Re: Re: Waiting for root file system... hang solved
Vidar Langseid wrote: Hi In article [EMAIL PROTECTED], Johannes Wiedersich wrote: Cameron L. Spitzer wrote: [snip upgrade instructions] Thanks for posting your experience! I am sure it will be useful to others. You could have a debate about whether this is an installer bug, a kernel package bug, a udev bug, or operator error. If I understand correctly, you upgraded the kernel and the new kernel would not boot. Then it would be a kernel bug. That was my initial conclusion. But then I spent some time googling for the error messages. A lot of people have had this same hang, and most of them got there by some other path than I did. So I think it may be a more general problem than that. Yes, and I is one of them who just experienced it. Just installed Etch but on first boot after installation I got the Waiting for root file system... My (old) mainbord (BE6) have two IDE hardisk controlles onboard. One generic IDE controller (piix) and a hpt366 controller. It seems that in default kernel in Etch that all IDE drivers are loaded as modules. If you then have multiple IDE controllers you are in trouble because the IDE kernel modules will be loaded in random order. I guess everyone with multiple IDE controllers will end up with the same problem and the debian installer doesn't warn you about this. Not sure how this should be fixed either. Using ext2/ext3 labels in fstab and grub config is one way I guess. Using UUID ( http://michael-prokop.at/blog/2006/08/11/stable-root-device-aka-uuid/ ) in fstab and grub config is probably another way. A 3rd solution would be to enforce that that one kernel module (piix in my case) is loaded before an other (hpt366 in my case) but that is not possible AFAIK. Or is it ? It seems to me the real issue isn't the association between drivers and partitions. The problem is that the association between device names and partitions isn't stable. Suppose I built a machine with a PIIX motherboard and an Adaptec add-in card. Then my PIIX motherboard shorts out, and the only available replacement is a VIA motherboard. If my device names to partitions association depended on PIIX getting loaded first, I just lost that stability. I've got a 50/50 chance I'll have to rescue that machine before it will boot correctly. Volume labels or UUIDs in grub/menu.lst and /etc/fstab is the only way I can see to fix that. Not sure if it is possible to fix this with udev rules. I suspect the IDE kernel modules are loaded before udev daemon is started though.. Udev is a secondary issue. It would be nice to have the association between device names and partitions stable. Then the device icons on your desktop wouldn't break when you replace your motherboard. That can be nailed down in the udev rules. Multiple sound cards can be dealt with in /etc/modules because sound is never in the boot path. NIC cards have the same problem as the broken motherboard. If I associate an ethN name with a MAC address in udev rules, then my network, and possibly my boot, break when I have to replace the NIC card. But maybe there is no way around that. It would also be nice to have a minimal set of permanent device nodes that survive without udev. Once upon a time you could boot a unix system with only /bin/sh installed. Now we need an initrd and an automatic module loader because there are too many device types to compile them all in. It seems to me that requiring for boot a user space daemon which is difficult to configure and poorly documented is a bad idea in general. Needlessly failure prone. The fact that it broke the upgrade from Sarge to Etch is a strong signal. Perhaps udev is worth it. In that case the Debian installer should use volume labels or UUIDs. Cameron
Re: Waiting for root file system... hang solved
In article [EMAIL PROTECTED], Andrew Sackville-West wrote: On Fri, Nov 02, 2007 at 03:33:06PM -0700, [EMAIL PROTECTED] wrote: In article [EMAIL PROTECTED], Johannes Wiedersich wrote: Cameron L. Spitzer wrote: [ when I upgraded Etch to Lenny, device names changed and the Lenny kernel wouldn't boot, but the Etch kernel still worked. ] If I understand correctly, you upgraded the kernel and the new kernel would not boot. Then it would be a kernel bug. [cls:] My friend in Los Angeles tried to install Ubuntu for a friend, and got stuck waiting for root file system in the middle of a fresh install from CD. When he booted his trusty Knoppix CD it revealed the root file system was just fine. I suspect udev device names are less persistent than we have assumed they are. yeah, this is probably *not* a kernel bug but more likely either a udev bug or initramsf-tools bug. Something got changed there in the device naming and that's not really the kernel's fault, so far as I know.=20 BTW, were you able to boot through the busybox? I didn't have to try. The Etch kernel+initrd still worked. As soon as the system was up I changed /etc/fstab and grub/menu.lst to use volume labels which make the Lenny kernel+initrd work too. I've had to learn my way around that having just reconfigured my laptop. The critical item is the contents of $ROOT. The value of $ROOT gets set by the kernel command line and if it doesn't match, then you have trouble. If you change that to the appropriate value you can then 'exit' busybox and the boot will carry on. I didn't know that. If I hadn't had a working option in GRUB I would have tried editing the kernel command line next. I've also rescued Debian by booting Knoppix, mounting stuff, and running a chroot shell. Once you're up and running, then rebuil the initrd's. I actually tried that before going to volume labels. Rebuilding the initrd puts the same old /etc/fstab in the new initrd image. That doesn't get you past the udev hang. I guess a more sophisticated update-initrd would alert you to the difference between the current mtab and the /etc/fstab contents. It wouldn't know whether the difference was intended, so it would have to ask what to do. And if I'd put a new device-names fstab in the Etch initrd then my Etch kernel wouldn't have worked any more. Come to think of it, I only learned about volume labels a few months ago, solving a similar problem. I installed an Etch web server on /dev/hde (a drive on an add-in ATA interface card), in a test machine that already had an Etch workstation on /dev/hda. And when I launched the hde kernel with GRUB it booted the workstation instead! I fiddled around with the udev rules but they were so poorly documented I wasn't confident I could upgrade the machine remotely. I've been using Debian for a long time. It's just *weird* to see anything broken like that. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Waiting for root file system... hang solved
I googled for the answer and didn't find it, and lots of people asking, so this is for the next person with the same problem. I installed a Debian Etch workstation, tasksel: desktop, guided partitioning, the six partitions (/ /usr /var /home /tmp swap) way. Works great. I upgraded to Lenny, the October 29 2007 weekly CD. This leaves a GRUB menu with an Etch kernel and a Lenny kernel. The Etch kernel boots just fine. The Lenny kernel hangs at Waiting for root file system... and eventually you get that (initramfs) prompt from the Busybox shell, which means it never found the root FS. Turns out this is because Etch thinks this PATA drive is /dev/hda, but Lenny thinks the exact same drive is /dev/sda. So the /etc/fstab and /boot/grub/menu.lst files have the root file system as /dev/hda1, which no longer exists. I hear ubuntu Fiesty has the same problem. The fix is to edit /etc/fstab and /boot/grub/menu.lst and change all the hard drive partition device names to unique file system volume labels. /dev/hda1 becomes LABEL=a-root or whatever you like. Then use e2label(8) to add volume labels to all the file systems. I also labeled the swap image. While you're at it, /etc/uswsusp.conf is wrong, too. (Can that file use a swap volume label instead of a device name? The manpage doesn't say.) This causes a pause during boot, where resume: is looking for the wrong swap device. Once all the partitions are labeled, you can make new initrd images. update-initramfs -u -k all You could have a debate about whether this is an installer bug, a kernel package bug, a udev bug, or operator error. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SAMBA ground-up tutorial?
