Here in Australia a V.34 modem is around $100, USR Courier is still mid $200s. Hamish -- Hamish Moffatt, [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] Latest Debian packages at ftp://ftp.rising.com.au/pub/hamish. PGP#EFA6B9D5 CCs of replies from mailing lists are welcome. http://hamish.home.ml.org -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: Installing Debian from WindowsNT Pt. 2
John, this is a generic answer to your question on how to transfer files from your NT box to a Linux (or other UNIX type) system. I am not familiar with everything that is part of the Linux Base, so I am going to assume that there are no utilities there for serial file transfer. If they are, you can ignore the first step. 1. You will need to get a file tranfer program of some sort onto the Linux box. This could be rzsz (the Zmodem receive/send tools), xmodem or ymodem. It does need to be a binary executable, so you will need to get it, copy to a floppy and then copy from the floppy to the Linux system. I will also assume that you know how to do DOS floppy access from Linux and that the tools are there. 2. Once you have the above installed, login over the Hyperterm serial line. It is a good idea to set this to run at the highest bps rate the hardware can support, so you get faster transfer times. On the Linux side, this would be done (if I remember correctly) in the /etc/inittab for the getty running on the port. In Hyperterm, pull down the file menu and select Properties, then click on the configure button to set up the bps rate to match your getty setup. 3. On the Linux side (at the shell prompt and assuming you got the rzsz pgm), type rz to receive a file. In Hyperterm, pull down the transfer menu and select Send File... which will bring up a dialog box where you can select the name of the file to send. The default protocol is zmodem, change it to whatever you are using on the Linux side if necessary. Select OK and the file should now go over to the directory you logged into (or cd'd to after login). Again, this is a fairly high level overview (with details where I remember and can check on my local system - for the Hyperterm part at least). My Linux box is at home so I cannot check details for you on that end. Good luck. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Thursday, May 14, 1998 12:40 PM To: debian-user@lists.debian.org Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Installing Debian from WindowsNT Pt. 2 After searching and searching the FAQ's and HOW-TO's I found that I could enable the COM port on the Linux box in the /etc/inittab file. I can now log onto my Linux box from hyper terminal on the Windows machine but I still can't access or send files from the Windows machine to the Linux box. I have been through and through all the info I can find on the WEB with no hope. Everything assumes the I have access to the net, but I don't. I also don't have a CD Drive. I have a Pentium motherboard with 16 M RAM, a 600M hard drive with Debian Linux installed and a 400M hard drive I installed as a second drive AFTER Linux was installed. I've never used Linux or UNIX before so I need step by step assistance. I realise this is going the hard way but my System Administrator won't allow a Linux box to be connect to the network. He is afraid that Linux will bring down his precious WindowsNT network?!?!?!? Is it possible to mount the second hard drive I added after installing Linux or will I need to re-install Linux. ( It's no big deal at this point, I have nothing but the Base Floppies installed right now) I've already got several people bugging me for access to a 'true Operating System' but I keep telling them I have to get the system installed first. Thanks again for any assistance you can give. Cheers, John Gay -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: Linux the hard way
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Thursday, May 14, 1998 4:40 PM To: [EMAIL PROTECTED] Cc: debian-user@lists.debian.org Subject: Linux the hard way Bob, I downloaded minicom and lrzrz via floppy and used dselect to install them. I then logged in from hyperterminal and sent a file using zmodem protocol after issuing the rz command from the Linux box. SUCCESS!!! The Congratulations... highest baud rate my terminal will allow is 19200. As you cam imagine, this will be rather slow :- I think the next step is to download Just remember what it would have been like in the days of 300 and 1200 bps modems, and be thankful you can do 19.2K bps. the packages from the debial site to my windows machine. Is there a free utility I could use to copy the entire directory/ies rather than right-clicking the files one by one? I want to re-create the directory structure on my windows machine so I can then transfer this structure to my Linux box so I can hopefully use dselect to install the rest of my Debian Linux system semi-automaticly. Does this sound like the right steps, or am I missing something here? If you have a better suggestion, please let me know. Thanks again for your help, I feel like I'm on the way to finally learning to use a Real operating system. Cheers, John Gay My only other thought would be to get a dos/Win version of tar, download the Debian files to your NT machine in the desired locations and tar the whole thing up into one file based on the parent directory location. Then transfer this single (very large) file and hope there are no power failures (though zmodem will check for partial transfers on a restart and seek to the position of the last correctly transferred data - thank goodness ;-). On your Linux system, use tar to extract the files. There may be other utilities that would do recursive transfers, but I do not of them. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: first script executed
-Original Message- From: nico [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 20, 1998 2:02 AM To: Hamish Moffatt Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Re: first script executed Hamish Moffatt wrote: On Wed, May 20, 1998 at 09:45:20AM +0200, nico wrote: Do you know what is the name of the first script executed after the install of the distrib ( the one which call adduser to give a passwd to root and create a new user, dselect ...) and how i can configure it ? I forget what it's called, but I'm pretty sure that it is deleted after it is used. Why would you want to run it again and what do you mean `how can i configure it'? I'm trying to realize my own debian installation. That's why i would like to customize it. In fact, i don't want tu re-run it but to change it before it is executed... If any1 have a clue. Hi, Nico, I do not know the answer, but do have a couple of ideas. You could mount either the CD or the base floppies and try to track the order of execution of scripts. I would check out the root for a .profile first. Xenix installs used to come up as a root login and start out with this file. Also check for rc files in /etc, especially ones with odd names. The second thing I have had good success with is changing to an alternate virtual terminal during the install and running ps. This not only gives me a list of the processess currently runnning but also gives the expanded arguments being used. Of course, this means doing a normal install again so you can track things. I have found that using these techniques can also be a powerful (and sometimes surprising) way to learn the system. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: NFS mount problem
-Original Message- From: Nick Gilliam [mailto:[EMAIL PROTECTED] Sent: Friday, May 22, 1998 4:09 PM To: debian-user@lists.debian.org Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: NFS mount problem When attempting to mount nfs I get the following error: mount qms1:/home /mnt mount clntudp_create: RPC: Program not registered ---snip--- This error indicates that the RPC software on the remote host is not running (required for NFS servers) which probably means that the server is not set up to be an NFS server yet. Check the rc scripts (look in /etc/init.d) for one related to NFS and see if you can manually start the daemons. I am afraid I do not know the Debian rc setup that well yet, so I cannot be more specific. On a Solaris system, there are files called nfs.client and nfs.server. The first is started by default, the second has to be set to start by telling the init program to go to run state 3, rather than run state 2, as its initdefault in /etc/inittab. I expect something similar on the Debian Linux side. I hope this helps you get NFS running. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: NT and Linux
-Original Message- From: King Lee [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 27, 1998 12:29 PM To: debian-user@lists.debian.org Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: NT and Linux Hello, I got into a discussion with a system administrator of a website. The system administrator wishes to use NT because it supports software raid 5 (raid without a special controller). I thought if it works, there would be a terrible performance degradation. The system administrator said only if a disk goes down would there be a performance hit. There will be some performance loss, since the system CPU will need to handle the RAID algorithm. Software RAID also means buying the SW to support it or having it come with the system (as it does with NT). It is still necessary to purchase the disk farm and perhaps more HBA's to distribute the load and improve redundancy. There are no restrictions that I know of about the interface or disk types used. One issue is how much I/O is written at a time and the size of the stripes of data written to each disk. As an example a 5 disk array using a stripe size of 16K will have a big performance hit (whether implemented in SW or HW) if a write of less than 4*16K (64K) is made. The RAID system must then read the unchanged data from the disks, make the needed changes, calculate the new parity and write the whole thing back. This takes CPU cycles and will affect system performance at some load levels. And even if the writes are full stripes, it still needs to calculate the parity and write the stripes to disk. Another issue is the type of I/O being done (random vs. sequential) but this impacts I/O performance in either SW or HW RAID. The more random the I/O, the beter the chances are that several writes (or reads) will land on different disks in the array, reducing seek time issues. I/O performance will also improve as more threads are run. Does anyone here know anything about The questions I have are 1. Has anyone here had any experience or knowledge about software raid. How good is it? I have used it, but not recently and not in a production environment. It does/did work. (I'm a test engineer so I beat the hell out of it. I had no failures or problems.) 2. Does Linux support hardware raid 5 Basically, any system can support hardware RAID at any level, since the RAID functions are handled by the RAID controller. But then there needs to be some way to configure the RAID subsystem. This can be done by either a serial interface to the RAID subsystem controller, using a terminal emulator, or by special software using a SCSI pass through to send information to the controller over the SCSI bus. This assumes a SCSI subsystem of course. The subsystem manufacturers are building high performance systems, so the dollar outlay can be large (4 or 9 GB 7200 RPM Ultra SCSI disks in a cabinet running Ultra SCSI to the host, supporting a large number of drives (7 or more)). The serial method is fast and easy but does not scale well to large numbers of systems, where the SCSI base scales nicely but is more difficult to implement well. I think this guy is looking for an excuse not to use Linux. King Lee -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: Nethack save files: was: angband save problems
I have not been following the beginning of this discussion thread, so this may have been said before. The reference to nethack caught my eye, since I had tried it out and had save problems, too. What I found is that both the save file and its parent directory must have write permission in order for this to work. If they are both set to group id games and then the executable is made set GID with group games, things worked fine. This was a week or two ago and I do not remember if the group was acutally named games, but the procedure is what counts, here. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -Original Message- From: Britton Leo Kerin [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 27, 1998 6:28 PM To: debian-user@lists.debian.org Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Nethack save files: was: angband save problems Nethack seems to have the same sort of problem. Is this perhaps a systematic error in debian's arrangement for games and other things that have to save data of this sort, or is there something we're missing? I seem to remember a big argument about where to keep certain kinds of 'variable-config-save-sorta-thingies' somewhat recently, could that be related to thses problems? Yes, that is how you do it. However, as to whether or not that would help, or is advisable, I'm not sure. Your executable is set GUID. Hrm. But the files are not set to group writable. Maybe that is the problem. Permissions are the same as my save file (name 1000.mcv21 and mcv21 instead of apc27) They're all in angband/lib/save, right? Nope /var/lib/games/angband/save/ Interestingly, character dumps fail as well Matthew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: Getting ttyx
Several ways, including piping the tty output to sed or cut or awk. But my preference would be: basename $(tty) # assumes you are using bash, ksh etc. or basename `tty` # back quotes for sh --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 27, 1998 11:50 PM To: Debian User List Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Getting ttyx What is the best way to get the current console's tty? I know tty does the job, but how do I find the name WITHOUT the /dev/ ? i.e. tty1, tty2, tty3 etc? I want to use it with bl in a script. Michael Beattie ([EMAIL PROTECTED]) -- - Curiousity may kill the cat, but a 12 gauge is quicker! -- - Debian GNU/Linux Ooohh You are missing out! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: adding users via scripts
-Original Message- From: Nathan E Norman [mailto:[EMAIL PROTECTED] Sent: Sunday, May 31, 1998 10:11 AM To: Debian User List Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: adding users via scripts We are in the process of moving users from a BSDi box to a Linux box. BSDi has an adduser script which accepts encrypted passwords. The Debian adduser script does not. From reading through the code for each, it seems that BSDi's adduser script manipulates the password file directly. ( I'm not a Perl guru, btw :) So, I'm wondering if anyone knows of or has developed an adduser type script for Linux that will accept encrypted passwords on the comamnd line. I don't feel up todeveloping this myself, though I'll give it a try if I have to. In the short term, I'd probably add the users with one script and change the entries in /etc/shadow with another. Searching for ideas, advice, etc. Thanks. How about just adding the 'passwd' command to the scirpt? --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: NT and Linux
-Original Message- From: King Lee [mailto:[EMAIL PROTECTED] Sent: Thursday, May 28, 1998 11:29 PM To: Leandro Guimaraens Faria Corcete Dutra Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Re: NT and Linux On Thu, 28 May 1998, Leandro Guimaraens Faria Corcete Dutra wrote: King Lee wrote: 1. Has anyone here had any experience or knowledge about software raid. How good is it? 2. Does Linux support hardware raid 5 snipped The article from www.osnews.com did say that software raid takes up CPU cycles, but it did not say how much. It would seem that if the CPU must check for errors on each byte from disk and performance would take a big hit. Perhaps the kernel checks for errors only if it knows that a disk died, and normally there would not be a hit. Does anyone know about CPU hit of software raid. Why would anyone buy expensive raid hardware if software does the same without too much penalty? King Lee First, the CPU not only checks for errors on reading, it must also calculate the parity on writes. In RAID5, spanning 4 disks, for example, 1/4 of the storage is used to hold parity info. Data is written in stripes of some size, one stripe per disk, in a round robin sequence. One stripe will be parity. In the above 4 disk example, if a stripe were 16K in size, there would be 48K of data and 16K of parity. In RAID5, the parity stipe will rotate between disks, so no single disk is loaded with all the parity (this improves performance over RAID4(I believe) where all parity is on one disk). If a disk write is less than 48K, the system must read 48K from the disks, make the needed changes, recalculate parity and write the resulting 64K back to the disks. If the size is 48K, this read of data can be dispensed with. The system must then only calcualte the parity and then write the 64K. This means CPU cycles are needed for SW RAID. I do not know the impact in terms of actual numbers, but I can say the main issue is scalability. In SW RAID, the more RAID subsystems created, the greater the impact on CPU performance. In HW RAID, there is no additional impact. So even if SW RAID for a single RAID5 subsystem matched HW RAID for the same config, there will certainly come a breakeven point, where additional capacity causes CPU performance degradation in the SW RAID setup. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: NT and Linux
Hi, King, my comments follow your questions, below. I hope this helps. Bob King Lee asks: Thanks Bob McGowan for your very informative reply. I gather that 1. Software raid is OK if problem is I-O bound, i.e., CPU would normally be idle waiting for I-O. I would agree with this analysis. If the CPU is doing nothing, it might as well be calculating parity for RAID. :-) 2. If we have multiple subsystems, we increase the the I-O bandwidth, and now the CPU may not be keep up with the I-O. In general, increasing I-O turns I-O bound problem into CPU bound program. I would also expect this to be true, though I have no evidence to support the idea. 3. Software raid 5 may be OK for workload with lots of reads, but run into trouble if workload does lots of writes. Not necessarily. Remember, when reading the data, you still have to read a stripe from all the disks and verify the parity, so there is still some overhead. Also, if there are lots of writes, there may be a higher chance of ordering the I/O requests to take advantage of writing a full set of stripes, reducing the frequency of the read/modify/write cycle, which will reduce I/O load. 4. Software raid 5 is more efficient for large files. Generally, the answer to this is: it depends ;-) Are you talking reads and/or writes. What combination? How random? Etc. Also, this question (and the third, to some extent) are getting away from the original question comparing SW and HW based RAID technnology and are getting into the more specific issues of RAID efficiencies, which DO NOT depend on whether the RAID is SW or HW. Generally, in RAID5, writes will always be more expensive than a regular disk. If you have a read/modify/parity calcualtion/write scenario, it is worse, but even the data collection/parity calculation/write sequence takes more time than a pure write. The efficiency of RAID5 is in its read characteristics, for random access. Large numbers of random read requests will distribute across multiple spindles, improving I/O due to redcution of seek delays and an overall reduction of read requests PER SPINDLE. There will also be less wait time for unrelated requests. This implies that the more disks you can put in the array, the better the performance. And this may be where SW RAID could be better than HW RAID, since SW based arrays can span multiple controllers. The controllers also do not need to be the same interface type either. You can mix IDE, SCSI, etc. HW RAID systems generally have some limits on the number of disks you can have, based on the number of internal buses and bus width (ie a two internal narrow SCSI channel system would be limited to a maximum of 14 hard disks). If you are concerned about write performance more than read performance, you might want to consider using a mirror set of some sort (RAID1 and RAID6 [AKA RAID10]). Since there is no parity calculation, write performance is very close to a standard disk's. The disadvanage is that 50% of the capacity is lost. Is the above more or less correct. King On Mon, 1 Jun 1998, Bob McGowan wrote: On Thu, 28 May 1998, Leandro Guimaraens Faria Corcete Dutra wrote: snipped The article from www.osnews.com did say that software raid takes up CPU cycles, but it did not say how much. It would seem that if the CPU must check for errors on each byte from disk and performance would take a big hit. Perhaps the kernel checks for errors only if it knows that a disk died, and normally there would not be a hit. Does anyone know about CPU hit of software raid. Why would anyone buy expensive raid hardware if software does the same without too much penalty? King Lee First, the CPU not only checks for errors on reading, it must also calculate the parity on writes. In RAID5, spanning 4 disks, for example, 1/4 of the storage is used to hold parity info. Data is written in stripes of some size, one stripe per disk, in a round robin sequence. One stripe will be parity. In the above 4 disk example, if a stripe were 16K in size, there would be 48K of data and 16K of parity. In RAID5, the parity stipe will rotate between disks, so no single disk is loaded with all the parity (this improves performance over RAID4(I believe) where all parity is on one disk). If a disk write is less than 48K, the system must read 48K from the disks, make the needed changes, recalculate parity and write the resulting 64K back to the disks. If the size is 48K, this read of data can be dispensed with. The system must then only calcualte the parity and then write the 64K. This means CPU cycles are needed for SW RAID. I do not know the impact in terms of actual numbers, but I can say the main issue is scalability. In SW RAID, the more RAID subsystems created, the greater the impact on CPU performance
RE: set clock to GMT?