In article [EMAIL PROTECTED], Chris Lale wrote: [EMAIL PROTECTED] wrote: [This message has also been posted to linux.debian.user.] I've got a Debian bastion host at a small nonprofit. Windoze. We're using CUPS on the Debian box. We want to share all the printers across all the hosts on the LAN. Many thanks to everyone who replied on and off list. Best clues so far: 1. the 2nd ed O'Reilly Samba book is on line now. 2. CUPS in Etch has been compiled with support for smbspool(8). You need that for the Linux box to use a printer offered by a Windoze box. smbspool is in the smbclient package. The howto on the NewbieDOC wiki [1] may get you started. Perhaps you could contribute your experience to the wiki page when you have got your system working? [1] http://newbiedoc.berlios.de/wiki/Installing_Samba_Linux/Windows_networking Definitely. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sneakernet .deb packages by usb thumb?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], charles norwood wrote: On Sun, 2007-03-18 at 19:32 -0700, [EMAIL PROTECTED] wrote: My desktop debian system is at home with no available internet connection. Work provides internet with no linux. is there any way to aquire .deb packages from the repository and take them home on a usb for install/upgrade? Put them on the USB drive at work. Copy them to /var/cache/apt/archives at home. APT will think you downloaded them (apt-get -d) yesterday. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Speeding up boot time
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Michael Pobega wrote: I'm looking to speed up my Debian Etch boot speed, but I have no idea where to start. Install the uswsusp package. Don't shut your system down. Just suspend it to the swap partition with s2disk. When you turn the power back on, boot the same kernel as last time. When it discovers the suspended system it comes up really fast. Just don't change your hardware while it's suspended. A customer did that with her Mac for six years. Suspend to disk, resume, never rebooted. She asked me to revise her OS, and I ran fsck first. It found one wrong inode number. Does that mean she had an uptime of six years? Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: thunderbird: how to backup mailbox and restore quickly
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Chris Bannister wrote: On Sun, Mar 04, 2007 at 11:31:34PM -0800, Cameron L. Spitzer wrote: A very large file in /var/mail was created instead. Any idea? That would be the default delivery. Your procmail file should end with this :0 /dev/null to prevent that. Won't that lose mail if any recipe fails? It would if this were procmail being used as a mail delivery agent. Here we are using formail and procmail together to pluck individual messages out of an Inbox file. The mail has already been delivered. We're sorting it post-delivery. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
correction, Re: thunderbird: how to backup mailbox and restore quickly
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Yuwen Dai wrote: Now I move both Inbox and the expected result to my home dir. But command $ formail -s Inbox |procmail /home/yuwen/rcfile Should be $ formail -s procmail /home/yuwen/rcfile Inbox No pipe between formail and procmail. Formail will invoke procmail over and over, once per message. With a suitable rcfile, splits large Inbox into multiple mbox files, sorted any way you like. Move them to where you want them in Thunderbird's mail storage hierarchy. Thunderbird will create .msf files to match. Also, end your rcfile with this recipe :0 /dev/null to prevent it from writing messages that didn't match onto /var/mail/yuwen Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: thunderbird: how to backup mailbox and restore quickly
In article [EMAIL PROTECTED], Yuwen Dai wrote: [I wrote] $ formail -s mboxfile | procmail recipefile recipefile looks like this: :0: * ^Received: .* myserver.example.net ;.* Dec 2006 /home/me/December I created a similiar rc file like this: :0: * ^Received: .*;.*Nov 2006 /tmp/November The expected file /tmp/November didn't create. Maybe Procmail won't deliver in /tmp because everybody can read and write there. Maybe ^Received: .*;.*Nov 2006 doesn't match anything in the mail. Test that with egrep. A very large file in /var/mail was created instead. Any idea? That would be the default delivery. Your procmail file should end with this :0 /dev/null to prevent that. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: thunderbird: how to backup mailbox and restore quickly
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Jörg-Volker Peetz wrote: Yuwen Dai wrote: I'd like to backup part of my Inbox, e.g., the emails of last year as my Inbox becomes larger and larger. I also want to restore and read the backup quick in thunderbird if need. What's your suggestion? How about creating a new folder, e.g., Inbox-2005. Then move all mails to be backed up into this folder. Archive this folder with soumething like tar czf Inbox-2005.tgz ./.thunderbird/.../Inbox-2005* (The three dots ... have to be replaced by the exact path) Afterwards this folder can be removed. In case you need it back, just stop thunderbird, un-tar the backup archive file, and restart thunderbird again. The last time I looked, Thunderbird stored mail in ~/.mozilla-thunderbird/[stuff].default/Mail/[account name]/ Each folder was a text file in the old, ugly, standard mbox format. It also maintains a Foldername.msf (message summary file?) file that gives kind of an overview of the corresponding mbox, but if mbox.msf is missing it will just read the mbox and generate a new one. You can bust an mbox file into individual messages with formail -s. You can sort them any which way with procmail. Do something like this while Thunderbird is not running: $ formail -s mboxfile | procmail recipefile recipefile looks like this: :0: * ^Received: .* myserver.example.net ;.* Dec 2006 /home/me/December :0: * ^Received: .* myserver.example.net ;.