-Original Message- From: Rafael Cordones Marcos Subject: Re: set clock to GMT? On Thu, Aug 27, 1998 at 09:19:26AM +0100, Vincent Murphy wrote: OK, sorry for not making myself clearer in the first place. -deleted discussion on GMT By the way, (for anybody listening) when I live my PC on for several days I have found that the hardware clock and th system time differ in HOURS. Is that OK? Should I use cron to update the hardware clock every now and then? Rafa, Generally, I think setting the hardware clock from software is not a good idea unless you have gotten the current software time from some other reliable source. My understanding of the software clock is that it will always drift by some amount due to system activities. The hardware clock should be more accurate than the software clock. If your hardware clock is going off by hours, I would tend to suspect the hardware clock itself (the chip or its power supply). Then you would want to use the hardware clock to periodically update your system software clock. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com
RE: find replace?
Subject: find replace? Hi, is there a console utility to find and replace in a file? Preferably on which can find replace from a bunch of files as passed by 'ls' or something? Thanks, Timothy The sed stream editor is the tool to use. To give it the list of files to process requires some shell knowledge and a little additional information from you. Do you want each processed file to be concatenated on stdout or should each file be saved as an individual processed file? And which shell are you using? --- Bob McGowan i'm: bob dot mcgowan at artecon dot com
RE: Serial Console
From: Tim Sailer [mailto:[EMAIL PROTECTED] Sent: Friday, September 11, 1998 9:18 AM Subject: Serial Console Has anyone actually gotten the serial console stuff to work with the newer kernels? It works to a point, with lilo allowing input from the serial port, and the initial kernel messages going to the port, but right after it initializes the swap, all I/O stops at the serial port. When the system is rebooted, the shutdown messages go to the port. I'm clueless at this point... Tim, It's been a while since I did this (and on the pre-2.0 kernel) but one thing you did not mention... IF ;-) I remember correctly, Linux has some devices specifically set for console and perhaps system tty (like /dev/console /dev/syscon /dev/systty from my Solaris UNIX system). When I tried the serial console setup, I renamed these files (so I could recover if needed) and then set up links with the original names to the serial device being used as the console. IHTH and Good Luck, --- Bob McGowan i'm: bob dot mcgowan at artecon dot com
Re: Relabel partition didn't work
Manon Metten wrote: On 4/5/07, *Douglas Allan Tutty* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: On Thu, Apr 05, 2007 at 07:06:51AM -0500, Manon Metten wrote: Linux debian 2.6.18-4-486 #1 Mon Mar 26 16:39:10 UTC 2007 i686 GNU/Linux I want to change the name and access point of a partition on my second hd. It's labeled /xyz now (coz I could think of no better name when installing etch). I tried this: e2label /dev/hdb4 xyz e2label /dev/hdb4 store e2label /dev/hdb4 store So seemingly the label has changed. I edited /etc/fstab accordingly: changing /dev/hdb4 /xyzext3defaults0 2 to /dev/hdb4 /store ext3defaults 0 2 Then I rebooted, only to find this message popping up during boot time: mount: mount point /store does not exist There was no further error during boot time and kde was up and running. Then I'd reset everything and could mount and access /xyz as before. What did I do wrong and how do I change /xyz to /store? You're confusing disk lables with mount points. Your fstab doesn't have disklables in it. You told mount to mount /dev/hdb4 on /store, so it looks for the directory /store, which doesn't exist. So backup: What are you trying to do? Doug. Hi Doug, After fiddling around with sarge for some months, I installed etch a couple of weeks ago. I've done lots of reading and I'm following this list for a while, but I'm new to debian anyway. I have 4 partitions on my 2nd hd, swap, /tmp, /var /xyz. /xyz is the 4th primary partition on my 2nd hd. When installing etch, at some point partman asked for a mount point for that partition and I entered /xyz. I use this xyz partition only to store some .iso's and other large files, so I want rename/relabel it to /store. If I do mkdir /store, that would create the dir/mount point on my 1st hd where / is located (correct?). Thus the question is: how to change /xyz to /store on my 2nd hd, so I can do eg. something like this: mv ~/*.iso /store so it moves *.iso from ~ on my 1st hd to /store on my 2nd hd. Greetings, Manon. @ Joe Heart What in the world is e2label? It doesn't show up in my search. I fould e2label in /sbin, read the man and tried to use it. What is it today, national People from other distros day? Huh? Sounds like a warm 'welcome to debian' to me. I guess you don't have the monopoly of wisdom too. BTW: it's SHE, not he. Anyway, greetings to you too, Manon. Hi, Manon, There's nothing to do to the second hard disk. Think of the mount point as directions to the kernel on where to go look for a disk. But this is a logical reference to a physical device. So, first, some background on devices. Generally, the device names are pretty much fixed in a system, over time (this does not hold very well for SCSI devices, particularly when using pluggable storage such as USB, but that's another story). So, for a pair of IDE type hard disks, you will have the device names 'hda' and 'hdb'. The partitions on these devices then add numbers to the basic name: hda1hdb1 hda2hdb2 And so on. The above is complicated by the presence of an extended partition, but it doesn't sound like you have one. So, for your 2nd hard disk, you should have disks named 'hdb1', 'hdb2', 'hdb3', 'hdb4'. These are names the system gives to the devices, which can be found in the /dev directory. These names have nothing to do with labels. Debian will, by default, make labels for disks based on the mount point given during installation. Anything done after that will not get a label unless you put one on it. And though labels can be used to help get around the pluggable device problem mentioned above, the default setup just uses the disk names and directory mount point names. This next piece can be a bit confusing, but I think is essential to understanding how all this works: Each and every disk partition has its own 'root' directory. This is what users of DOS/Windows systems have to put up with: C: has a \, D: has a \, A: has a \, on and on This is hidden in unix/linux systems by mounting. What a mount does is to tell the kernel that when a mount point is accessed on disk1, move to disk2 root and process from there. This might look better as a simple picture: BootDisk HardDisk2 / point)- / etc usr (mount bob jim jane What you work with is simply /usr/bob or /usr/jane, without concern for which disk it's on. So, you want to change the mount point? Create a directory (store in your case) somewhere in the filesystem. Others have suggested /mnt or /media, but there are really no
Re: backup-restore question
Chris Parker wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Question being is it possible to restore a backup from a win32 (2003 server) ntfs filesystem via tape onto a linux machine. I want to be able to test our backups to make sure all is restorable. If this is possible how to do it into an ext3, ext2 or other linux native filesystem? thanks all for your comments. Chris Parker -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGFXiGfnAfmpxxbc4RAhuMAKCiSKbpynef1+bzcFq3zWq/1r4JIACglBXh 3ebzus1I6thOutGaUvjPPnw= =3FUr -END PGP SIGNATURE- Short answer, 95%+ probability, no can do. Longer answer, it depends on the backup/archiver tool used. For example, if you perchance have Cygwin installed on your windows system and used one of the tools that are common to it and UNIX/Linux (tar and cpio come to mind), the answer is an unqualified yes. More likely you used a Windows tool, which probably has a proprietary archive format, so it can't be read from a Linux box. Or, you may have used a tool that's proprietary but that has both Windows and Linux versions. This may work. Then, there's the question of whether the backup is an image of the filesystem itself or just copies of the files. If it's an image (a 'true' image, such as a 'dd' would make), then you might be able to use dd to read the tape into a regular file on the Linux box and then mount that using the loop back device feature: dd if=/dev/tape of=/path/to/file bs=10k # or something like this. mount -o loop -t ntfs /path/to/file /path/to/mount/point Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: permission of shadow file and upgrade the kernel
Douglas Allan Tutty wrote: On Mon, Apr 09, 2007 at 12:47:23PM -0700, ann kok wrote: Hi all why the permission of the shadow file in debian is 640? ---deleted 1. What do you think the permissions of shadow should be? The only user who needs to read /etc/shadow is root, that is the whole point of having shadow passwords. ---deleted Doug. One might wonder why it isn't just 600, if the only user needing access is root? The answer may be in the permissions and owner/group: -rw-r- 1 root shadow It would appear there are (or could potentially be) tools that need to only read the file. Rather than make them set uid to root, which would give them rw permission, they are set gid so they have ro permission, which limits the damage they could potentially do. Bob smime.p7s Description: S/MIME Cryptographic Signature
compilation issues with a Gnome based app.
Hi, all, I'm trying to build an application called gASQL, version 0.6_0.2.94 and it's bombing out in the configure script. Bottom line is that it wants to use 'gnome-config' to find out about the parts that are installed, but when run the way the configure script sets it up, it fails to find things like bonobo, which so far as I can determine are there. I have been able to get pkg-config to provide the information needed, and have munged around in the the configure script, trying to figure out what else might be going on. So, I find that configure tries to compile a C program to figure something out about the Gnome environment, and fails as it can't find the include file 'bonobo/gnome-object.h'. I used the Debian package search page to try and find this file (in all four releases) and it came up empty. A Google search does find references to it, so it must be in some dev package I can't find. Any ideas on where to go from here? Thanks, Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Xen Upgrade to 32-bit PAE on Etch: Advice Please
$ apt-cache search hypervisor libc6-xen - GNU C Library: Shared libraries [Xen version] xen-hypervisor-3.0.3-1-i386 - The Xen Hypervisor on i386 xen-hypervisor-3.0.3-1-i386-pae - The Xen Hypervisor on i386 with pae xen-linux-system-2.6.18-4-xen-686 - XEN system with Linux 2.6.18 image on i686 xen-linux-system-2.6.18-4-xen-vserver-686 - XEN system with Linux 2.6.18 image on i686 It looks like the last two items listed are the generic install names you want, that depend on the latest versions of kernel and hypervisor. Also, it looks like they both depend on the PAE version of the kernel. Bob Grok Mogger wrote: Hey, I am currently running the 32-bit non-PAE version of Xen on Debian Etch. I installed all this with the Xen packages via aptitude. I want to upgrade to the 32-bit PAE enabled version of Xen. I think this *should* be as easy as aptitude install xen-hypervisor-3.0.3-1-i386-pae followed by an aptitude remove xen-hypervisor-3.0.3-1-i386. Then check to make sure that my /boot and my grub menu look like they should, and change all the DomU config files to use the new PAE enabled kernel which will be installed, reboot and voila. Done. Never done anything like this before though, so I thought I'd ask here first to make sure I know what I'm talking about. =) I also have one other question. Is there a package I can install that will make sure I always have the most up-to-date version of the Xen Hypervisor? With the kernels I'm used to seeing something like a linux-image-686 that always depends on the latest version of the kernel. So if I install that, I can make sure that my kernel is always up to date. I don't seem to see any such package for Xen though. Thanks, - GM smime.p7s Description: S/MIME Cryptographic Signature
Re: Impact on changing the shell of default system accounts
Felipe Rocha wrote: Hello, I would like to know which is the impact on changing the value of the default bash '/bin/sh' of the system accounts. If it doesn't have problem, which option ('/bin/false', '/usr/sbin/nologin' or '/dev/null') best fit on this situation? Here you have a list of accounts from '/etc/passwd' that I would like to change: daemon:x:1:1:daemon:/usr/sbin:/bin/sh bin:x:2:2:bin:/bin:/bin/sh sys:x:3:3:sys:/dev:/bin/sh sync:x:4:65534:sync:/bin:/bin/sync games:x:5:60:games:/usr/games:/bin/sh man:x:6:12:man:/var/cache/man:/bin/sh lp:x:7:7:lp:/var/spool/lpd:/bin/sh mail:x:8:8:mail:/var/mail:/bin/sh news:x:9:9:news:/var/spool/news:/bin/sh uucp:x:10:10:uucp:/var/spool/uucp:/bin/sh proxy:x:13:13:proxy:/bin:/bin/sh www-data:x:33:33:www-data:/var/www:/bin/sh backup:x:34:34:backup:/var/backups:/bin/sh list:x:38:38:Mailing List Manager:/var/list:/bin/sh irc:x:39:39:ircd:/var/run/ircd:/bin/sh gnats:x:41:41:Gnats Bug-Reporting System (admin):/var/lib/gnats:/bin/sh nobody:x:65534:65534:nobody:/nonexistent:/bin/sh fetchmail:x:104:65534::/var/lib/fetchmail:/bin/sh Thanks, Felipe Rocha You will not be able to do a 'switch user' to those accounts. That is, even as root, an 'su daemon' will fail, since the named shell would simply exit. This could be a big problem, depending on how the system handles running startup scripts (I haven't looked to see exactly how Debian handles it). Many systems will use 'su XXX' in rc scripts, where XXX is one of the 'system' names, to start programs related to that system 'service'. For example, an 'su mail' might be done before starting 'sendmail' (or other MTA), so it does not run with root privilege. This reduces the impact of security problems to things owned by 'mail', rather than opening the whole system up to the exploiter. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Impact on changing the shell of default system accounts
Bob McGowan wrote: Felipe Rocha wrote: Hello, I would like to know which is the impact on changing the value of the default bash '/bin/sh' of the system accounts. If it doesn't have problem, which option ('/bin/false', '/usr/sbin/nologin' or '/dev/null') best fit on this situation? Here you have a list of accounts from '/etc/passwd' that I would like to change: daemon:x:1:1:daemon:/usr/sbin:/bin/sh bin:x:2:2:bin:/bin:/bin/sh sys:x:3:3:sys:/dev:/bin/sh sync:x:4:65534:sync:/bin:/bin/sync games:x:5:60:games:/usr/games:/bin/sh man:x:6:12:man:/var/cache/man:/bin/sh lp:x:7:7:lp:/var/spool/lpd:/bin/sh mail:x:8:8:mail:/var/mail:/bin/sh news:x:9:9:news:/var/spool/news:/bin/sh uucp:x:10:10:uucp:/var/spool/uucp:/bin/sh proxy:x:13:13:proxy:/bin:/bin/sh www-data:x:33:33:www-data:/var/www:/bin/sh backup:x:34:34:backup:/var/backups:/bin/sh list:x:38:38:Mailing List Manager:/var/list:/bin/sh irc:x:39:39:ircd:/var/run/ircd:/bin/sh gnats:x:41:41:Gnats Bug-Reporting System (admin):/var/lib/gnats:/bin/sh nobody:x:65534:65534:nobody:/nonexistent:/bin/sh fetchmail:x:104:65534::/var/lib/fetchmail:/bin/sh Thanks, Felipe Rocha You will not be able to do a 'switch user' to those accounts. That is, even as root, an 'su daemon' will fail, since the named shell would simply exit. This could be a big problem, depending on how the system handles running startup scripts (I haven't looked to see exactly how Debian handles it). Many systems will use 'su XXX' in rc scripts, where XXX is one of the 'system' names, to start programs related to that system 'service'. For example, an 'su mail' might be done before starting 'sendmail' (or other MTA), so it does not run with root privilege. This reduces the impact of security problems to things owned by 'mail', rather than opening the whole system up to the exploiter. Bob After checking /etc/init.d/*, I found that on my etch system, only one file appears to have an 'su ...' in it, so my previous comments don't look like they are too important for a Debian startup. Still, I'd be awfully careful before trying to change them. Perhaps a test install, in a 'chroot' environment or under an x86 emulator would be a good thing to use, to test for possible impacts in basic functions. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: dumb questions about video editing and camcorders...
Michael Fothergill wrote: --- Dear Debianists, thanks for all the useful comments on video editing etc. I do have a few other questions that I have accumulated after investigating this a little further. I have nosing into the various different kinds of camcorders that are about these days. There seems to be quite a few different types. From what I could see it looked as though getting one with a hard drive on it might be the way forward rather than a tape or DVD. Your comments on this choice would be appreciated. I would suggest getting a camera with a hard drive. Higher capacity to cost ratio for one thing. Also, simply higher capacity, which means longer movies/multiple movies, before having to move data to other media. Also getting a camera with a large lens generally seems to be helpful for a number of reasons including motion and light sensitivity. The more zoom power, the better, so long as it's *optical* and *not* digital. Forget digital zoom, it offers little in return for the loss in quality. I'm not familiar enough with video editing to know how this applies to video images, but with still photos, you can do the digital zoom in your photo editor (Gimp, or whatever) and have it under your full control, rather than letting the camera do it for you with minimal control. In darkroom terminology, you're 'enlarging' the image. With optical zoom, you will maintain full digital image quality throughout the zoom range (remembering that at longer focal lengths, hand shake can contribute to loss of quality - so image stabilization is also a good thing to have). --- Would most people who have Nvidia cards and Etch on their machines and that subscribe to this user list be using the proprietary driver? I use the proprietary driver module because of its vast improvement in things like Googleearth. --- bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Oh, yeah: Unable to switch to TTY from X
Amy Templeton wrote: I probably should've put this in the last email I sent out, but it is kind of off-topic for that and it just occurred to me. Since installing Debian (a while back), I'm unable to get back to a TTY after I invoke startx. --- I presume you mean that using the 'Alt-Ctl-F#' keys fails to switch to a tty console, but since you don't explicitly state what you're trying, I'd like confirmation before saying anything more. ;) Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: audacity number fonts
Andrew Sackville-West wrote: On Fri, Apr 20, 2007 at 08:06:17PM +0100, Alan Chandler wrote: On Friday 20 April 2007 19:27, Andrew Sackville-West wrote: On Fri, Apr 20, 2007 at 04:47:19PM +0100, Alan Chandler wrote: I recent update to debian sid has had a strange side effect. When I run audacity, most of the text is fine, but where there are numbers they are all shown as 0. Anyone else experiencing this? looks fine here. I sometimes feel I have a completely different version of Debian to everyone else :-( I know the feeling. I'm going through that with oocalc lately. on one machine certain files won't open and crash the program, but work fine on others. bah. to your problem. I'd guess its more of a font issue than an audacity issue. I know that doesn't really help, but maybe it will steer you in the right direction. A I'm actually having font problems with Audacity, as well, but have not had time to figure anything out. In my case, menu titles and sub-menu lists are practically unreadable. This also impacts some TCL/Tk based apps like tkman. So my problem sounds a bit more general in nature. I have a second system, basic etch with KDE and Audacity, in which everything is fine, so, when I get time, I'll be comparing between the two to see if anything obvious can be found. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: opinion sought: lvm vs. evms
Miles Fidelman wrote: I run a couple of rack-mount servers, each containing 4 fairly large SATA disks. I'm getting ready to rebuild one of the boxes as a development/staging environment. I'm considering using Xen, with an etch Dom0, and wondering what's the best way to manage the underlying disk space for maximum flexibility, with a level of reliability. (Actually, my preference would be to run opensolaris with ZFS as Dom0, but that's not quite there yet). I'm torn between using md to raid the disks and running LVM on top of that, or using EVMS to handle it all - and looking for opinions on the pluses and minuses of each (and, for that matter, opinions on how well EVMS is still supported). Thoughts, suggestions, opinions? Thanks very much, Miles Miles, I have not got a lot of experience with EVMS but did install it on my home etch system. It 'works' in that I was able to configure my setup successfully and have been using the single RAID5 device I created without issues (it's my home directory filesystem). However, there are problems with a couple parts of it: One, when the system tries to mount the managed volumes during boot, I get a mess of errors (I don't recall exactly, they have something to do with references to undefined elements). Second, the EVMS device sometimes fails to mount at all during boot, because the system thinks it's already mounted. So there are some sort of cleanup issues during shutdown. I've had to add a temporary init script to check and see if my home filesystem is mounted and if not, attempt to mount it again. All in all, if it worked as advertised, I think it would be a great tool. And, perhaps, if it were used on a pristine, newly installed, system, things might go better. But as it stands, I'll be reconfiguring my system soon and will be removing it, for now. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: to allow root logins or not?
Greg Folkert wrote: On Fri, 2007-04-20 at 15:53 -0500, Default User wrote: Gee, I hate to ask another question, but - During an Etch install, it asks if I want to allow root logins. If not, no root account is set up (I guess as a security measure), and all admin access is done by sudo. Now I normally do almost all admin work as sudo, but is there a downside to not having an actual root account. That is, might there be instances when something really should (or must) be done in a real root account (not sudo), and with no root account, the user is hung up? Not recommended, but possible: sudo su - or if X is needed: sudo sux - Make sure you sudo aptitude install sux for the second one. As I understand the OP's question/statement, the question in the install appears to be whether to create the 'account' for root or not. It apparently does not say it will create an account without login privileges. Since both 'su' and 'sudo' commands require an account for validation of the user name and determination of privilege, I would assume the root account is in fact created but with a shell such as '/bin/false' or with an 'impossible' password, so as to prevent logging in while still allowing use of either 'su' or 'sudo'. This would effectively prevent login to single user mode via 'sulogin', since it requires a password. In those cases where the startup process drops a user into a root level shell without prompting for a password (this can happen very early in the boot sequence, for example if an fsck of / fails and the disk is unreadable), all would be well. Given that sometimes root level access is required through 'single user' mode, I'd strongly encourage allowing root login. Just be sure to use a very good password and don't abuse the privilege of being root. If you really want to not have root login, you would need to have a rescue CD or equivalent method of booting the system, to fix anything that couldn't be handled by a normal startup from the boot hard disk. Though these may be quite rare, they can happen. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: UUID vs /dev
Michael Pobega wrote: On Tue, Apr 24, 2007 at 04:03:13PM +0200, Andrea S. Gozzi wrote: I noticed that ubuntu 7.04 switched to the UUID method for drives/partitions identification in fstab. I heard rumors that debian will soon do the same (or already does). Even with google I couldn't find any *impartial* comparison of the two methods for identifying hardware. Since it seems to be what most distributions will do from now on, I wanted to discuss its benefits and/or flaws. Personally I don't see any point in using UUID=xxx instead of /dev/xxx (also beacuse to get the ID you need 'vol_id -u /dev/xxx'), but there are surely pros I'm not aware of... Andrea I think UUID is used because it is better to use UUID to recognize the drives than /dev. For example, if you're trying to have your USB drive automount to /mnt/usb, you'd use something like /dev/sdb1 in fstab, correct? What if you plug in an external drive? That will be picked up at /dev/sdb, and following that the USB thumbdrive will be picked up as /dev/sdc. The reason for UUIDs is to make it so that the computer can recognize the drives by their device ID rather than the order they were plugged in. I may be wrong though. This is what I've been told. This is correct, but is a 'high overview' level description. A few more details, maybe, will help understanding what's going on (and the only downside I know for it). First, the UUID and LABEL methods work in basically the same way, but the UUID is generally considered better, particularly for drives that migrate between systems (such as USB or Firewire). There is less of an issue with colliding values between multiple systems using the UUID. On booting, the system scans devices and determines the UUID (or LABEL), and creates symlinks from the UUID/LABEL directories (/dev/disk/by-uuid or /dev/disk/by-label) pointing to the actual device node associated with the UUID/LABEL. So the system can find a device node, using the LABEL or UUID. Once found, the system uses the device node. It's this that leads to the single 'downside' that I've seen, of using UUID/LABEL: Once mounted, the system 'forgets' the UUID/LABEL, so output of the 'mount' command lists device nodes. Same for 'df'. And, KwikDisk (KDE applet) will show two items for a particular mount point, the content from fstab (LABEL/UUID) and the device node actually mounted and used to access the device. It's pretty quickly obvious (for the moderately experienced user) which element to use to umount the device, but the multiple entries can become confusing if you have several disks or partitions set up this way. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: the mysql postgresql question.
Dave Patterson wrote: Hi all - is it possible to go with one database system for all package dependencies? Package foo depends on mysql for install, Package bar allows postgresql or mysql but requires one or the other. Package umpty-scratch prefers postgresql. Is it necessary to have both database systems installed in order to have foo,bar, and umptyscratch; Or can I pipe a dependancy somehow? Cluebricks welcome. Cheers, Dave It depends ;( Your 'foo' package depends on mysql. This means there's a front end program that interacts with the DB, and a schema used to build the DB structure used by the front end. This schema probably has mysql specific functionality, which could (probably does) translate into mysql specific functionality used in the front end. The other two cases described sound like the schemas are generic and so can be used with any SQL compliant database. So you should be able to install mysql, followed by the three packages, and have everything work (perhaps with a bit of configuration effort for the second and third packages). But if you want/require PostgreSQL for those two, you'll still need mysql for the first. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: hda:lost interrupt!!!
Owen Heisler wrote: On Mon, 2007-04-30 at 08:48 -0700, Andrew Sackville-West wrote: On Mon, Apr 30, 2007 at 05:26:13PM +0530, shyam narayanan wrote: hi all, I had a perfect debian installation and as i boot up i am getting the error messsage hda:lost interrupt hda:DMA interrupt recovery hda:dma_timer_expiry:dma status == 0x24 these three messages are coming in loop each taking some time :( mine is an intel pentium D processor and HDD is a 40gb seagate of model ST340810A I'd look for potential hardware failure of that hd. Yeah, prepare for that drive to completely melt down soon, so you aren't taken by surprise if it actually does. I had a system with a chipset fan that would die intermittently, allowing the chipset to overheat. By the time I figured out what was happening and got the fan replaced, the chipset had been damaged. I got DMA errors (like yours) a lot on the second IDE controller. Other than that, I didn't have any problems. I would suggest also doing some tests (memtest86) and disabling DMA on that drive. Also, perhaps a bad cable could cause that...? Unlikely I suppose. I've had 'lost interrupt' problems on a couple of machines, but not related to hard disks (at least for the second system). I recently purchased an HP AMD64 laptop, planning to install Debian's 64 bit version. During install, all was OK, but on booting the new system I got intermittent hangs. Googling came up with several possible work arounds. The one that works, in my case, is to boot with the kernel option 'noapic' ;( Now, I get errors about IRQ 7, nobody cared, and the system disables it. According to KDE info center, int7 is related to the USB ehci controller. This disabling happens whether or not there's anything connected to the USB ports. Since the system doesn't stay up for long if booted without that option, I cannot determine if this IRQ would get disabled in any case, or if it's because of the option. In any case, it would seem to me that possible bugs in the relevant kernel drivers or IRQ routines could contribute, as well as buggy or otherwise broken APIC chips. It would be nice to know if there are any diagnostic tools that could be used to help isolate the problem. I don't know much about memtest86, but the name would imply it may not be helpful in this case. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: replacing /usr with a new mountpoint
Martin Marcher wrote: Hello, my setup is in a 30GB partition with LVM on top. now i had something like the following initially set up / 1GB /home 3GB and a few other non standard mountpoints ok i found that although this is just some minimal system for testing the / partition is to small (more precise /usr is eating up too much space) so I set up a new LV for /usr rsynced mounted and modified fstab to fit the changes. of course I'd like to regain the space that the /usr directory on the / partition uses. Could I just telinit 1 umount the /usr mountpoint empty out the /usr directory remount again and telinit 3 back to normal? thanks for the help I'd suggest booting from a rescue or live CD, mount your system disk root on /mnt and delete contents of /mnt/usr. On reboot, all should be well. There's no guarantee that going down to state 1 will in fact kill off every process that has something open in the /usr directory. In theory, it should, and what you suggest should work, but I've seen a number of cases where this doesn't happen, completely. Besides, there are a lot of programs in /usr/bin that could turn out to be useful, and using the CD will provide them. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: replacing /usr with a new mountpoint