* Nov 2006 /home/me/November ... $ Local=.mozilla-thunderbird/bletch.default/Mail/Local Folders $ mv December November $Local Sure beats sorting them in Thunderbird's GUI. You'll have to look at your mail headers to figure out regular expressions to match the messages you want to pick. I'm using the final Received line because the date is more reliable there than in Date: lines sent by random Windoze users. Notice the cumbersome quoting to preserve the space in the directory name. When you start Thunderbird it will create Local Folders/November.msf etc. Similarly you can move obsolete mbox files out of ..mozilla-thunderbird and compress or make tarfiles or whatever. Just do it while Thunderbird is not running so it doesn't get confused, and don't forget to remove the corresponding .msf files. They're just mbox files. In a pinch you can open them with mailx and save groups of messages to a new mbox file (or onto the end of an old one) with its s command. Good old mailx. I haven't tried this in a while. Test it first, and let me know if it doesn't work any more. Thanks Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: No Mouse in X
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Dave Walker wrote: The Problem: When I run X (startx) from any user account, it begins, loads Gnome (just today downloaded it) but the mouse does not move. I found out that Ctrl Alt Backspace quits X, thankfully, so I can return to the command line. My Analysis, such as it is: I have run dpkg-reconfigure xserver-xfree86 many times to check the mouse settings, and everything looks fine. If I look at the file: /etc/X11/XF86Config-4 it shows two entries for mouse one after the other as below... Are you installing Sarge 3.1 or Etch 4.0? My mouse broke when I moved to Etch, and replaced XFree86 with Xorg. Solution was to install the hal and xserver-xorg-input-mouse packages. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Installing Etch with GUI on T20: no screens in X
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], john gennard wrote: I have an IBM T20 Laptop and want to put on it Etch (Kernel 2.6.18-3) with Kde. The installation without GUI is fine, but Kde will not launch. Error messages in /var/log/kdm.log indicate that the installer assumed the laptop has a Synaptics Touchpad and this cannot be detected. /proc/bus/input/devices show that the device is in fact a 'TPPS/2 IBM Trackpoint' Googlings for the Trackpoint show a confusing picture which I don't understand. However, as I have considerable arthritis in the hands and fingers, I have been using an external PS/2 mouse, so I can do without the 'fiddly' Trackpoint. I commented out references to Synaptics Touchpad in /etc/X11/xorg.conf, but Kde still does not launch and the error messages )in Kdm.log) now say:- (EE) No devices detected Fatal Server error: no screens found. At the moment, to get into Etch, I have to boot into single user mode, otherwise the Laptop locks solid. I had a similar problem. It turned out my system was missing packages hal and udev. X.org wants to use keyboard and mouse via hardware abstraction layer, but the dependency was missing. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: hardware problem? continual errors with drive and kernel
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Roberto C. Sanchez wrote: On Fri, Feb 09, 2007 at 01:50:40PM -0800, Ric Otte wrote: =20 I was wondering if anyone else had had problems like this and had any suggestions. I don't know much about hardware and don't begin to know wh= [Kernel oops in stable code, suggesting cpu-mem issues] Get a Knoppix CD or DVD and run memtest on your system for a couple of days. Have you *ever* seen memtest catch a pattern sensitive memory failure? Memtest is good for finding stuck bits and stuck address lines. But there is another class of failure: walking wounded chips with electrostatic discharge damage, bad signal integrity, failed or inadequate bypass caps, cracked traces with high resistance bridges. These can give unreproducable soft memory errors when you hit just the right address sequence with just the right pattern in RAM. I've had bad motherboards that could run memtest for days, at all temperatures. But they'd give a kernel oops or signal 11 before they got to the end of a kernel compile. GCC generates a more chaotic pattern set than memtest does. I had one where sliding the scroll bar up and down on an xterm would do it. Desktop-class PC hardware doesn't provide fault isolation tools. You just have to replace parts one at a time. First measure the supply voltages. You can chase an inadequate power supply for days; it makes everything else look flaky. If you find 12V, 5V, and 3.3V within 5% of nominal, with the CD spinning and hard drive seeking (fsck -nf), move on to other things. Next look at BIOS and correct any overclocking or aggressive timing settings that some fool may have left on the board. You've already eliminated the hard drive. Try the cable, just to rule it out. I haven't seen a flaky CPU in years, they seem to fail all the way when they fail. Memory can be bad, especially due to ESD. But my bet is on the motherboard. Motherboards get micro cracks, bad solder, wrong value bypass caps, all kinds of hard to isolate stuff. People don't handle them delicately enough, and they're barely strong enough not to crack when you clamp those cooler retainers down. Even Intel and Tyan and Asus. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
X.org mouse broke on dist-upgrade, solved ?
[This message has also been posted to linux.debian.user.] A colleague did a dist-upgrade, sarge to etch, and everything still worked except the X Window System. dpkg-reconfigure left him with No screens found X unusable. The standard fix for that, boot Knoppix and use the xorg.conf file it generates, worked, except his mouse was dead. Turns out the dist-upgrade brought in udev and hal and module-init-tools, but no new kernel image. He was still running the 2.4 Sarge kernel, no udev, no hardware abstraction layer. apt-get install linux-image-2.6-486 update-grub fixed the problem. Is dist-upgrade supposed to bring in the right kernel/initrd? Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: getting broadband working on my new AMD64 box....