Douglas Allan Tutty wrote: On Thu, May 03, 2007 at 11:54:10AM +0200, Martin Marcher wrote: On 5/3/07, Douglas Allan Tutty [EMAIL PROTECTED] wrote: Somewhere in the debian documentation is a warning that after going to single-user mode a return to multi-user is not guaranteed to work. too bad i'm trying to do all of that without actually rebooting (more a matter of because it should be possible not a requirement) Reboot into single user (with the -s option if there isn't a grub menu item already) so that you know noting under /usr is being used, mv /usr to /oldusr, fix fstab so that the new usr mounts on /usr, then shutdown -r. Of course be careful not to use any binaries that reside under /usr. Stick wit straight bash and other stuff under /bin. Use the full path to make sure. all of this is done and the system already works with the new /usr mountpoint I'd just like to regain the space without rebooting - to be honest this is the whole point of this exercise. I'm not understanding. Do you mean that you mounted /usr over /usr without emptying it? If so, and you insist on non rebooting, then at least stop X and as much else as you can (as a precaution), then umount /usr, which will now show your full /usr directory tree, mv /usr /oldusr, mkdir /usr, fix owners and permissions to match /oldusr, remount /usr, and if everything is working, rm -rf /oldusr. Note that existing running apps that have files from /usr open will continue to work since open files are not unlinked until they are closed. Which will prevent being able to umount /usr in order to mv the underlying /usr to /oldusr. This is why it's necessary to go to single user mode, which 'should' kill any process with open files in the /usr tree. Good luck, Doug. Since it's necessary to go single user anyway, what's the difference between getting there from runlevel 2 versus rebooting to it? All users need to be told to save work and log off, in either case. The only diff I can see would be for a large server system that could take forever to reboot. Anyhow, as an exercise, you might want to consider going to runlevel 1 as noted earlier, then using 'ps' to see if there're any running processes that should have died but didn't. And, use 'fuser' to see if any of these are using the /usr tree in any way. You should also save a copy of the 'ps' output to a file for future reference. You can then kill whatever may need killing to allow the umount to succeed. Then do the 'umount/mv/remount/remove' steps described earlier to regain the space and return to runlevel 2. If there were processes normally started by rc scripts going to level 2, where the process is already running, you should see errors for those scripts, though it's possible for some to fail to detect there's already a running process, resulting in two copies executing. Getting the interplay of rc scripts for various runlevels *and* runlevel transitions right is an arcane art, quite difficult to master (I make no claims as to mastery, just the difficulty of achieving it;). So, if you find that some programs end up with two copies running (you can check by comparing with the file created from the 'ps' output, above), you can manually kill them. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: order of IDE drives in lenny
Towncat wrote: I needed to re-install my Debian system due to a disk failure. (Used to be etch, now lenny.) I have a new primary IDE master to which I installed, and I also have an IT8212 IDE card in the machine. For some reason when booting, the IT8212 gets assigned the /dev/hda-hdd, while the onboard controller is assigned /dev/hde-hdh. During installation the onboard controler was / dev/hda-hdd, and therefore the root file system is not at its right place when booting, so the system does not start. How can I tell the kernel the order of loading the controllers? I temporarily removed the IT card, but I will need it and the drive attached to it. Now, however, the onboard IDE is /dev/hda-hdd, and the system is running. You will need to use either the LABEL=... or GUID=... options in your boot and fstab files. There have been several posts on this topic, going into the gorry details (several from me, try searching the archives for them). A quick synopsis: you need to be sure there is a usable 'label' on your disk partitions (programs to check/create labels vary per FS used). But in all cases, if you have a root partition with a label of just slash (/) you will need to change it to something like '/root' in order to use it. GUID values are automatically created, but are very long values. They are most useful for USB devices that migrate from system to system, as they should be unique even across systems. You will need to add the proper line to the boot configuration for the root partition, so the kernel uses it, and use labels/GUIDs for all mountable partitions mentioned in /etc/fstab. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Random Crashes
Douglas Allan Tutty wrote: On Mon, May 07, 2007 at 04:54:18PM +0200, Gregor wrote: Eduardo M KALINOWSKI wrote: So I took of the old 512Mb ram module, because it should be the one with problems, since the crashes happened already when I had only that one. The system still crashed. Just to be sure, I put it on again, and only this one, and the system also crashes. The motherboard has two slots for RAM. I tried both modules in both slots, and I did notice that when a module (either one) is in one of the slots, the system crashes just after boot --- at most I can type the password and let KDE start, but it crashes before KDE is fully loaded. With a module in the other slot, then the system is usable most of the times. So, am I really unlucky to have two memory modules with problems, or what else should I suspect? Motherboard? Processor? What would be the possible ways to diagnose the problem? To me it sounds like a hardware fault somewhere along the path to the memory slots with more problem on one than the other. I'd say swap the processor but most people don't have a spare hanging around (ditto spare MBs), and there's the heat-sink issue. Do you have a spare system that takes the same kind of memory you can try your sticks in? As I see it, the problem with relying on something like memtest is that it tests the whole memory system not just the sticks; a faulty MB on the memory path can show as bad memory. As far as what to suspect, there's really only three things: MB, CPU, memory sticks. Suspect all. Doug. I recently purchased a Turion x64 dual core based HP laptop, and have loaded the AMD64 build of Debian stable (etch). The system installed OK but on booting the newly installed system, it would freeze (hard) at unpredictable points. After much googling, searching the HP FAQ's for their Debian support (not available for any laptops, but on servers - however, there are known issues with hangs like this), and so on, it looks like the problem is kernel support/interaction with the APIC system. The HP FAQ is here: http://h20219.www2.hp.com/services/cache/442408-0-0-225-121.html No way to tell if this is your issue or not. What worked for me, which at least lets me boot and use the system, was to add 'noapic' to the boot command line (I use grub, so that's in /boot/grub/menu.lst, similar edit for lilo with the 'append=noapic', IIRC). There were other boot options mentioned (I don't have the details at hand, I'm sorry to say), but none of them helped in my case. Hope this is helpful. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: etch in amd64: system freezes
Marcelo Chiapparini wrote: Dear list, I have installed etch in a new AMD64 machine. My problem is the following: after a very little activity in the Gnome desktop, the system freezes (even an open window terminal freezes) and only a hard reboot bring the system back. If during the boot I interrupt the window manager with ctrl+alt+f11 and switch to a terminal, with alt+f1 for example, the system works perfectly. The problem seem to be with the graphical interface. I ran memtest86+, and everything is ok, no error was reported. My system is the following: CPU: Athlon64 2043 MHz Mem: 1024 Mb Chipset: VIA K8T800 pro (Abit AV8 motherboarad) RAM: 204 MHz (DDR408) / CAS: 3-3-3-8 /DDR-1 (128 bits) Video card: ATI Radeon Mobility 9200 (AGP) I googled the internet and looked the debian arquives too (debian-user and debian-amd64), but I didn't find any clue. I will appreciate very much any help. Thanks in advance Marcelo I recently purchased a Turion x64 dual core based HP laptop, and have loaded the AMD64 build of Debian stable (etch). The system installed OK but on booting the newly installed system, it would freeze (hard) at unpredictable points. After much googling, searching the HP FAQ's for their Debian support (not available for any laptops, but on servers - however, there are known issues with hangs like this), and so on, it looks like the problem is kernel support/interaction with the APIC system. The HP FAQ is here: http://h20219.www2.hp.com/services/cache/442408-0-0-225-121.html No way to tell if this is your issue or not. What worked for me, which at least lets me boot and use the system, was to add 'noapic' to the boot command line (I use grub, so that's in /boot/grub/menu.lst, similar edit for lilo with the 'append=noapic', IIRC). There were other boot options mentioned (I don't have the details at hand, I'm sorry to say), but none of them helped in my case. Hope this is helpful. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: order of IDE drives in lenny
Bob McGowan wrote: Towncat wrote: I needed to re-install my Debian system due to a disk failure. (Used to be etch, now lenny.) I have a new primary IDE master to which I installed, and I also have an IT8212 IDE card in the machine. For some reason when booting, the IT8212 gets assigned the /dev/hda-hdd, while the onboard controller is assigned /dev/hde-hdh. During installation the onboard controler was / dev/hda-hdd, and therefore the root file system is not at its right place when booting, so the system does not start. How can I tell the kernel the order of loading the controllers? I temporarily removed the IT card, but I will need it and the drive attached to it. Now, however, the onboard IDE is /dev/hda-hdd, and the system is running. You will need to use either the LABEL=... or GUID=... options in your boot and fstab files. There have been several posts on this topic, going into the gorry details (several from me, try searching the archives for them). A quick synopsis: you need to be sure there is a usable 'label' on your disk partitions (programs to check/create labels vary per FS used). But in all cases, if you have a root partition with a label of just slash (/) you will need to change it to something like '/root' in order to use it. GUID values are automatically created, but are very long values. They are most useful for USB devices that migrate from system to system, as they should be unique even across systems. You will need to add the proper line to the boot configuration for the root partition, so the kernel uses it, and use labels/GUIDs for all mountable partitions mentioned in /etc/fstab. Bob Duh... that first cup of coffee doesn't seem to have helped. That's 'UUID' not 'GUID'. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: access to DVD movies
[EMAIL PROTECTED] wrote: Debian users, http://xinehq.de/ suggests xine_dvd_plugin for access to DVD movies. xine_dvd_plugin is not visible in Debian. What in Debian has a similar function? Thanks, ... Peter E. Desktops.OpenDoc http://carnot.pathology.ubc.ca/ You may also want to check out VLC, ...the VideoLAN project's media player. It handles audio/video files, DVD/VCD/... and network streams, in many formats. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Quote or double quote error in crontab
Valdir Marcos wrote: In crontab, I have: 00 01 * * 1-7 root/usr/bin/sarg -f /etc/squid/sarg-diario.conf -d `(date --date 1 day ago +%d/%m/%Y)`-`(date --date 1 day ago +%d/%m/%Y)` And this line generate the following message: From [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] Sun Apr 22 01:00:01 2007 Envelope-to: [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] Delivery-date: Sun, 22 Apr 2007 01:00:01 -0300 From: [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] (Cron Daemon) To: [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] Subject: Cron [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] /usr/bin/sarg -f /etc/squid/sarg-diario.conf -d `(date --date 1 day ago + X-Cron-Env: SHELL=/bin/sh X-Cron-Env: PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin X-Cron-Env: HOME=/x X-Cron-Env: LOGNAME=x Date: Sun, 22 Apr 2007 01:00:01 -0300 /bin/sh: -c: line 1: unexpected EOF while looking for matching ``' /bin/sh: -c: line 2: syntax error: unexpected end of file When I execute this exact line in the prompt there is no error message. The command is executed perfectly. How can I solve this problem? I'm not sure about the 'why' but I may have a solution. Put the command line in a script file and put the script file in the crontab. This has several advantages: 1. You can avoid problems like this one. crontab files are quite picky about their contents. 2. You can do all the fancy scripting stuff you're used to using in normal scripts (because it's in fact just another normal script). 3. You can edit the script at any time without having to deal with the crontab command. No worries about accidentally clobbering the crontab content, changing something incorrectly, etc. 4. You can run the script directly to test it. Immediate feedback is available. Generally, most real world admins use this technique (at least the ones I know;). Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: cron and CRON
Valdir Marcos wrote: Take a look: # pstree init─┬─atd ├─bash ├─cron─┬─firebird │ └─sendmail ├─cron───cron─┬─sendmail │ └─sh───curl ---deleted extra content # ps aux | grep cron root 14094 0.0 0.1 1764 820 ?Ss 00:21 0:00 /usr/sbin/cron root 14256 0.0 0.1 1848 708 pts/0R+ 01:42 0:00 grep cron r# ps aux | grep CRON root 12156 0.0 0.1 2056 936 ?SMay05 0:00 /USR/SBIN/CRON root 14098 0.0 0.1 2056 940 ?S00:28 0:00 /USR/SBIN/CRON Why is there two processes with cron? I don't see anything like this on my system. Also, I notice that the prompt for the second 'ps aux' is 'r# ' rather than just '# '. This could just be a cut and paste error, but if it's real, then you've done something between the two ps runs that we need to know about. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: [ in /usr/bin Question
bdeferme wrote: Tom Grove [EMAIL PROTECTED] wrote: Hello all...I am new to this Debian thing :-) I used it in the Woody days but moved over to the FreeBSD world for the last few years. I recently installed Testing (Lenny) and see the left bracket in my /usr/bin directory and do not know what it is. When I ls -al it I get: -rwxr-xr-x 1 root root 24752 2007-01-30 13:51 /usr/bin/[ This leads me to believe that it was installed with the base system or some package because I just installed the system earlier this week. Any help is much appreciated. Thanks. -Tom Grove Weird, Why don't you try to run it and see what it is? P.S. Don't run it as root, just to be sure :-) [EMAIL PROTECTED] www.paradize.be --To UNSUBSCRIBE, email to [EMAIL PROTECTED] a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] I'm not sure why 'test' and '[' exist as separate files, today. In the 'old' days, '[' and 'test' were simply hard links to the same file. This allowed the code to look at the name used to run it and set some specific default behaviors. The most important of these was that the name '[' required a final argument of ']'. This was done to make shell scripting cleaner and easier to read. Today, most shells have these built in, so there's no exec overhead to worry about, so performance is improved. But this now means you have to worry about quoting and escape processing. Old Bourne shells didn't have these built in, so it didn't try to interpret them and escaping was not needed. Other responses to this question have included a number of command line examples with various amounts of escaping applied. I have some comments on those: 1. ls /usr/bin/[ and ls /usr/bin/\[ are identical (the escape is not needed). This is because the shell only treats '[' specially when it is stand alone. And ls itself doesn't process special characters. 2. When the command in question does do regex processing, it is much easier to escape the string from the shell with single quotes than with multiple backslashes. dpkg -S '/usr/bin\[' is much easier to read (and type correctly), than dpkg -S /usr/bin/\\\[ Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: [ in /usr/bin Question
Douglas Allan Tutty wrote: On Wed, May 09, 2007 at 09:00:58AM -0700, Bob McGowan wrote: bdeferme wrote: Tom Grove [EMAIL PROTECTED] wrote: Hello all...I am new to this Debian thing :-) I used it in the Woody days but moved over to the FreeBSD world for the last few years. I recently installed Testing (Lenny) and see the left bracket in my /usr/bin directory and do not know what it is. When I ls -al it I get: ---deleted content--- dpkg -S '/usr/bin\[' is much easier to read (and type correctly), than dpkg -S /usr/bin/\\\[ No wonder I can never figure out what a shell script is trying to do, either way it looks like a cat on a keyboard. Give me python and fortran77 any day. Doug. As I don't use python, I have no direct experience here. I do use Perl, however. It is also 'easier' than the shell, assuming you put the script in a file. So, a question: Can python 'run' code directly from the command line, as Perl does with the -e option? perl -e 'while(){print}' for example. If so, you'd have the same quoting issues you have with a shell script, since you would need to protect the python part of the input from shell interpretation. Since the shell is both a command interpreter (runs other applications) and a programming language, it gets very complicated when it is working with another program that also uses wild cards or regex or simply general programming language constructs like parens or braces. Putting things in a file makes things even easier, assuming the program in question will read files (as Perl, python, awk, C, Fortran, Pascal ..., do). Unfortunately, this doesn't help a lot with the shell, the same problems exist in the shell script file as exist on the command line. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Are volume labels a file-system thing?