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Michael Fothergill wrote: [new install, Internet not working] My old box is an AMD Duron 1200 Mhz 32 bit machine which is running Etch i386. The internet connection works fine on it. I use NTL broadband cable [NTL's cable box at eth0]. heh. This is in my rbldnsd file: 86.0/11 ; ntl.com cable modem users .UK (0-31) ; 86.2/15 ; cable/broadband.ntl.com CPE ; 86.6/15 ; cable.ntl.com CPE Apparently some of your neighbors are hosting spam bots. I also looked in the network configuration box in the network admin icon. I found that the DNS server numbers for NTL are 194.168.4.100 and 194.168.8.100. There are still thousands of wide open name servers on the net. My advice is find a few that work, and make a static /etc/resolv.conf. Then when your consumer broadband vendor's name servers fail you can still see. Get everything else working. Then go back and try the name servers NTL *wants* you to use. This way you are only troubleshooting one thing at a time. Acquiring name servers via DHCP is convenient, but it's one more thing that can fail. To find some open name servers, try the ones shown in Whois or host -t ns, for a dozen random domains. (*cough* mmu.ac.uk *cough* bbc.co.uk *cough* tees.ac.uk *cough*) I also ran ifconfig as follows: [EMAIL PROTECTED]:~$ /sbin/ifconfig | grep -A 1 eth eth0 Link encap:Ethernet HWaddr 00:02:44:33:AD:C3 inet addr:86.22.11.90 Bcast:255.255.255.255 Mask:255.255.252.0 -- eth1 Link encap:Ethernet HWaddr 00:0A:E6:0C:AE:D4 UP BROADCAST MULTICAST MTU:1500 Metric:1 My question is: what is the best use of this information when I disconnect the old box and plug in the new one to help it hook up to NTL? I could just put the DNS numbers in the network admin window for example. That might not work. The new box has a different MAC address (HWaddr) and NTL may want to give it a different IP address. If you're supposed to use DHCP, use it to get the right IPA. Just be skeptical of the rest of the stuff it tells you. The other thing you might ask the old box about is /sbin/route -n | grep G to find out where the gateway is for destinations beyond your network segment. If your link is up, you can ping it. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Handing over boot sequence from one drive to another
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Douglas Allan Tutty wrote: On Sun, Jan 28, 2007 at 12:54:46PM +0100, MeneM wrote: Could anyone tell me how to copy my current debian installation to a different external iomega jazz drive, I recently copied an etch install to a new drive. I had root, swap, /usr, and /var on their own partitions. Copied each to empty file systems on the new drive. Copying root is tricky because you don't want to copy /proc or /sys or /tmp, you want empty mount points with the same ownership and permissions. If you're new to Linux, a new install would be easier. and be able to boot from it in case of emergency? That depends on whether the external drive can be booted by BIOS. If not, the easiest thing would be to leave a copy of /boot on the old drive, and just make an entry in /boot/grub/menu.lst that knows where the new root is. Nobody said your /boot directory has to be on the same drive as your system. I installed etch on an old system where the BIOS couldn't recognize a 10 GB drive. I had an old 540 MB drive lying around and put /boot on that. BIOS doesn't know there is a 120 GB drive next to it, and doesn't care. I suppose you could put /boot on a CD. I'm figuring; reading some articles after some merry googling, that I need to copy /boot /etc /usr/sbin /usr/bin /sbin /dev and a bit of /usr to it and then make it bootable with fdisk right? You forgot to make mount points for /proc and /sys and an empty /tmp directory. You forgot /var/. See how this is not a good idea? Bootable means there is an MBR where BIOS can see it, that points to a /boot directory that BIOS can see. Linux does not care about the boot flag in the partition table. You see I'm trying to go down this route, but I cannot get the jaz drive to boot. There's no BIOS option for it. So I have to boot off CD or something and then hand the boot process over to the jaz drive. but _how_? You need GRUB to be able to see vmlinuz, initrd, and grub/menu.lst. But only stage 1 needs BIOS. Make a boot floppy. When it works, put an El Torito boot catalog on it and burn it to a CD. Floppies go bad. A pre-requisite to booting the kernel on the jazz drive is that the jazz drive be bootable. OP said that's not an option. Old BIOS. So what __will__ your bios boot? Some suggestions for testing. Get the hd-media kernel and initrd.gz from Etch. What's an hd-media kernel? Package search doesn't see it. Put the hd-media directory on it. What's an hd-media directory? My etch install doesn't have that. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: The Lost Sheep and Networking (3c509)
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Matt Zagrabelny wrote: When installing, debian-installer cannot recognize my card, and gives me a list of modules. In it is ``3c509'', but without a letter. Relatively few 3C509 (Primat chip) s were made. They had only 3Com's proprietary jumperless discovery mechanism. Then ISA Plug-and-Play came out and it wasn't compatible. The revised chip is Pandora, 3C509B, with both the 3Com and MSFT/Intel schemes. We sold millions and millions of those. The 3c509.c driver handles both. However, it does say ``3Com Etherlink III'', so I select it. =20 DHCP fails, despite the fact that my router has DHCP enabled, and many other computers in my house are using it. Setting up static IP does not work either. However, the corresponding light on my router does light up, and, what's more, when I run ``dhclient eth0'', my router claims that the computer connected to it, and got the IP 192.168.1.102 (I can tell by the MAC address that it is my computer). However, the computer itself is simply saying that it got no DHCPOFFERS. 3C509[B] is an ISA card, and its interrupt assignment is stored in a tiny EEPROM. They can be changed to fix a conflict, or if there is more than one 3C509B in your machine. If your router got your MAC address but its offer never came through, you can send but not receive. That is a symptom of the interrupt on the wrong line. Perhaps it was changed early in your 3C509B's life. If your motherboard has an ISA Plug and Play BIOS, it is supposed to work around that and find the right IRQ line. But many BIOSes are broken. That is why Linux has its own ISA Plug and Play. Did you try it? =20 I am at a total loss here. It is clear that the card is working, because, at the very least, the MAC Address is being transferred. However, I cannot connect anywhere, including to 192.168.1.1. =20 Any ideas? There is a 3c5x9setup program in the nictools-nopci package. Perhaps you can put the IRQ (and I/O port base) back to the factory condition. I think it was IRQ 10 and base address 0x300 but don't bet on it. Otherwise, figure out what they are and tell the driver about it in modules.conf. Historically, 3c509 was compiled into the kernel, not loaded as a module. That was because you wanted to run the register location discovery before the machine was really running. That is, at kernel load time, not at modprobe time. But the card is obscure enough now, and few ISA devices exist outside the southbridge chip any more, so the hazard is less. The description in modules.conf(5) is ambiguous and there is no example of a *complete* options line, so I cannot tell you a valid syntax for the line in that file that you need. If you have a kernel with 3c509 compliled in, the kernel command line argument would be eth=0x300,10 try knoppix, see if there are any parameters being passed to the module that corresponds to the interface. if you believe in voodoo, move the nic to a different pci slot. That would be some powerful voodoo. I don't think your motherboard would survive. Another reasonable option would be to retire the 3c509 due to its support issues (incomplete modules.conf manpage) in Linux, and get a PCI NIC. Almost any PCI NIC will outperform any ISA NIC. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Adding a new HDD - how do I move / everything ?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Abu Zaher wrote: I'm currently running my Sid on a 40GB hdd which is running out of life, In a few days I'll buy a new 80GB. Now what is the best way 2 dump my whole / to my hdd so that I can boot and do all my stuff from that hdd? I mean the above discussion is for /var/lib/mysql, will the same work for / ? I've copied whole Debian installations. There are many ways; I'll just show one. I assume your whole system is on one drive /dev/hda, and you're replacing it, and the slave (/dev/hdb) on the first ATA channel is where you have temporarily attached the new drive. 1. Boot your old system with the new drive at hdb. 2. Observe your old partition table with fdisk -l. 3. Observe your disk utilization with df. Write it down. 4. Partition the new drive the same way, except make the partitions bigger. I like cfdisk for this. Keeping the same partition table assignments means you won't have to edit /etc/fstab on the new drive. 4a. If you have time, run badblocks over each new partition. It won't report any errors, but this gives the drive a chance to spare out any sectors which may have gone bad in shipment and storage. It also brings it up to operating temperature. 5. Build file systems and a swap area on the new drive. Swap area should be at least twice the size of the real RAM you expect to have over the life of this drive. More than three times is a waste. If you're using a customized kernel, make sure it knows how to mount the type of file systems you chose. 6. (optional) Format a brand new floppy disk and make a selfcontained GRUB disk out of it. fdformat /dev/fd0 cd /usr/lib/grub/i386-pc cat stage1 stage2 /dev/fd0 floppycontrol -f cat stage1 stage2 | cmp - /dev/fd0 Floppies are often defective so you have to verify them like that. 7. Insert your Debian install CD and reboot. shutdown -r now At the boot prompt, hit Enter. When the installer asks you to select a language, ignore it and hit F2, Enter. You should get a root shell prompt. 8. For each file system on disk, make temporary mount points, mount the old and new file systems, copy the old to the new, and unmount both. Let's say you have root on hda2, /usr on hda6. and /home on hda7. This may look like a shell script, but you can just type it carefully into the shell prompt. for i in 2 6 7 do mkdir /mnt/hda$i /mnt/hdb$i mount /dev/hda$i /mnt/hda$i mount /dev/hdb$i /mnt/hdb$i cd /mnt/hda$i cp -a . /mnt/hdb$i cd / umount /mnt/hdb$i /mnt/hda$i done 9. Now you have a complete copy but it's not bootable. Shut down your computer and move the new drive to the master position on the first ATA cable. You'll probably have to adjust the jumper. 10. Boot the CD again. This time type rescue at the boot prompt. The installer will set up in RAM and turn on the network, but then you will get an Enter Rescue Mode dialog where it will ask you where your root file system is. Select the appropriate partition (part2 in this example) and hit Enter. The next dialog offers to reinstall the GRUB boot loader. Select that and it will give you some options to make the new drive bootable. Install GRUB to the master boot record. 11. Reboot without the CD. GRUB should find your menu.lst, kernel and initrd, and your new system should come up just like the old one. You might want to run grub-install and update-grub just so you have the same version as before, but it's not really necessary. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Adding a new HDD - how do I move /var/lib/mysql ?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Roberto C. Sanchez wrote: On Wed, Jan 10, 2007 at 02:46:29PM -0500, Kevin Mark wrote: probably the simplest way is to just make a new /var on the new drive and move all of it there. -add new HD to computer -partition HD -format HD -mount new HD as /mnt/newvar -stop all processes (like single user mode) -copy all of /var to /mnt/newvar -move /var to /oldvar (not sure this works) -create /var -edit fstab to add mount point to make new partition point to /var -reboot and hope it works A couple of points: - This will be lots easier and less painful if you do it all in pass [?] [all in passing?] [all in one pass?] [huh?] from a LiveCD (though Knoppix is my favorite, I'd recommend the Ubuntu LiveCD since it suppors LVM--see next point) I'd recommend against the live CD for this. If this is a production system you want to use kernel and utilities that were built together. If you'll be running the file system with Sarge, don't build it with mke2fs from Sid. Besides, the date and time zone are probably wrong on your live CD. Set the time correctly before you make a file system or copy stuff around. To replace /var with minimum downtime, 1. shut down, add new drive 2. boot as usual 3. fdisk and build new file systems 4. telinit 1 # shuts down everything that's got files open on /var 5. mount new file system as /newvar, cp -a /var /newvar 6. cd / ; mv var oldvar ln -s newvar var ls /var 7. if /var looks right, telinit 2 # starts it all up again 8. dont forget to put /newvar in /etc/fstab 9. when you're sure things are good, rm -rf /oldvar Don't run telinit 1 if you don't have access to the console, but then how did you add the drive? Notice I made the symlink without the leading slash. This comes in handy if you need to boot a live CD and run your system from a chroot some day. Symlinks should usually be relative paths. In any case, I wouldn't just move all of /var. If the problem is /var/lib/mysql, then just move that. Make a partition or a logical volume /extra and copy /var/lib/mysql to /extra/lib/mysql. Then you don't have to go through the telinit run level change, you can just stop and start mysql. - Save yourself some trouble and setup LVM and then create a single volume group and put your logical volume on that. That sounds like good advice. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Ghorgeous Blond TRANASEXUAL Pvosing In Byikini Outdoor
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], a spammer wrote: [hashbuster] Take that! http://from.mayroty.com/ Bcusty Tzanned Brunette [garble raunch] The previous junk was apparently brought to you courtesy of the Python Video/Webfinity/Dynamic Pipe spam gang. The barnyard porn site is hosted on btnaccess.com a/k/a pccwglobal.com. See http://www.spamhaus.org/sbl/sbl.lasso?query=SBL4197 and http://www.spamhaus.org/sbl/sbl.lasso?query=SBL49941 They are hiding behind a (fake) colocation provider Rackco.com which seems to be the spam gang itself. If you have any contacts with PCCW, please yell at them loudly. They have been ignoring all complaints and earning Spamhaus escalations. Not good for business. Sorry, but they're gonna keep hitting this list until they can't find hosting any more. cls -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
APM no kernel support ? Use Software suspend 2 ? how ?