Hendrik Boom wrote: I just discovered that the commands for reading ans setting volume labels have 'e2' in their names, like e2label. Does this mean that I can label my partitions only if I put an ext2 of ext3 file system on them? Or is there some other mechanism I should know about? Labels are specific to each file system type, because the superblock is where it's stored, and each FS has a different format for that area on disk. You will need to do a bit of research, but each FS type should have a program available to create/edit labels. As you discovered, for ext[23] the program is 'e2label'. For XFS, it's 'xfs_admin'. I did some man searching, using the 'SEE ALSO' section, several times before looking for some of the specific programs in the /sbin and /usr/sbin directories. Upshot is that once you know the general form used for FS specific commands (e2 or xfs_ prefix, for example), you can just do: ls /sbin/xfs* /usr/sbin/xfs* And then use man on whatever command looks most promising. If you're lucky, the name will actually have 'label' in it ;) Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: hardware freezez and no clue where to start searching
AndiSHFR wrote: Hi. Maybe someone can give us a hint or direct us in the right direction. We have a debian testing system. uname -a shows: Linux sv-vmhost02 2.6.18-4-amd64 #1 SMP Mon Mar 26 11:36:53 CEST 2007 x86_64 GNU/Linux The machine seems to freeze every n seconds for some seconds. It can be easily seen when doing disk io. ---deleted examples--- Any idea where to start searching for? Thx Andi You may need to look into kernel boot options, such as 'noapic', 'nolapic' or 'pci=off'. These have all been suggested in various places for problems relating to AMD 64 kernels hanging in various ways, though none of the ways I've seen match what you're reporting. In those other cases the system hangs hard and has to be forcibly reset. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: [ in /usr/bin Question
Douglas Allan Tutty wrote: On Wed, May 09, 2007 at 05:08:33PM -0700, Bob McGowan wrote: Douglas Allan Tutty wrote: No wonder I can never figure out what a shell script is trying to do, either way it looks like a cat on a keyboard. Give me python and fortran77 any day. As I don't use python, I have no direct experience here. I do use Perl, however. It is also 'easier' than the shell, assuming you put the script in a file. So, a question: Can python 'run' code directly from the command line, as Perl does with the -e option? perl -e 'while(){print}' Yes, to some extent. Python uses the indentation of the program to indicate flow. Therefore, if your command is just a string of commands to run, it works. However, if it involves looping or decision it won't since python (by design I think) will not do the perlish and bashish one-liners. So: python -c a=5; print a; b=10/5; print b; print a * b works. OK, python has a switch that lets you do this. So, since shell quoting can also be used to escape newlines, you can do: python -c a=5 print a b=10/5 print b print a * b Which means that you could also, if you *really* wanted/needed to, write a short (several lines) script with proper indentation, looping, etc. Or, as Ron Johnson suggested, use a 'pseudo file' (in UNIX shell language, these are called 'here documents, I suppose because the standard input is redirected from 'here'). But even with them, care is needed to prevent (or allow) shell interpretation of the contents. In any case, though interactive processing is possible, putting the code in a file is much more useful, even for so called 'throw away' one liners, which I've often found I needed to use again and end up having to re-write. for example. If so, you'd have the same quoting issues you have with a shell script, since you would need to protect the python part of the input from shell interpretation. Yes, understandable, but since such use is literally throw-away code, there is no future maintenance to consider. On the other hand, there's no reason to submit to the limits of a one-liner when you can just as easily type: python This is also true for Tcl/Tk and can be done with Perl, using the right options (though in Perl, multi-line input is harder). and enter the python shell where you can do anything you want in python, indents, import modules, whatever. If you really want, you can operate the whole system from within the python shell as you would the bash shell. To make that easier, issue from os import * from os.path import * So that, to run a program, you issue a system call: system('lynx') But, these are intended for prototyping and testing code. They're not generally considered 'command interpreters'. This is what the shell is for, with the intent of making interactive use and command execution easy (no 'system' call, no parens, no quotes, just type the command name and go). That the shells became programming languages too only means there was a need that could be handled that way. The other scripting languages followed, as the need (and power requirements, as in speed, control structures, and so on) for them increased. The joy of python is that absolutly nothing is cryptic unless you want it to be. I understand that some languages tend toward being cryptic. But some (most of, I think) of that crypticness is due to programmers trying to be 'cute' or showing off or ... (as you said: ...unless you want it to be.) I am a firm believer in commenting code, and in being as 'obvious' as possible in the code itself. Some might call it compulsive but it's sure saved me a lot of heartache, not to mention sweat, blood, tears and time. This is true, regardless of the language used. Since the shell is both a command interpreter (runs other applications) and a programming language, it gets very complicated when it is working with another program that also uses wild cards or regex or simply general programming language constructs like parens or braces. I think that the only such program I use is ls. I don't do sed, awk, etc, for the same reason I don't program in bash or perl. Its too cryptic. There is no doubt that regex is very powerful but I just can't remember it long enough for any code that contains it to be meaningful and therefore maintainable. I'm not sure what you mean by '...the only such program I use is ls.'? The 'ls' does no regex or wild card processing, anything of this type is interpreted by the shell, first. For example, assuming no file name ending in 'x' is in the current directory: ls *x ls: *x: No such file or directory This assumes you're using a Bourne compatible shell. C shell error is from the shell (ls never gets run, even though the shell prints the name as part of the message): ls *x ls: No match. On the other hand, in my __Python_2.1_Bible__ is an example: rgrep.py
Re: exporting a variable to global shells
Jan-Florian Hilgenberg wrote: Hi guy's, first I am german, so ignore my bad english please ;-) i want to get a variable out of a child shell in it parent shell, the sense is, that I want to use the ProxyServer of my school automaticly if it is pingable, the script isn't hard but the variable isn't fully global after exporting it my script: #!/bin/bash ping -c 1 192.168.4.4 http://192.168.4.4 export http_proxy=192.168.4.4:8080 http://192.168.4.4:8080 echo $http_proxy #for debugging When the script is exiting the echo stdout's the proxy adress, but if I am back in my parent shell, the http_proxy variable is empty. The script should be run by the command post-up in the /etc/network/interfaces. ---deleted network stuff answered in other posts--- bye As others have noted, exporting variables is a one way process, children get what the parent sees, not the reverse. However, you can get data from a 'child' back to the parent in cases like this using what's called 'command line substitution' (big words to describe a fairly simple process ;) or by using functions which run in the parent context. First, using substitution: Assume your script file is called 'getproxyip', do these things: 1. remove the 'export' command but *leave* the variable in place. 2. leave the debugging line in place (it's no longer debugging but is the return value. 3. in the parent script, where you use your script, change it to be: HTTP_proxy=$(getproxyip) You can also use 'backticks' (accent grave), but they make things much harder to read: HTTP_proxy=`getproxyip` The advantage of using the backticks is that they are more 'portable', but I don't think that's an issue here. Second, using a function: This is written in the parent file, or is 'sourced' by the parent file. In either case, the function runs in the same process environment, so anything that is not 'local' to the function can be seen by the parent as well. function getproxyip { # your code here # Simple assignment works find, so your code drops in, no changes. } If you write this in a separate file, the main script *must* access it by 'sourcing' the getproxyip file: # setup function to determine proxy address: . getproxyip # The 'dot' command is also called 'source', so this also works: source getproxyip # Use one or the other. Using both won't break anything, just uses # extra resources. I hope this helps. Have fun!-) Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: hda: lost interrupt
Dallas Clement wrote: Hello All, I just installed the Debian Etch 4.0 on my system. It will not complete bootup however. It keeps hanging on hda: lost interrupt and eventually fails to boot entirely. I previously had Debian Etch Testing installed. It has been working fine for the past month up to the moment I tried installing 4.0. I tried re-installing 4.0 three times all with the same result. My system is an Intel Core2 Duo. I'm trying to install Debian Etch 4.0 AMD64. Does anyone know how to recover from this error? Thanks, Dallas My laptop Turion 64 requires using 'noapic' to boot and run. You can add this to the kernel 'command line' (at the boot prompt) for a one time test to see if it helps. If it does, edit your boot loader 'menu' to append this string to the entries for your kernel. You will need to read up on how to do these things for your particular boot loader. Other options you might try (some that I've heard about but not tried) are 'pci=no', 'acpi=no', 'acpi=no-idle' and 'nolapic'. I have no information on the possible down side of using any of these, but hey, if it lets you boot and things run, you're ahead of the game ;) I hope this helps. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: escape characters in sed
Hans du Plooy wrote: william pursell wrote: Your version is missing the final ', so I'm guessing you meant: sed 's/\\)/ /' Yes, sorry, several mistakes on my part... which will replace occurences of \) with a single space Not what I had intended, I just wanted to replace ) with a space. Actually I wanted to replace the ( with a space and get rid of the trailing ), So, you'd want to use: sed -e 's/(/ /' -e 's/)//' Which says replace left parens with a space, right one with nothing, with a single run of 'sed'. : eg: $ echo '2nd backslash and first paren replaced: \\))' | sed 's/\\)/ /' 2nd backslash and first paren replaced: \ ) Note that this is exactly the same as: $ echo '2nd backslash and first paren replaced: \\))' | sed s/\)/\ / 2nd backslash and first paren replaced: \ ) Thanks, now I kinda understand why. Hans Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: exporting a variable to global shells
cga2000 wrote: On Fri, May 11, 2007 at 12:33:12PM EDT, Bob McGowan wrote: snip 3. in the parent script, where you use your script, change it to be: HTTP_proxy=$(getproxyip) .. you can take it one tiny step further by using an array: . child: cz=($c0 $c1 $c2)/* .. $c4 .. etc.*/ echo [EMAIL PROTECTED] . parent: cz=($(child)) /* note added outer parentheses .. */ $p0=${cz[0]} $p1=${cz[1]} $p2=${cz[1]} .. I think you need a fairly recent version of bash to do this .. Dunno about other shells. Thanks, cga Actually, unless you need an array, this is overkill, if the objective is to get a series of single values in a series of simple variables (p0 ... p?). Assuming these assignments: x=a y=b z=c p=d echo [EMAIL PROTECTED] # where cz is (a b c d), from cz=($x $y $z $p) and echo $x $y $z $p generate the same results. And you can read data into a series of variables using the 'read' built in command: read x y z p The only issue being, if you don't have total control on the output of the command being read, and it generates more than 4 fields, the 4th to last field all get stuffed into p, so in that case you'd want to do: read x y z p rest So: $ echo a b c d | read x y z p $ echo $x $y $z $p a b c d And read has been part of Bourne shells for ages, so there's less of a backward compatibility problem. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: escape characters in sed
Alex Samad wrote: On Mon, May 14, 2007 at 08:59:49AM -0700, Bob McGowan wrote: Hans du Plooy wrote: william pursell wrote: Your version is missing the final ', so I'm guessing you meant: sed 's/\\)/ /' Yes, sorry, several mistakes on my part... which will replace occurences of \) with a single space Not what I had intended, I just wanted to replace ) with a space. Actually I wanted to replace the ( with a space and get rid of the trailing ), So, you'd want to use: sed -e 's/(/ /' -e 's/)//' any reason not to do sed -e 's/(/ /;s/)//' ---XXX--- No, except perhaps clarity. And, the fact I was not aware that you could do it that way. ;) Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: exporting a variable to global shells
Alex Samad wrote: On Mon, May 14, 2007 at 09:13:32AM -0700, Bob McGowan wrote: cga2000 wrote: On Fri, May 11, 2007 at 12:33:12PM EDT, Bob McGowan wrote: snip 3. in the parent script, where you use your script, change it to be: and more snipped So: $ echo a b c d | read x y z p $ echo $x $y $z $p a b c d what about something like set -- 1 2 3 4 5 6 7 for x do echo $x done change 1 2 3 4 5 6 7 for what every number of variables you want ? ---XXX--- Reminder: the original poster wanted to have data passed backward (up the process tree) from a child script to a parent. Collecting the output of a command using 'command substitution' is most easily handled as a single variable or line of data. You can make the method of passing the data out of the child as complex as you like, but in the calling script: lineIn=$(child ) puts one long string in $lineIn, including newline characters, spaces, tabs and so on. The result, using your 'set...loop', would put newline separated numbers into the returned string. Further processing would then need be done on the string to break it up into pieces, or not, as dictated by the needs of the caller. My 'child | read a b c d' is a quick way to read in the multiple lines output by your method, or to break a long line up on any white space characters. Without using the arrays suggested by another poster, which may not be supported by all versions of Bourne type shells. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Preventing delayed USB writes
Brendan wrote: On Tuesday 15 May 2007, Ron Johnson wrote: On 05/15/07 15:52, pedxing wrote: I would like to configure things so that, for instance, when I (ok, actually my wife) use konqueror to copy songs to my mp3 player, when the copy dialog says 100%, I can immediately unmount the device without having to wait for a delayed write. I have looked at the mount man page, but it says the sync option (which I think is what I want) is only available for ext2/3 and ufs file systems. [snip] I would rather not experiment on my wife's new mp3 player, but it would be a pain to force her to the command line every time she wants to transfer music to it... Does she just yank it out? IMNSHO, I would say to you: remember that Linux comes from a true multi-user heritage from back when 9-track (reel-to-reel) tapes ruled the data center, and were frequently dismounted and mounted. Thus, break the single-user mindset and proper unmount tapes when you are finished using them. In GNOME, when you right-click on a USB stick icon there is a choice to Unmount volume. I'm sure than KDE has something similar. You and your wife should just train yourselves to this new reality. Or, look into setting up syncSo that yanking something out doesn't do damage. I love that you advise Train yourselves to this new reality. Yeah, no. But, based on prior conversations in this thread, and web pages referenced, using the auto sync feature may be a bad thing, actually damaging the flash device beyond use. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: bash scripting question
Tyler Smith wrote: Hi, I've got a question about a short bash script I wrote. I need it to --snipped-- #!/bin/bash lab_num=41 for map_name in aest_90 bush_90 carol_90 comp_90 \ hirs_90 roan_90 swan_90 vir_90 ; do lab_let=$(echo -n $(printf \\x$(echo $lab_num))) echo $lab_let $map_name ${map_name}.ps ; echo $((lab_num++)) /dev/null ; done --snipped-- Some general comments, mostly aimed at making your code cleaner without changing what it does. First, both 'echo' and 'printf' put their results on standard out. Your call of 'printf' is inside command substitution, so its STDOUT becomes the command line for 'echo' which just prints to its STDOUT. Why the double print out? Just do: lab_let=$(printf \\x$(echo $lab_num)) Next, the 'echo $lab_num' is not needed, $lab_num can stand alone: lab_let=$(printf \\x$lab_num) And, the double quotes escape things, too, so the double backslash is not needed: lab_let=$(printf \x$lab_num) Then, the line where you increment lab_num can also be simpler. In bash the $((...)) alone on a line will replace itself with the result (command substitution, again). But, leave off the leading $ sign, and it just does the increment: ((lab_num++)) So, cut and pasted from a bash shell: $ lab_num=41 $ lab_let=$(printf \x$lab_num) $ echo $lab_let A $ ((lab_num++)) $ lab_let=$(printf \x$lab_num) $ echo $lab_let B Thanks, Tyler Since you're using bash, you may also find it convenient to put your hex digits into an array, which you can then subscript into with decimal numbers, to build the hex values needed to print other characters. This would need two loops, the outer to increment the 'tens' digit, the inner to increment the 'ones' digit, but it would do the trick. For example: x=(0 1 2 3 4 5 6 7 8 9 A B C D E F) tens=0 digits=0 while [ $tens -lt 3 ] do while [ $digits -lt 16 ] do echo ${x[$tens]}${x[$digits]} ((digits++)) done digits=0 ((tens++)) done The result is: 00 01 02 . . . 2D 2E 2F Change the 'tens' and 'digits' as needed to get the right starting value. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: bash scripting question
Tyler Smith wrote: On 2007-05-17, Bob McGowan [EMAIL PROTECTED] wrote: Some general comments, mostly aimed at making your code cleaner without changing what it does. First, both 'echo' and 'printf' put their results on standard out. Your call of 'printf' is inside command substitution, so its STDOUT becomes --snipped stuff-- Oh, great, thanks. I added the echo to stop getting the complaint about unknown command, but this is better. So, cut and pasted from a bash shell: $ lab_num=41 $ lab_let=$(printf \x$lab_num) $ echo $lab_let A $ ((lab_num++)) $ lab_let=$(printf \x$lab_num) $ echo $lab_let B Much improved! This would need two loops, the outer to increment the 'tens' digit, the inner to increment the 'ones' digit, but it would do the trick. For example: x=(0 1 2 3 4 5 6 7 8 9 A B C D E F) I knew there was an array form in bash, but I couldn't find it. I'm working from the O'Reilly book classic shell scripting, and the only reference to arrays is in relation to awk scripts. This is a big help. Thanks alot! Tyler You're welcome ;) Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Preventing delayed USB writes
Daniel Burrows wrote: On Thu, May 17, 2007 at 05:21:17PM -0700, pedxing [EMAIL PROTECTED] was heard to say: On May 17, 8:50 pm, Douglas Allan Tutty [EMAIL PROTECTED] wrote: To me, ideally, nothing would be written to the flash at all until either sync or umount. Yes, I wait. Yes, I also wait. Because KDE doesn't give the user any indication of how long to wait, the user has to drop to a terminal, run a sync command, and wait for the prompt to come back to get any indication of when it is ok to remove. You may want to take a look at Gnome. In Gnome, removable devices are handled like this: an icon appears on the desktop when the device is inserted. When you're done with the device, you click on the icon and select safely remove. A window pops up saying removing device... or somesuch, and when it goes away it's safe to remove your device (and the icon on the desktop is gone). I don't use Gnome much except as an xterm launcher, so I can't say more than that, but it looks like this aspect is pretty well-designed. I couldn't believe that KDE doesn't do something like this nowadays -- Gnome has been doing this for at least a couple years IIRC -- so I downloaded and installed KDE on my desktop to try it out. Apparently the only way to get an icon for your removable device is to click on the services tab (intuitively represented by ... a flag?), which shows Out of curiosity, just where does one find this services tab?? attached devices under storage media and lets you safely remove the device. A little green triangle disappears after the device is unmounted, which I didn't even notice until I'd mounted/unmounted it several times. So, the functionality is present in both UIs, but KDE has chosen a rather obtuse way of getting at it, IMO. Daniel Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: [OT] Best of UNIX/Linux Books that you can't stop reading
Deboo ^ wrote: Can eveyone list the best of UNIX/Linux books. In other words the books that doesn't let you leave it. My choices are: 1. Unix Power Tools (I think there's no second to this one and this is the King of all Unix/Linux books) 2. The Linux Cookbook: Tips and Techniques for Everyday Use by Michael Stutz Please add to the list. Regards, Deboo To better understand kernel related features (and also an old classic): The Design of the Unix Operating System Maurice J. Bach, Prentice-Hall 1986 smime.p7s Description: S/MIME Cryptographic Signature
Re: Mount a Windows partition
Wayne Topa wrote: Matthias Brennwald([EMAIL PROTECTED]) is reported to have said: Dear all I installed Debian Etch on a Dell D800 laptop. There's another partition with Windows (an some of my files) on it. How can I mount this partition? I tried the following so far: - sudo apt-get install ntfsprogs - sudo mkdir /mnt/windows - sudo ntfsmount /dev/hda1 /mnt/windows -o fmask=0111,dmask=0 fusermount: failed to open /dev/fuse: No such file or directory fuse_mount failed. Unmounting /dev/hda1 () I don;t mess with ntfs but a quick scan of the package lists cane up with the ntfsprogs package. It includes /usr/bin/ntfsmount and the depends maght help ie: libfuse2 (= 2.5), libntfs9 (= 1.13.1), fuse-utils ( 2.5.0) :-) HTH, YMMV, HAND :-) Wayne You may need/want to add the 'fuse-utils', as well. You then need to load the kernel module: modprobe fuse This is (I believe, based on searching in /lib/modules) part of the kernel package, so it doesn't need to be installed, but you do need to load it, and make arrangements for it to be loaded on each boot. Edit the file '/etc/modules' and add the word 'fuse' on a line by itself, at the end. Of course, the modprobe and edit must be done as root ;) Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: ask for suggestion: which filesystem suits for both Linux and Mac OSx
Marko Randjelovic wrote: Ron Johnson wrote: On 05/23/07 20:40, Ken Hu wrote: ¼ Wed2007-05-23 ¼ 21:25 -0400Douglas Allan Tutty Ð0 What filesystems can your Mac OSx read and write? Well , I think the filsystem my mac uses is HFS, but I can find no way to mount HFS on Linux. Of course my Mac can read cdrom or dvdrom, but what I need is to plug my usb external hard drive to my mac just as my original post said. The hfs hfs+ drivers might not (probably aren't??) built by default. You'd have to roll your own kernel. Not too difficult. They are built as module in Etch kernel (2.6.18-4). So your root partition cannot be hfs unless you recompile the kernel. Also, i am not sure about boot partition, because i don't know if grub supports it. If you need boot partition, i would try lilo. The OP was interested in support for an external USB hard disk, to be shared between a Mac and PC. So, probably not to be used for booting. The modules are hfs.ko and hfsplus.ko. If you want them available all the time, you can add 'hfs' and 'hfsplus' on separate lines in the /etc/modules file. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: adaptec sata installation
Guido Driesen wrote: dear..., How do i install debian on a system with a adaptec 1420sa controller and a maxtor 160 gig sta drive. The install programm freezez after a minute or so Thanks guido We'll need more info before being able to answer you. 1. What is your motherboard? 2. What is the CPU (AMD/Intel/???, 64 bit, 32 bit...)? 3. How much memory? 4. Which Debian (etch/stable, testing, old stable...)? 5. Anything else you think we need to know. -- Bob McGowan Symantec, Inc. smime.p7s Description: S/MIME Cryptographic Signature
Re: Sed advice needed
Piers Kittel wrote: Hello all, I need to delete some words out of a large file containing information about packets I'm analysing. I know I can use sed to do this, but haven't really used it before, so am a bit unsure of how to do it. Two example lines are as of below: 181,1324.014027,111.111.111.111,111.111.111.111,RTP,Payload type=ITU-T H.261, SSRC=2008229573, Seq=54520, Time=1725612773, Mark 185,1324.078941,111.111.111.111,111.111.111.111,RTP,Payload type=ITU-T H.261, SSRC=2008229573, Seq=54521, Time=1725616276 I need to convert the above to the below: 181,1324.014027,111.111.111.111,111.111.111.111,RTP,54520 185,1324.078941,111.111.111.111,111.111.111.111,RTP,54521 What's the best way to do this? I've been reading the man pages of sed, cut and awk but I can't quite figure out how to do this. Any ideas? Thanks very much for your time in advance! Regards - Piers --To UNSUBSCRIBE, email to [EMAIL PROTECTED] a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] To rephrase your question and data, so I'm sure I understand and you know what I'm doing: You want to print items 1, 2, 3, 4, and 5, below, as is, with only the Seq value from item 6, enclosed in double quotes. 1: 181, 2: 1324.014027, 3: 111.111.111.111, 4: 111.111.111.111, 5: RTP, 6: Payload type=ITU-T H.261, SSRC=2008229573, Seq=54520, Time=1725612773, Mark If my description is correct, this will do the trick, assuming that all the lines in the file have the same spacing and relative positioning: sed 's/Pay.*Seq=\(.*\), Time.*$/\1/' What I mean by the above assumption would be, for example, that a Seg number of 386 would look like Seq=386 rather than Seg= 386, or other formatting differences. The above sed substitution finds the part of the string starting with a double quote followed by Pay, then any characters repeated up to the portion with Seq=, followed by anything at all up to a comma space followed by the word Time followed by anything up to the end of the line. The portion between the \(...\) is remembered and is accessed in the substitute by the \1. So, this part of the line: PaySeq=#Mark is replaced with: # Where is the sequence number for each line. -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: slurping a file
Hugo Vanwoerkom wrote: Hugo Vanwoerkom wrote: Hi, I want to download an file that consists of many pieces that are referred to by the main index.html. One way is to do it manually, that takes days and is prone to errors. Is there a slurp application that walks through the file and downloads the pieces? --- Responses regarding wget and Perl with LWP --- Thanks guys!! I knew about those 2 ways too :-( Goes to show you what age does... Hugo If you don't mind working in a graphical mode using your browser, and if the browser is Firefox or a derivative, you could also use the 'Down Them All' (DTA) extension. Once installed, clicking in a non-link area of the page brings up a menu with DTA options. Selecting the 'DownThemAll!...' item opens a window listing all the links on the page. You can select some; select all and delete some; add or remove filters to modify items listed;, etc. Then click start and go get a cup of coffee (or your preferred beverage of choice) and enjoy a movie while it all gets downloaded ;) -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: hwclock /dev/rtc
Paul Wise wrote: Hi all, [Please CC me, I'm not subscribed] Can anyone tell me why hwclock --directisa works on this laptop (Dell Inspirion) but /dev/rtc does not (linux 2.6.21 from Debian sid)? # hwclock --directisa --show Fri Jun 1 15:36:11 2007 -0.377827 seconds # hwclock --show select() to /dev/rtc to wait for clock tick timed out # lspci | grep ISA 00:1f.0 ISA bridge: Intel Corporation 82801GBM (ICH7-M) LPC Interface Bridge (rev 01) # lspci -n | grep 00:1f.0 00:1f.0 0601: 8086:27b9 (rev 01) Is there a module udev isn't loading that it should be? Or do I need to write a driver for it? $ lsmod|grep rtc rtc12372 0 If you get no output, then the rtc driver has not loaded. But you don't need to write one, try: modprobe rtc first. Then, if hwclock works, you may need to add the driver name to the /etc/modules file. -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: printer setup
Andrew Sackville-West wrote: On Mon, Jun 04, 2007 at 05:04:23PM +, [EMAIL PROTECTED] wrote: Back again. I finally have etch gnome up and running again but am still having problems with my hp printer. I've tried using gnome desktop to setup the printer and everything seems to be ok until I print the test page and nothing happens. When I looki at the que iI see status: job stopped. so the job is there? just stopped? or is the printer stopped with no job in the queue? I've tried cups but it asks for a user name and password but does not accept mine and I have to cancel out of cups. I assume that since I'm not on line this option will nt work. this option has nothing to do with whether you are on line or not. Cups, on your local machine, provides a web interface for its configuration. That's why you can access it through localhost:631. localhost is the current machine you are sitting at. The user and password your need would be 'root' and root's password. Or you can add yourself to the lpadmin group useradd your-user-id lpadmin and then logout/login and then you will be able to modify the printer as yourself instead of as root. I've also tried /usr.sbin/lpadmin -p deskjet -vdirect usb://hp/deskjet%203900?serialx... -m deskjet.ppd. I get lpadmin: unknown argument usb://hp/deskjet hmmm... I've not tried this method, so can't speek to it, but if you could provide the verbatim output that would be great. it might be a quoting issue for the usb://... portion, try it with quotes around it. can you provide the contents of the cups logs (/var/log/cups/*) I've spent many hours reading info from gmone, cups and debian and still find myself knocking my head against the wall. anyone know what I can do? keep trying? Odds are its something really simple but we're losing it in the transmission here. rejoice in the fact that when its all done, you'll be a cups expert ;) A This is a shot in the dark, but I've seen this issue several times. If your printer configuration is setting the paper size to A4 and you're printer is is actually US letter size, the job will stop, waiting for the correct paper type. You'd need to use one of the CUPS configuration tools (I prefer the browser based method), select the printer configuration and verify the paper size there. -- Bob McGowan Symantec, Inc. smime.p7s Description: S/MIME Cryptographic Signature
Re: console too long after stopping x
Tyler Smith wrote: On 2007-06-05, Tyler Smith [EMAIL PROTECTED] wrote: Hi, I'm using fluxbox without a dm. When I exit from fluxbox and return to the console, the text scrolls off the bottom of the screen so that the active prompt is not visible. ... The only things that are now changed from default are a call to consolechars in .bash_profile: consolechars -f Uni3-Terminus14 As a follow-up, I determined that the problem can be fixed by re-running the consolechars command after I quit x. This happens every time I login, so that's why logging out and in fixes things. I know how to automatically run a command when I login. How do I automatically run a command when I stop X? Alternatively, is there something else I can do with my config files to fix this? Thanks, Tyler I can't answer the question regarding base causes, but to get a program to run after you quit X, add the command to the end of your .xinitrc file in your home directory. -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: hwclock /dev/rtc
Mumia W.. wrote: On 06/04/2007 06:12 PM, Paul Wise wrote: [please CC me, I'm not subscribed] On Mon, 2007-06-04 at 13:36 -0500, Mumia W.. wrote: $ lsmod|grep rtc rtc12372 0 [...] No I didn't write that. Bob McGowan wrote that. Please be more careful with the attributions. Apologies to Mumia and to Paul: I habitually removal of the original poster's address and only reply to the list. So, habit bit even though the explicit request was made to reply directly. -- Bob McGowan Symantec, Inc. smime.p7s Description: S/MIME Cryptographic Signature
Re: console too long after stopping x
Tyler Smith wrote: On 2007-06-05, Bob McGowan [EMAIL PROTECTED] wrote: I can't answer the question regarding base causes, but to get a program to run after you quit X, add the command to the end of your .xinitrc file in your home directory. I'm not sure how to do this. I didn't have an .xinitrc, so I copied the one from /etc/X11/xinit/xinitrc to my home directory, and added the line as follows: ++ #!/bin/sh ## comments snipped ## . /etc/X11/Xsession consolechars -f Uni3-Terminus14 ++ However, this doesn't seem to have changed anything - I logged out of x, back in, and then out again, and the problem remains. Thanks, Tyler I just checked my KDE environment and it doesn't use .xinitrc either. Sigh ;( I do know that .xinitrc is used by the base X11 environment, with the 'xinit' program, which can be used to start additional X servers on other virtual terminals. But, not for starting the first X server from a login manager (KDM, GDM, etc., also known as Display Managers). For KDE specifically, there is a program called 'startkde' which has a man page that mentions At the end of a session, the scripts found in ~/.kde/shutdown will be executed. I expect there's a way to have the same happen for a Gnome environment but I don't have any expertise with it. Sorry for the confusion. -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: partition table type
Johannes Wiedersich wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 jeffry s wrote: i am also wondering what all those partition type mean. can anyone point any documentation about the difference of amiga, bsd, dvh, gpt, mac, msdos, pc98, s390, sun, loop? It's just a different format of how the start of your hard disk is organized. I don't know about particular differences. If you can't find it on google, it is probably so obscure that you won't ever need it. Look at the article in wikipedia on partition types. You only need a different partition type than the default, if you have an other OS on the same hard disk that doesn't understand the default Is this the whole story? I would have thought that the base hardware would also have something to do with it. For example, the BSD OS versions designed to work on Intel architecture would need to have disk partitioning that the BIOS could work with, just as for msdos. And modern PC BIOS support for gpt is becoming common, too. But, a disk that is fully mac oriented, for example, could not just plug in to a PC system and boot. The system would recognize the disk as available, but would not be able to do anything with it, as the partitioning is totally foreign (though, with the Intel Macs, this may no longer be completely true). But when you say amiga or s390, I'd think you were explicitly loading an amiga or s390 version of an OS (which could be Linux) on Amiga or S390 hardware. The basic hardware and it's equivalent of a BIOS would then dictate the partition type to use. msdos. You should check the documentation of all your other OSes to find a partition type that is supported by all of them. It's good to know that debian also supports obscure stuff :-) Johannes -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) iD8DBQFGZlZSC1NzPRl9qEURAl6UAJwPZLa1+0ePigtrysD76vP1ECVEUACfTYDQ WcZ+eexljs/nOmD4pTs39Y4= =DfpO -END PGP SIGNATURE- -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: partition table type
Mike McCarty wrote: Bob McGowan wrote: Johannes Wiedersich wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 jeffry s wrote: i am also wondering what all those partition type mean. can anyone point any documentation about the difference of amiga, bsd, dvh, gpt, mac, msdos, pc98, s390, sun, loop? It's just a different format of how the start of your hard disk is organized. I don't know about particular differences. If you can't find it on google, it is probably so obscure that you won't ever need it. Look at the article in wikipedia on partition types. You only need a different partition type than the default, if you have an other OS on the same hard disk that doesn't understand the default Is this the whole story? I would have thought that the base hardware would also have something to do with it. What do you mean by base hardware? For example, the BSD OS versions designed to work on Intel architecture would need to have disk partitioning that the BIOS could work with, just as for msdos. And modern PC BIOS support for gpt is becoming common, too. The BIOS has nothing to do with disc partitions. The BIOS works either at an absolute physical level using geometric address: cylinder, head, sector, or at a logical physical level (LBA) using logical sector number. It has no other conception of the disc. But, a disk that is fully mac oriented, for example, could not just plug in to a PC system and boot. The system would recognize the disk as Of course not, if by fully MAC oriented you mean having a PPC on it. available, but would not be able to do anything with it, as the partitioning is totally foreign (though, with the Intel Macs, this may no longer be completely true). The partitioning doesn't enter into it. But when you say amiga or s390, I'd think you were explicitly loading an amiga or s390 version of an OS (which could be Linux) on Amiga or S390 hardware. The basic hardware and it's equivalent of a BIOS would then dictate the partition type to use. I don't know how boot load works with those machines. Here's how it works on PC architecture machine The processor comes out of reset in real mode, and sets the IP to point to the reset location (I forget the address, but perhaps :8000). The BIOS is now in control, running from ROM. It does basic tests of functionality and preliminary initialization of hardware. It may switch to priviledged mode type operation. It enumerates the PCI bus, and if it finds bridges, initializes them and enumerates sub PCI busses. If it finds peripherals it recognizes, like USB I/F or hub, it initializes them. Then it looks for things like video cards, keyboards, etc. on the back side. If all looks good, it beeps and then starts to look for something else to boot. If it switched away from real mode, it switches back. The BIOS now looks for bootable devices, using some order for device type and device ID. Typically, it looks for CDROMs, floppies, and hard discs, in that order. Possibly for USB sticks. I am familiar with how floppy and hard disc boot works, and how some CDROM boot works. For a floppy or hard disc, and for some CDROMs, the first absolute sector is loaded into memory at location 07C0:. If the last two bytes are the special marker AA55, then this indicates bootable device. If not, then the device is skipped. If the marker is present, then the processor jumps to the first byte of the loaded sector, which then assumes control of boot. For a floppy, and for certain CDROMs which just have an image of a bootable floppy on them, this is the BR. For a hard disc, this is an MBR, and the code in the MBR must find an appropriate partition. For a hard disc, the code in the MBR scans the portion of the MBR which contains the PT looking for partitions (also called volumes, not to be confused with LVM) which have the active or bootable bit set in them. If none is found, or if more than one are found, then an error has occurred. Else, the code in the MBR has found what purports to be a bootable partition. It loads the first sector from that partition into memory at location 07C0:, and looks at the last two bytes. If they are AA55, then it presumes it has found the beginning of a volume, which contains exactly the same information as the first sector of a floppy. IOW, it has found a BR, and we are in the same situation as with a floppy and certain CDROMs. At this point, either the BIOS or the MBR code, whichever has control, has loaded a BR, and jumps to the first instruction. The BR has in it a thing called the BPB, which describes the geometry of the floppy disc, floppy disc image on CDROM, or of the partition of the hard disc. The code in there looks for a file system, tries to find the appropriate boot images, and loads, initializes, and starts that boot image. What happens during this is OS dependent. For MSDOS, for example, MSDOS.SYS (actual OS) and IO.SYS (BIOS extensions) get loaded and COMMAND.COM gets run
Re: problem
walter wrote: On Wed, 2007-06-06 at 10:06 +0200, walter wrote: I've lost a lot of time and plastic trying to download your dvds, cds. All corrupt. It seems you or somebody else don't want people know about your distro. What were the problems you experienced ? Regards, Bart Martens I can download ubuntu livecd, gentoo livecd, etc. with `direct download from ftp' without problems in a couple of hours. I can do the same with torrent. I know how to burn a cd. With your distro it's imposible for me do the same, direct, torrent, jigdo or whatever I use. I've tried downloading your dvds with emule, I've downloaded with emule and burned hundred of iso images without problem. The two times the debian dvds had corrupted packets (six dvd in total). And the md5sum say OK! This is my experience and I have not need of lie you. I am a bit tired that each time I write to a linux developers they treat me like an idiot (That's why I became ironic). I repeat I know how download and burn a cd. If you don't know where's the problem, your problem, it is difficult I do. And it would be a good new for you that somebody complains about can't download your distro. After all I am giving a favor telling you about this. I've downloaded both CD and DVD images from the Debian servers at various times and have no problems whatsoever burning or using them. However, I now prefer to use the netinst image and download only the packages actually needed or wanted. Have you tried installing that way? If the MD5 sum is OK, I would assume the problem is not with the download or the server, it's at your end, somehow. Though it is possible that the original has corruption and that the MD5 sum is based on that corrupted image and so reports a good download. However, if this is true, you would not be the only one (that I know of) making this complaint. By the way, this particular mailing list has a lot of people on it who may well be developers, but don't work on Debian. We are users, as the list name suggests. -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: dd_rescue works to copy an NTFS ...