[This message has also been posted to linux.debian.user.] I had sarge on a Compaq Armada 3500 notebook. No ACPI, but APM suspend to disk was working with an upstream kernel with APM compiled in. Fresh etch install on the same system, 2.6.17-2-686. modprobe apm gives apm: BIOS version 1.2 Flags 0x03 (Driver version 1.16ac) apm: overridden by ACPI FATAL: Error inserting apm ([path]/apm.ko) no such device. I added acpi=off to the kernel command line in GRUB, but I got ACPI anyway. Is there some boot time kernel argument to shut off ACPI in this kernel? Should I use Software Suspend 2? Is it in Debian's kernel already or do I have to make my own? There's a Debian page in the Suspend2 Wiki but it was written when sarge was testing and it's not clear what still applies. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Replacing Gnome with KDE
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], [EMAIL PROTECTED] wrote: Thank you, everyone who offered suggestions and observations. I've got KDE working, basically by using aptitude to remove the remaining bits of gnome, and then (re)installing the kde meta-packages. Just for amusement, try installing icewm too, and see how a WM that loads in about a tenth of the time feels to you. And then you could move on... Having just attempted to replace GNOME with KDE, and ending up with an empty Control Center, I'd be willing to give icewm a try. What do you use for sound daemon? I need to run Skype and the Flash player (with sound, inside Firefox or Seamonkey) at the same time, and that seems to take a daemon in front of /dev/dsp. aRts? esound? jackd? Something else? Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: GNU/Linux glory
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Kevin Mark wrote: On Sun, Nov 26, 2006 at 11:47:15AM +0200, Tshepang Lekhonkhobe wrote: Hi, I'm currently running gedit and gnome-terminal for about a week now without any noticeable side-effects after uninstalling them (these what is to notice? that you are running a program from ram that is not installed? or that it is running fine for a week? It's not running from ram, it's running from the file system. An executable that's busy isn't removed when you remove its last link in the directory structure. Its inode still exists, tying up all the relevant disk blocks, and can't be reused until all the processes that were executing it have exited. Unix had that feature when I started using it in '83. It seems to me Debian's upgrade-while-running behavior kind of depends on it. If dpkg(8) removes /sbin/init, the old one better keep running until the installer creates the new one and the old one exec(2)s it. Likewise for the sshd(8) you did it through. -- Cameron http://notwindoze.blogspot.com/2006/11/worlds-worst-unix-clone.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How can I verify hardware compatibility?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Grok Mogger wrote: I'm planning on buying components individually and piecing together a computer. I'd like to install Debian Linux on it. I'm not buying any of the latest and greatest stuff, and the hardware I have picked out is pretty standard fare, nothing too fancy. So I'm pretty sure everything will just work (TM), but I'd still kind of like to verify each piece before I just buy everything, slap it together, install and pray. My experience has been CPUs, host bridges and wired Ethernet just work. Brand new video hardware, wireless, and modems don't, unless you happen to get lucky. When modems moved from the ISA slot to PCI, the convention of having them emulate an external Hayes command set modem connected to a serial port went away. You won't find a new internal hardware modem at any US retailer today. They just aren't made any more. Even the controller-based with DSP models won't just work because they don't bother with looking like UARTs any more. They've got proprietary interfaces, and you can't tell from the packaging which chip they use, so you can't be sure you're getting one we have a driver for. If you want a dial-up modem that just works, get an external modem. Most of the motherboards I've seen in the last two years with built-in video have used Nvidia, or the Intel or VIA northbridges with the video built in. Intel just works, Nvidia and VIA don't, because the drivers are proprietary. Knoppix will make graphics, but it's using the basic VESA dumb frame buffer driver, not the drawing engine in the chip. It's barely fast enough for office work and not fast enough for games. So visit the XFree86 site for Debian 3.1 sarge or the Xorg site for 4.0 etch and make sure your xserver knows how to operate the particular video hardware that comes with the motherboard you're looking at. Otherwise you'll have to make do with software graphics, or stick an older video card in an AGP or PCI slot until the free drivers catch up, and there's no guarantee they ever will. I haven't found a PCI wireless card that just works yet. The best you'll do is find one whose Windoze driver works with the Linux NDIS wrapper. You might actually do better with wireless on a USB dongle. Low end printers and all-in-ones probably won't work. Even the printers that say works pretty well at linuxprinting.org are going to have little anomalies like spitting out a junk page in front of each new print job. I had a customer with an HP all-in-one with a supported HPIJ driver that did that. You're safest getting a printer with Postscript or PCL5 *in the printer* not in the printer driver. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Mailman with Apache 2 on Sarge
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], SAJChurchey wrote: the e-mail is getting sent to the proper command through the aliases (In this particular case, I'm trying to subscribe by sending an e-mail to mailman-subscribe), but the owner doesn't receive an e-mail and neither does the subscriber (I've set the list up to send confirmation to the user). I've checked /var/log/mailman/smtp-failures, and I'm getting the following errors: Nov 09 11:13:15 2006 (20371) delivery to [EMAIL PROTECTED] failed with code -1: (-2, 'Name or service not known') Nov 09 11:13:15 2006 (20371) Low level smtp error: (-2, 'Name or service not known'), msgid: [EMAIL PROTECTED] If those are the actual log messages, the problem is Mailman is trying to send to a domain that does not resolve. Does your hostname command say lists.mydomain.mytld? Does the string lists.mydomain.mytld appear in Mailman/Defaults.py? If so, you have to override it with a real name that actually resolves, in mm_cfg.py. Getting Mailman to send mail the first time is tricky. For some reason known only to the authors of the Python library, its mail seems to come out of the external interface on its host system. That is, the circuit it opens to talk to the SMTP server has a source IP address that is not localhost. So you have to make sure that source IP address is in mynetworks in /etc/postfix/main.cf. I've been reading through the Mailman FAQs and have tried several of the interactive python tests, and I've had no problem sending e-mail as the list user. I'll try sending a subscription e-mail using this method, but as far as I can tell the Postfix server is receiving and delivering the mail to the proper command. What kind of troubleshooting can I do on /var/lib/mailman/mail/mailman? If you are mailing stuff into Mailman through Postfix and /var/lib/mailman/mail/mailman, then that is not the problem. For me, getting Mailman to *send* has been the not obvious part. Any other problems I should look for? Mailman is documented but there is little troubleshooting information. It was not obvious to me that there are *three* kinds of Mailman list. The first list you create is special. It's the one that will send out password reminders and you really can't use it for anything else. The documentation suggests that you are just creating a test list but that is not true. It also assumes you will call that list mailman, which isn't a bad idea. The other kinds are regular and umbrella. It may not be obvious to you that you have to run mailmanctl start, and install the crontab, as the mailman user, not as root or postfix or nobody. You should also watch the error log. Mailman's internal data structures are delicate, and when they get corrupted those logs grow fast, because they are filling up with Python stack traces. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: document processing