Thiago Santos Faria Xavier Teixeira wrote: Hi, does anyone knows if dd_rescue works to rescue an NTFS partition? Thanks in advance, -- --- Thiago Santos Faria Xavier Teixeira --- Based on the description, which indicates it works like 'dd', except it will retry on failure, skipping failed sectors, the answer is 'yes'. This is because dd and, by extension, dd_rescue, do byte for byte copies of a 'raw' device. This means there are no dependencies on filesystem types, data content, partition type, etc. In fact, you could use either of these on the whole device (hda rather than hda#) and it would copy MBR, partition table, and partitions (regardless of FS type) faithfully. You could then put this on a disk of equal or greater size, being aware that if the destination is larger, you *will* lose that extra space because the dd copies the original disk geometry info to the new disk. But, qualify this with, if it skips a sector, you have no idea if the sector is used or not, if it is used, if it's in a system critical file or not, and so on. You will have to run the appropriate filesystem repair utility (fsck, checkdisk, ...) to try to fix things so files at least are readable. But the data in those missing sectors is lost, period. Caveat emptor ;) -- Bob McGowan Symantec, Inc. smime.p7s Description: S/MIME Cryptographic Signature
Apache2 and PHP5 configuration problem
I've got both these installed on my Debian etch system. I've enable the 'server-info' and 'server-status' URL's and based on the info in them, the mod_php5 module and configuration are loading. I have Apache set up so I can access it without problems, so far as HTML files are concerned (/var/www/apache2-default/index.html and /var/www/index.html both serve correctly). However, when I try to use a .php file, a file download dialog pops up. So, I'm clearly missing something in the Apache2 configuration. Can someone point me the right way on this? Thanks, -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: Apache2 and PHP5 configuration problem: SOLVED
Atis wrote: On 6/7/07, Bob McGowan [EMAIL PROTECTED] wrote: I've got both these installed on my Debian etch system. I've enable the 'server-info' and 'server-status' URL's and based on the info in them, the mod_php5 module and configuration are loading. I have Apache set up so I can access it without problems, so far as HTML files are concerned (/var/www/apache2-default/index.html and /var/www/index.html both serve correctly). However, when I try to use a .php file, a file download dialog pops up. So, I'm clearly missing something in the Apache2 configuration. Can someone point me the right way on this? #a2enmod php5 Regards, Atis As noted in my first post, the Apache info from 'server-info' tells me: Module Name: mod_php5.c Content handlers: yes Configuration Phase Participation: Create Directory Config, Merge Directory Configs Request Phase Participation: Content Handlers Module Directives: php_value - PHP Value Modifier php_flag - PHP Flag Modifier php_admin_value - PHP Value Modifier (Admin) php_admin_flag - PHP Flag Modifier (Admin) PHPINIDir - Directory containing the php.ini file Current Configuration: -- Which I thought implied that all the links for enabling modules were correct. But, I had done them manually, so just to be sure I removed the two (php5.conf and php5.load) and used 'a2enmod php5' as suggested. No joy. I'm still prompted to download the file rather than having the server execute it. Am I missing some other Apache directive needed to allow access to or execution of PHP scripts? -- After some additional poking around and talking with one of my co-workers: My document root is /var/www, with the default RedirectMatch that lets apache2-default work. Turns out this was actually the problem. Commenting out the redirect and putting all the files in the document root that needed to be there, fixed the problem. This probably means there were missing directives for the setup I had before, but that's moot, for now. -- Bob McGowan Symantec, Inc. smime.p7s Description: S/MIME Cryptographic Signature
Re: shell or fileutils commands and whitespaced filenames
Michael wrote: thx Bob. (Is there any list you're not subscribed :) Using 'null' is slightly shorter than my 'sed' command...but somehow i can't help it's still not really straight enough. Doing complicated things just to have quite common formatted filenames handled, is so *uncool* for a modern shell ! Could you supply a reference for this modern shell you're referring to? I don't know of any shell (UNIX, Linux or MS Windows) that doesn't require some form of quoting in order to handle white space in file names. Thanks. Bob If there's really no other way, then do you know of a different shell can maintain whitespace names just automatically ? I'd even move to csh if i knew it works better there. --micha [EMAIL PROTECTED] (Bob Proulx): Try using find and xargs. Here are some examples. find . -print0 | xargs -r0 chmod a+rX smime.p7s Description: S/MIME Cryptographic Signature
Re: Mr. Mad Duck: about etch-stable fluxbox?
Douglas Allan Tutty wrote: On Mon, Jun 11, 2007 at 09:10:59AM -0500, Gayle Lee Fairless wrote: I looked on the maintainer webpage which seems to indicate that an IBM Thinkpad 390 Pentium MMX (233 MHZ) needs to have a patch installed for fluxbox to work. Since an MMX is somewhat slow, I was hoping to consider fluxbox or something like it instead of the heavy-duty GNOME or KDE. BTW, I do have a copy of Martin Kraft's book just in case I missed something in it. Please tell me. Thanks for your advice. I enjoy using Debian from time to time on my other systems. You could also try icewm. I run it on my PII-233 with 64 MB ram desktop. Who is Martin Kraft? Doug. I think that would be: *Martin Krafft* has been a faithful supporter of Debian since 1997, working as a developer and a PR person, and fielding user questions on mailing lists. He has experience administering mid-sized networks and providing user support, and is responsible for numerous university servers and a 40-node cluster of Debian machines. Krafft is currently working on his Ph.D. at the Artificial Intelligence Laboratory of the University of Zurich. From O'Reilly web page: http://www.oreillynet.com/pub/au/2325 Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: problem: nvidia module not loading after updating Xorg
arijit wrote: Hi all, I've just updated Xorg packages though synaptic. But after system restart, i am unable to start X-server. Error shown : unable to load module: nvidia right now I've switched to 'vesa' but I want to use 'nvidia'. Also when i use nvidia-settings, i get this error: - $ nvidia-settings ERROR: NV-CONTROL extension not found on this Display. ERROR: Unable to determine number of NVIDIA GPUs on ':0.0'. ERROR: Unable to determine number of NVIDIA Frame Lock Devices on ':0.0'. - Here's the some output from Xorg.log:- - II) LoadModule: nvidia (WW) Warning, couldn't open module nvidia (II) UnloadModule: nvidia (EE) Failed to load module nvidia (module does not exist, 0) - How can i solve this? I was using nvidia-glx, NVIDIA binary XFree86 4.x driver version: 1.0.8766-4 -- /Arijit Sarkar/ /Kolkata, India/ Evolution on Debian testing Did you also upgrade the kernel? If so, you need to re-run the nvidia installation script to either install a pre-compiled, or to compile and install a custom, Nvidia driver. -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: /lib/init/rw
Hans du Plooy wrote: Hi guys, What does this do? jduplooy:~# df -lh FilesystemSize Used Avail Use% Mounted on /dev/hda2 52G 42G 7.7G 85% / tmpfs 503M 0 503M 0% /lib/init/rw--- ?? jduplooy:~# ls -lha /lib/init/rw/ total 4.0K drwxr-xr-x 2 root root 60 2007-06-05 19:03 . drwxr-xr-x 3 root root 4.0K 2007-02-07 18:36 .. -rw-r--r-- 1 root root0 2007-06-05 19:03 .ramfs Unmounting it seems to have no effect. I'm running Etch on several machines: desktop and server, even mix of AMD64, 486, 686 and K7 stock kernels. Thanks Hans On my system, /lib/init/rw also contains a hidden directory: drwxr-xr-x 2 root root 80 2007-06-04 16:31 .mdadm Which contains two empty files: -rw-r--r-- 1 root root 0 2007-06-04 16:31 md0-uevent -rw-r--r-- 1 root root 0 2007-06-04 16:31 md1-uevent This matches my software RAID setup. Based on the content of two *.sh files located in /init/lib, I'd think this is a system startup directory, used as part of the system's initialization (initrd, etc.). So, maybe it's OK to umount it, presuming that it's no longer used. Then again, since I really don't know, perhaps not. ;) -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: Cursor problems
Todd A. Jacobs wrote: Periodically, my mouse cursor will turn black and/or pixelated, and I haven't been able to find a reason for this or a fix other than to restart X. It seems to happen most often when browsing with firefox under KDE, but I see this behavior more often (and more randomly) under Gnome. So, two questions: 1. Does anyone have any idea what's corrupting the cursor? 2. Is there a way to fix it short of CTRL-ALT-BS? This isn't a new problem for me, and has been a problem with Etch since before it went stable, and continues to be a problem for me under Lenny. Suggestions welcome. I have an old Sony PCG-XG laptop that had major problems with the cursor under X. In this case, I was able to solve the problem by using an Option line in the Device section of xorg.conf, for the video card. The card is a Neomagic MagicMedia 256XL+, so this option may not work for you, or you may have to search out a similar action but different name. OptionSWCursor true This forces my card to always use a software cursor, even if the driver thinks there's a hardware one available. I hope this at least helps get you looking in the right direction;) -- Bob McGowan Symantec, Inc. smime.p7s Description: S/MIME Cryptographic Signature
konsole from script errors
I've got a simple script to start KDE konsole with a specific geometry: /usr/bin/konsole --geometry $coordinates The value for 'coordinates' is either passed in via an argument or set set to a default value: 671x504+0-0 I just used this script to start a konsole session from a session started using Panel-Terminal Sessions-Shell. With the ampersand backgrounding the konsole process, the invoking shell exits and the konsole process attaches to init: bob 2586 1 0 08:24 ... /usr/bin/konsole --geometry 671x504+0-0 This works OK but after the initial execution only, and in the konsole window from which I ran the script, I got these errors, once, so far: konsole: WARNING: Unable to use /usr/share/apps/konsole/screen.desktop konsole: WARNING: Unable to use /usr/share/apps/konsole/mc.desktop konsole: WARNING: Unable to use /usr/share/apps/konsole/sumc.desktop The files exist, are readable: -rw-r--r-- 1 root root 3876 2006-10-01 10:32 mc.desktop -rw-r--r-- 1 root root 4737 2006-10-01 10:32 screen.desktop -rw-r--r-- 1 root root 6427 2006-10-01 10:32 sumc.desktop and have content: screen.desktop: UTF-8 Unicode text mc.desktop: UTF-8 Unicode text sumc.desktop: UTF-8 Unicode text Any ideas why they can't be used? Bob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Detecting disk drives from an installation script
Orestes leal wrote: On Wed, 20 Jun 2007 16:29:12 +0200 Michelle Konzack [EMAIL PROTECTED] wrote: Am 2007-06-14 19:00:47, schrieb Orestes leal: $fdisk -l | grep dev | gawk -F' *' '{ print $1 }' | gawk -F'/dev/' '{ print $0 }' | grep '/dev/sd' particiones.txt Stripped echo -e ${FDISKOUT} | some_command ... ^^ Do not forget it. :-) Why is '-e' required? The -e tells 'echo' to -e enable interpretation of backslash escapes, but there are no backslash escapes in the output of the 'fdisk -l' to worry about. What's important is that you have the double quotes around the variable expansion, which preserves all white space characters 'as is' (or if you prefer, as they were in the original fdisk output - tabs, spaces and newlines). The entire content of the variable is passed to echo as a single (very large;) argument with embedded white space characters intact. Without the double quotes, echo sees multiple arguments. The arguments were created by the shell, using the white space separators to break them up. echo never sees any white space, so it it prints each argument separated from the next by a single space, which puts everything together into one long, unformatted string. Thanks. Thanks, Greetings and nice Day Ore. Orestes [EMAIL PROTECTED] -- Bob McGowan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: backgrounding a script
MiKe McClain wrote: Is there a way for a script to background itself? Thanks, Mike Yes. But you gotta be careful, it's very easy to get a run away process series. The basic idea is that the script has to rerun itself in the background and exit. The way I do this is to set a flag in the first pass and *export* it, so the second pass can skip the first pass code. Like this: #!/bin/bash # Got to be careful, we don't want to go into an infinite loop of scripts. if [ ! $DoNotBGMe ] then # Set this so we don't try to rerun this script more than one time. export DoNotBGMe=yes # Background self. First, get full path to self. This usage of 'type' # may be specific to the bash shell. me=$(type -p bgself) echo Before backgrounding: $$ $me exit fi echo After backgrounding: $$ echo I'm running! # End of script You can remove the 3 echo commands, they're there for illustration only. The code you want to have actually run goes after the 'if' statement. Bob McGowan PS: Forgot to mention, you may also not need to do the 'me=$(type...)' part, I do it so I can verify I'm actually running the script I think I'm running. Problems of this type can happen if you explicitly set the PATH being used by script (something you should do if the script is to become a production level tool, used by many users). bob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: /etc is gone
Karl E. Jorgensen wrote: On Mon, Jun 18, 2007 at 08:54:30PM -0400, j j wrote: Hello I have been have trouble with my debian box(64 studio). i just discovered that /etc directory is missing. I dont think i deleted it, but it seems that I must have. Perhaps not. Things to look at: - you haven't mounted something on /etc, have you? Although insanely stupid, pls don't be offended. When weird things happen, it's time to be insanely paranoid... (/proc/mounts should tell) - You haven't chroot'ed yourself, have you? (I know. Paranoia...) - Anything in the system logs? (things like re-mounting read-only or filesystem corruption would be very interesting...) I am surprised that you don't mention any other problems - if /etc/ was really really gone, I'd expect loads of other problems. Hence the somewhat paranoid checks... Not least of which would be no password file, hence no way to log in. Plus, no rc scripts to bring the system up, either. So, I'd suggest the system has booted into some sort of maintencance mode. Or, you're simply in maintenance mode for some other reason, running from some sort of RAM disk? More info is needed, as printed during the startup, particularly just before the point where you get a shell prompt. Hope this helps -- Bob McGowan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: setting in /etc/profile in debian
Liam, I think you have a good point. So how do I change its default if I want the file to be sourced when I login using GDM. Thanks Yong --- Liam O'Toole [EMAIL PROTECTED] wrote: On Wed, 20 Jun 2007 17:55:37 -0600 [EMAIL PROTECTED] (Bob Proulx) wrote: yong lee wrote: [...] I have to manually do source /etc/profile to see the new changes after I open a new terminal window everytime. Even a reboot does not help. Does anyone have any idea ? You say that even a reboot does not help and that seems very strange to me because I am sure that you would log in after that point. :-) Are you using a shell other than a POSIX compatible one such as csh? [...] If the OP logs in using GDM, then (by default) /etc/profile is not sourced. -- Liam -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] As I don't use GDM or gnome, I can't tell you exactly where to do this. But, you can add to the terminal command in the menu (and anywhere else you want) the '-l' option (at least, you can in KDE, I think the same should work in Gnome), which tells the terminal program to run as if it were a login shell. It will then read the startup scripts in /etc, as you want. One thing not mentioned by others is which files from /etc a shell will use depends on the shell. The /etc/profile file is read by Bourne shells, such as bash or ksh. If you are using csh or tcsh, it will *not* read /etc/profile, it will instead read /etc/csh.cshrc (read by every shell) and /etc/csh.login (read by login shell only). Note on some systmes these csh startup files are called /etc/cshrc and /etc/login, respectively. If you're using csh/tcsh, changes to /etc/profile wiil do nothing. See the man pages, section FILES, for a full list of rc files for each shell. Bob McGowan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Turning off the %^* scroll wheel
Curt Howland wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Friday 22 June 2007, Jonathan Kaye [EMAIL PROTECTED] was heard to say: The mouse/touchpad is controlled in /etc/X11/xorg.conf. Look for a Section InputDevice and if you see something like this: Option ZAxisMapping 4 5 then comment it out. That should get rid of the scrollwheel behaviour. Mr. Kaye, many thanks. Unfortunately, there is no such option in the two mouse related sections: == Section InputDevice Identifier Configured Mouse Driver mouse Option CorePointer Option Device/dev/input/mice Option Protocol ExplorerPS/2 Option Emulate3Buttons true EndSection Section InputDevice Identifier Synaptics Touchpad Driver synaptics Option SendCoreEventstrue Option Device/dev/psaux Option Protocol auto-dev Option HorizScrollDelta 0 EndSection = I tried reconfiguring and changing to imps rather than explorerps2, no change. Is there another possibility? --snipped-- Which of these two is actually used in the 'ServerLayout' section? Whichever it is, you could try switching it to the other and see what happens. Or if both are there (as they are for my laptop) you could try commenting one out to see which is actually defining the wheel. Then edit the conf file to change the 'Option Protocol' for that device to be some other descriptor. Another thought, based on output from 'gpm -t help', the imps name is not listed, but imps2 is. But it seems to include some sort of wheel functionality. So you might try a very basic interface, such as 'ps2' or 'fups2' (described as being used for 'broken' PS/2 mice). Though the names are from gpm they should match names used by X11, since the two have to (and do in my experience) work together. -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: ssh on lan by hostname instead of dynamic ip
Chris Lale wrote: BartlebyScrivener wrote: Hello, I use unison and rsync on my home LAN. Everything works fine as long as I use ip addresses, for example, ssh [EMAIL PROTECTED] But my router (on home LAN) assigns dynamic ip addresses. So my scripts don't work if the router happens to assign a different address that day to one of my machines. Is there a way to use ssh and rsync using hostname instead of ip address. It seems to me I've had this working once or twice, but then it stops. Not sure if dnsmasq[1] might help - it can serve the names of local machines which are not in the global DNS. I've looked at it but never actually used it myself. [1] http://www.thekelleys.org.uk/dnsmasq/doc.html If you'd rather not set up dnsmasq or need something quicker ;) you may be able to set the router to assign the same IP to the system using the MAC address. I have a DLink that allows this. You then put the IP and host name in your /etc/hosts file. From what I can see in the man page for dnsmasq, it would basically do the same thing as the above, it just puts the configuration onto your local server rather than the router (my guess;-) -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
Re: /etc is gone
j j wrote: where do find all that is printed before i get a shell prompt? /var/log/? jj Probably not, if you're not getting a normal boot. In other words, if there are problems with a disk drive preventing the file system where /etc is from mounting (that's almost always /), then wherever /var/log is may also be having problems. Besides, without / and /etc/fstab, how would the system be able to figure out where to find /var/log? You'll just need to copy down on paper what you see on the screen. You can also try typing 'mount' and a carriage return to see what it says is mounted. Of course, this all assumes the system is having major problems early on in the boot process. You have not provided, as yet, enough information to figure out where the system stands in the normal sequence. Anything printed on your monitor and still visible could help. On 6/21/07, *Bob McGowan* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: Karl E. Jorgensen wrote: On Mon, Jun 18, 2007 at 08:54:30PM -0400, j j wrote: Hello I have been have trouble with my debian box(64 studio). i just discovered that /etc directory is missing. I dont think i deleted it, but it seems that I must have. Perhaps not. Things to look at: - you haven't mounted something on /etc, have you? Although insanely stupid, pls don't be offended. When weird things happen, it's time to be insanely paranoid... (/proc/mounts should tell) - You haven't chroot'ed yourself, have you? (I know. Paranoia...) - Anything in the system logs? (things like re-mounting read-only or filesystem corruption would be very interesting...) I am surprised that you don't mention any other problems - if /etc/ was really really gone, I'd expect loads of other problems. Hence the somewhat paranoid checks... Not least of which would be no password file, hence no way to log in. Plus, no rc scripts to bring the system up, either. So, I'd suggest the system has booted into some sort of maintencance mode. Or, you're simply in maintenance mode for some other reason, running from some sort of RAM disk? More info is needed, as printed during the startup, particularly just before the point where you get a shell prompt. Hope this helps -- Bob McGowan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] -- Bob McGowan smime.p7s Description: S/MIME Cryptographic Signature
List Archive problem?? (Was: Re: Can't boot after installation == ALERT! /dev/sda1 doest not exist. Dropping to a shell.)
debian wrote: Hello, I searched the archives, but some files are not found. Anyway, i found some others where you explained some things. I could boot with the linux rescue cd and mount my /dev/sda1 I changed the bootlogd to YES but nothing is ever logged because the system can never connect to SDA1, so how could it ever write something to it ? You also say in a message with such a fault that probably the right modules are not loaded? But how is that possible if the installation recognized the disk and wrote all the files and everything and succesfully completed the installation ? Thanks. Verus. -Original Message- From: Bob McGowan Sent: dinsdag 24 oktober 2006 17:52 To: debian-user@lists.debian.org Subject: Re: Can't boot after installation debian wrote: Hello, i have a new poweredge 1950 with an embedded dell perc 5/i raid controller There are 2 disks in it of 72GB and are configured as RAID 1. I perform a new installation with the debian testing netcd. Everything goes well, netcard is found, DELL Perc is also found. When partioning the HDD in the installation menu, these are the details: SCSI3 (2,0,0) (SDA) -72.7 GB DELL PERC 5/i i partition the harddisk, continue the installation and install GRUB at the master boot record. I reboot after the installation, the GRUB menu is shown and the system boots. Then after a while i get: ALERT!: /DEV/SDA1 does not exist. Dropping to a shell BusyBox v1.1.3 (Debian 1:1.1.3-2) Built-in shell (ash) enter 'help' for a list of built-in commands. /bin/sh: can't access tty; job control turned off. I also tried in raid 0 without the second disk but it remains the same. It seems that the installation sees SDA1 but it is nog recognized after installation. Can anybody help with this problem please ? thnx, Verus. Verus, Do you have any ATA hard disks or USB disk devices (including memory drives) attached to the system? If so, I think you're seeing a difference in how the BIOS scans disk devices, versus the kernel. If this is the case, I've made several posts on this subject in the past that provide details on how to set up LABEL based booting/mounting of the system. Search the archives for any posts by me, one with the subject Recent upgrade causes drive lettering scheme to alternate from hda to hde and two with the subject Debian AMD64 boots only at random: how to use labels/fstab/grub Bob Disclaimer : This e-mail is intended for the exclusive use by the person(s) mentioned as recipient(s). If you are not the intended recipient, please notify us immediately. This e-mail does not constitute any commitment for Sea-Ro Terminal N.V. or its subsidiaries except when expressly otherwise agreed in a written agreement between the intended recipient and the originating subsidiaries of Sea-Ro Terminal N.V. Sea-Ro Terminal and each of its subsidiaries each reserve the right to monitor all e-mail communications through its networks. This mail has been checked for viruses by Symantec and Trend Micro. I've replied with copies of the relevant posts directly to Versus. His problem finding them raises a question. I just searched on http://lists.debian.org/search.html for AMD64, the October-December date, and got 13 hits. None had any content, returning the error: Not Found The requested URL //srv/-srv07/msg00116.html was not found on this server. There were 20 hits for AMD64 in the July-September quarter. Again, all 20 links generate the above error. And, when I searched for 'Debian amd64 boots', a subset of the subject of one of the messages I referred Verus to, it did not find those posts at all, in either of the two quarters in which I sent them, though it did return the response I made to Verus question in which I referred to the posts. I don't expect that this is normal behavior, but then, I'm also not familiar with the code that maintains the list archives, so maybe there are delays or issues I'm not aware of. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Can't boot after installation == ALERT! /dev/sda1 doest not exist. Dropping to a shell.
debian wrote: Hello, Thanks for the information, I tried some things and I hope I am little bit closer to the solution. This is what I have done: Rebooted with a rescue cd and changed the label of /dev/sda1 to /root Jfs_tune -L /dev/sda1 /root When I check my settings with jfstune: Jfs_tune -l /dev/sda1 ... ... Volume Label: '/root' With cfdisk: with cfdisk SDA1BOOTPrimary Linux JFS [/root] SDA2Primary Linux Swap / Solaris So I guess this is correct. (I also changed label of swap partition but appearantly that's not shown, weird) Then I mounted my disk and changed /etc/fstab: fstab: #/dev/sda1 / jfs defaults,errors=remount-ro 0 1 LABEL=/root / jfs defaults,errors=remount-ro 0 1 #/dev/sda2 noneswapsw 0 0 LABEL=swap noneswapsw 0 0 I changed /dev/sda1 with LABEL=/root Then I changed menu.lst from grub: title blahblahblah root(hd0,0) kernel /boot/vmlinuz-2.6.16-2-686 root=LABEL=/root ro initrd /boot/initrd.img-2.6.16-2-686 savedefault boot I saved everything and rebooted the system. Then when I boot, I get this: ALERT! /dev/disk/by-label//root does not exist. Dropping to a shell! BusyBox v1.1.3 (Debian 1:1.1.3-2) Built-in shell (ash) Enter 'help' for a list of built-in commands. /bin/sh: can't access tty; job control turned off (initramfs) So, appearantly the system can't find /root label or can't link it with /dev/sda1 ? Maybe because /root is not in the /dev directory (according the error). I tried to make a link (ln -S /dev/sda1 .//root But error stays the same. So maybe the /root is wrong ? I change the label to test Jfs_tune -L /dev/sda1 test Cfdisk shows me 'test' for label I changed fstab and menu.lst and made a link in /dev/ From test to sda1 (don't know if this is necessary but just trying...) Result is the same error, not working. What do I do wrong in this label/boot thing ? Grtz, (and thnx for helping me out!) Verus. -Original Message- From: Bob McGowan [mailto:[EMAIL PROTECTED] Sent: donderdag 26 oktober 2006 17:53 To: debian Subject: RE: Can't boot after installation == ALERT! /dev/sda1 doest not exist. Dropping to a shell. Versus, My intent in sending you to the archives was to save bandwidth in the newsgroup. I did not know about the 'missing' messages, a problem which seems to reflect a time lag in getting the data linked in. So, again in the interest of reducing bandwidth, and repetitive posts, I'm sending the details to you directly. I hope the attached information is helpful. Bob Disclaimer : This e-mail is intended for the exclusive use by the person(s) mentioned as recipient(s). If you are not the intended recipient, please notify us immediately. This e-mail does not constitute any commitment for Sea-Ro Terminal N.V. or its subsidiaries except when expressly otherwise agreed in a written agreement between the intended recipient and the originating subsidiaries of Sea-Ro Terminal N.V. Sea-Ro Terminal and each of its subsidiaries each reserve the right to monitor all e-mail communications through its networks. This mail has been checked for viruses by Symantec and Trend Micro. OK, this should have worked. So I need some context, which is no longer part of the message. Which Debian system are you running (stable, testing, unstable)? I'm using testing, with udev support. It may be that my fixes are specific to the release I'm using. I'm not knowledgeable in the workings of udev, so other than saying you may need it for this to work, I cannot provide any further assistance with your issue ;( and will have to defer to others on the list. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: Good reader for /usr/share/doc/ and/or man pages ?
HXC wrote: That is a very extensive list, thanks a lot! I especially like the konqueror option :-). Do you happen to know a way to view the Debian specific readme's in /usr/share/doc/? On Friday 27 October 2006 17:05, Kamaraju Kusumanchi wrote: On Friday 27 October 2006 05:40, HXC wrote: I am searching for a reader that 'automatically' reads /usr/share/doc/ and/or man pages. It would be especially great it such a program would list the available (Debian) readme's available in the /usr/share/doc directory. Anyone knows such a program? Various ways of reading man pages, documentation man pagename vim, :Man pagename emacs, M-x man konqueror, man:pagename info pagename pinfo pagename install dwww, and go to http://localhost/dwww/ in any browser install man2html and enter the address http://localhost/cgi-bin/man/man2html in any web browser. khelpcenter yelp and finally dont forget google :-) My favorites so far has been konqueror, dwww YMMV. Please let me know if I have missed anything... raju -- Kamaraju S Kusumanchi http://www.people.cornell.edu/pages/kk288/ http://malayamaarutham.blogspot.com/ The description for dwww reads: Read all on-line documentation with a WWW browser All installed on-line documentation will be served via a local HTTP server. When possible, dwww converts the documentation to HTML. You need to install both a CGI-capable HTTP server and a WWW browser to read the documentation. I believe it will do what you want. Bob smime.p7s Description: S/MIME Cryptographic Signature
Re: xargs and max chars
michael wrote: Anybody got a quick solution to how to use xargs in corp with, say, find and ls when there's more chars than xarg (see -s) likes for a single run? eg if I run find . -name 'me*' -print | xargs ls -altd and find gives more than circa 20k chars then it appears that ls is run multiple times with the output concat-ed ie you get groups of date sorted files rather than a single list of date sorted files... (giving xargs the '-x' aborts the above cmd indicating find returns more chars than xargs can handle on a single run)... ta, michael Assuming that the 'ls -altd' is what you actually want, you might find the '-ls' option to 'find' useful. From the manual: -ls True; list current file in ‘ls -dils’ format on standard output. The block counts are of 1K blocks, unless the environment vari‐ able POSIXLY_CORRECT is set, in which case 512-byte blocks are used. See the UNUSUAL FILENAMES section for information about how unusual characters in filenames are handled. Not quite the same set of options, but the find will get you the dot files, so you don't need the 'a'. The only problem would be the 't' ;( but maybe you could pipe the 'find' output to 'sort', which shouldn't care about the line length. And the extra columns from the 'i' and 's' options could be removed with awk, sed or cut. I hope this helps. Bob smime.p7s Description: S/MIME Cryptographic Signature