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Douglas Tutty wrote: I'm revisiting how I make documents. I have been using lout since I started with linux in 2000 My primary use is for letters and notes but also larger projects. I don't like wysiwyg. I want to be able to make: ps, pdf, txt, html. I want something that is simple, probably a markup language, but without excessivly long tags or difficulty changing things like margins for non-html output. What do people find works well? LaTeX just gets better. Every time I do a non-trivial document in anything else, I regret it. Larger projects - LaTeX. For trivial documents, KWord is better than I expected; I like it a lot better than Openoffice Writer. If you don't like WYSIWYG and you have to give a slide show, check out mgp (MagicPoint). Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: argh! linux and floppies
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Mark Grieveson wrote: I did try to use superformat. It seemed to do the low level format OK, but then it got to running mformat and the drive just made a whole lot of noise and eventually gave me the following error: This is exactly what happens with me as well. One more thing. Scraping the mold helps the head contact the surface better, but it doesn't erase the remnants of the old tracks very well. Those old tracks impose background noise on the new ones and make read errors more likely. Use a bulk eraser if you can. The ones they sell for audio tape work fine. Scrape the mold, then bulk erase, then format. I don't know what fine adjustments superformat thinks it's making. It's been a long time since I looked at the NEC floppy controller but I don't remember it having very many knobs to twiddle. I suspect its reputation for doing a better format than fdformat may be due to it doing a mold scraping pass first. But the best advice is just avoid floppy disks if you possibly can. Fry's has a 1 GB USB flash drive for US$15 after the rebate. That's 700 floppies' worth and it fits on your keyring. They had a 128 MB drive for three bucks. Floppies are obsolete. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: argh! linux and floppies
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Kent West wrote: I have just 30 minutes ago tried to use three 3.5 floppies on two different machines, and can't get anywhere with them. I decided to put it on the back burner and read my email when I came across your post. Could be the media. I have twenty year old floppy disks with no defects. Most floppies I've bought in the last ten years went bad within eighteen months. The old ones had a fungicide to protect the glue from mold. They don't seem to do that any more. I stopped buying new floppies when I got a box where a third had to be scraped before they would work. Surplus places have old ones, unsold antique software still in the EULA wrap. Those work better. Try scraping the mold off with a couple of fdformat -n /dev/fd0u1440 passes before you try to write real data it. Then check it for errors with dd if=/dev/fd0 conv=noerror | sum or something. But if you salvage a bad disk that way, don't expect the data on it to last more than a week or two. Try cleaning the head in the drive with a qtip and rubbing alcohol. I've tried cfdisk and fdisk to look at the partition(s) (do these work on floppies?), and mformat, and mkfs.vfat, and fdformat, and all I ever There's a master boot record on the first sector. But the bytes at the end where the partition table goes aren't used. get is something like could not get geometry of device or Problem reading cylinder 0 or Unable to read /dev/fd0, etc. Fdisk and mformat are failing because they read before writing anything. An old floppy (more than a month since you wrote it last) has faded. You need to write a new set of sector marks with fdformat before it will read reliably. Cfdisk will still fail because there's no geometry to get. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Getting rid of ghostscript message
In article [EMAIL PROTECTED], Peter Thomassen wrote: Hi, I'm trying to convert a dvi2ps-generated file to TIFF, using $ gs -q -dNOPROMPT -dBATCH -sDEVICE=tiffg4 -sOutputFile=file.tif file.ps Try -dNOPAUSE and -dQUIET -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Getting rid of ghostscript message
In article [EMAIL PROTECTED], Peter Thomassen wrote: Hi, I'm trying to convert a dvi2ps-generated file to TIFF, using $ gs -q -dNOPROMPT -dBATCH -sDEVICE=tiffg4 -sOutputFile=file.tif file.ps Try -dNOPAUSE and -dQUIET -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: BIND name caching and forwarding
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Andrew Critchlow wrote: Hi everyone, Anyone good with Bind? Yeah, the guy who packages it for Debian. I was wondering that if I set up a dns = server just for local hosts would I have to include forwarders to the ISP d= ns servers for internet lookup? If you don't, your BIND is going to be querying the root servers to find out who's authoritative for .com all the time. That's abusive, if your ISP even lets you do it. Also does anyone know how I could configure= a caching-only nameserver? thanks= apt-get install bind9 bind9-host dpkg-reconfigure bind9 Or something like that. Let Debian do it. You might want to put something in netfilter so the general public doesn't use your BIND as their forwarder. Block INPUT to port 53. Then unblock INPUT to port 53 for networks where your friends are. Here's a more interesting problem. Say I've got about six thousand CIDR rules for rbldnsd. (Meaner than sbl-xbl, more useful than SPEWS.) I like what it does for Postfix and I want to keep it. Now say I want to run BIND on the same interface at the same IP address, so I can be authoritative for a few domains as well. Is there a clever way to make BIND be a forwarder for rbldnsd? Is this ridiculous wrt performance? Perhaps have rbldnsd listen on a weird port and have BIND query it on some private address via port forwarding? Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
power management in etch kernel ?
[This message has also been posted to linux.debian.user.] I've installed etch with default kernel on a laptop with APM BIOS but no ACPI. It runs a lot hotter than it did with my customized kernel on sarge. I added apm=on to the boot options to get apm -s to work. Still got heat and battery life issues that sarge didn't have. Are there more boot options to enable the slower clock and halt between interrupts kernel features? Should I be using some other kernel? Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: spamcop
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], s. keeling wrote: Mumia W.. [EMAIL PROTECTED]: On 09/28/2006 12:23 PM, Andrew Sackville-West wrote: [...] also, threadjacking, but its spam related... is anyone else getting a lot of these bounced email spam? I'm getting a TON of it lately. It all has a .zip or .com binary attachment, so obviously its a virus or I've gotten a couple of such messages with the virus removed. Follow it up. Complain to the bouncer that their SMTP is misconfigured and they're sending bounce messages to innocent third parties whose From: addresses were forged by spammers. My experience has been that anyone clueless enough in 2006 to run a broken server (Barracuda appliance, Qmail...), that sends backscatter to random addresses supplied by spammers and malware, is too clueless to understand a complaint about it. If you can even get a complaint through. Look their domain up in whois -h whois.abuse.net. postmaster (default, no info) every time. Send mail to [EMAIL PROTECTED] and [EMAIL PROTECTED] User unknown, or sorry, your spam report triggered our spam filter. Not just clueless, clue resistant. Clue repellent. Report them to RFC-Ignorant.org when that happens. Then add them to your local DNSBL if you can get away with it. Or your firewall. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: (no subject)
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Chris Humphries wrote: +-- | On Thursday, Sep 21, 2006 at 03:51:36AM +, s. keeling wrote: | operator [EMAIL PROTECTED]: | Could someone recommend a substitute for k3b? Do what k3b is doing manually (it just calls external programs to do it's work). That is what the text below is. There's another thing k3b does. It knows the squirrelly mappings from the SunOS scsibus,target,lun or portable devicename:scsibus,target,lun language in the cdrecord(1) man page to the device names on the Linux systems people actually use. On some distros, it (with KDE set up correctly underneath) knows whether there is a Hardware Abstraction Layer and what names to use so the right driver modules get loaded. If you don't know all that stuff and you try to use cdrecord directly, you either get lucky or cdrecord does not work. It's nice that S. Keeling got lucky using /dev/hdd on his particular system, but that device name will not always work. It would be good if there were an organized documentation someplace explaining to newbies what the difference between /dev/hdd and /dev/scd are, and how to tell whether your installation is using sg SCSI emulation or IDE-CD and how to get the automounter to let go of the CD drive when the distro left it on by mistake, and some of the other things that prevent cdrecord from just working. If it exists, I haven't found it. I use k3b and xcdroast because cdrecord doesn't always work by itself and I've never been able to figure out why. Cameron There are several scripts and things around that do various ways of doing the things below, k3b is just a nice fancy gui for it instead. -Chris | | mkisofs -R -o /scratch/iso/track_01.img \ |-pad -allow-leading-dots -max-iso9660-filenames \ |-r -relaxed-filenames /scratch/afio | | Then: | | cdrecord dev=/dev/hdd -eject \ |-tao -data /scratch/iso/track_01.img -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: alternative for rdiff-backup
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Kevin Mark wrote: --ABTtc+pdwF7KHXCz Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Wed, Sep 13, 2006 at 05:49:57PM +0100, Michael Ott wrote: Hello! =20 rdiff-backup crashed every second use. Hi Michael, it seems a wiser thing to do to try to help the maintainer of rdiff-backup as both of you have an interest in seeing it work, that is For what it's worth, I've been using rdiff-backup for a couple of years now. Every night my sarge system backs itself up to my colocation company's file server. Now and then one of my users deletes something by mistake and I bring it back for them like magic. What a great program. We did not use the rdiff-backup package from Debian because it's too old. Rdiff-backup wants to be the same version at both ends, and he installed from the upstream so I had to. But it's running over Debian's rsync and ssh. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: (OT) Prejudice against sendmail?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], [EMAIL PROTECTED] wrote: qmail has the least Debian support, due to uninteresting licensing issues, but has a large user community and excellent documentation (www.lifewithqmail.org, qmail.org). It runs nicely on Debian systems, and is unlikely to require upgrading. Sorry, no. On anything bigger than your personal mail server, Qmail is going to require replacing with a modern MTA. There's a denial of service bug. When the queue gets too big, qmail-send just spins around deciding what to send next and never gets around to actually sending anything. There's a backscatter bug, due to Qmail's inability to reject at SMTP time stuff it won't be able to deliver. It bounces spam at random victims instead of rejecting it. There's a mailbombing bug. Say you have a mailman list with a thousand subscribers, and 250 of them are Yahoo.com. A modern MTA will connect to Yahoo once, give 250 RCPT TOs, and one DATA. (Or maybe break them up into chunks of a hundred.) Qmail will connect to Yahoo 250 times, amd give one RCPT TO and the same exact DATA. Yahoo will block you for that. All three of these bugs are architectural. You can't fix them without such extensive changes that the result isn't Qmail any more. And there's nobody in charge, taking responsibility for those changes. Pick a patch at random from qmail.org. Chances are it's been abandoned by its author already. Pick any three large patches. Chances are they'll break each other. Qmail's not missing from Debian because of uninteresting licensing issues, it's missing because it's broken and can't be fixed. I ran Qmail for eight years. Two of those bugs finally forced me to switch to Postfix. Luckily I never ran into the DoS bug. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian on i486
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Leonid Grinberg wrote: Hello, I have obtained the proper module for the 2.4 kernel and my ethernet card (3c509), but modprobing it results in insmod: Unresolved symbol isapnp_find_dev_Rdae0a386 Google does not find any pages with that symbol. Any ideas? My guess is you're trying to use a 3c509.o that was made with isa-pnp.o or isapnp.o, and those have to be loaded first, and they're missing. If you're using modprobe and it can't resolve that by loading isapnp, you're missing something, or you're trying to use modules from two different kernels. As I dimly recall, 3c509.o can load all by itself in a kernel made without ISA Plug and Play. But if you make your kernel with PnP, 3c509 needs it. I never did understand that part of that driver. Unlike most ISA cards, 3C509 doesn't have a fixed I/O address. The driver kind of feels around for it at load time. When I was using 3C509s I compiled it into bzImage so they would do that early in the boot time and not need to get modprobed with other activity going on. Same with ISA Pnp. As I already said, you're gonna want to make a kernel just for that machine. The stock Debian kernels are too fat and you don't have memory to waste. Skip the whole initrd thing, it'll boot faster. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian on i486
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Leonid Grinberg wrote: The machine has both a floppy drive and a CD drive, but the BIOS does not recognize the CD drive as a drive, and does not allow me to boot off of it, let alone USB. That leaves only floppies. the machine is unable to recognize the ethernet card, and I can't get much further than that The machine already runs some old distro of Linux, Any ideas? Save yourself a big hassle. Move the hard drive to a newer machine and do the base install there. Then edit etc/fstab as needed and move it back. Make a raw GRUB floppy with the new machine. Use it to boot the old one the first time, then install syslinux or grub or lilo or whatever in place. That should get you around any broken BIOS. I do laptop installs that way. That's *much* easier than fiddling around with Debian install floppies and non-booting CDs and unsupported Ethernet cards. If the newer machine has a Debian system on it already, and it fits on the old drive, you don't even need to go through the base install. Just make a file system on the old drive. You can cp -a a Debian system from one drive to another. (Remember not to copy /proc/ and /tmp/, and clean out /var/cache/apt/archives first.) Then edit etc/fstab, boot the copy with your raw GRUB floppy, and uninstall everything you don't need. While you're at it, the newer machine is faster, so make a kernel for the old machine there. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Will VNC run on a laptop with 32MB? What to do with OLD laptop?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Rick Reynolds wrote: J F wrote: Will VNC run on a laptop with 32MB? WOuld be a nice remote desktop in my living room to my bedroom linux machine or windows machine. I've got an old laptop. I was wondering what the best thing to do with it is? It has no CD and only 32MB of memory. I found X windows is swapping constantly to disk if you move a window on another machine I have. Best thing to do with that 48 MB laptop is add RAM while it's still available. You're not going to be happy with the X Window System in 48 MB. I installed Sarge on a 100 MHz laptop with 64 MB + 1MB video. Fluxbox was usable with Dillo. Mozilla ran without swapping too much, but only on low-graphics pages. Abiword was usable but slow to launch. The same system with 48 MB swapped constantly, and absolutely wasn't usable. The culprit was the X server itself. My conclusion, 64 MB is the lower limit for XFree86 3.x. Same machine runs Mozilla adequately with 128 MB. At the bottom end, RAM matters more than CPU speed. I wonder if you can use a swapfile on an NFS-mounted remote partition. That might be usable with a Cardbus (32-bit) 100 Mbps NIC. I doubt it would be adequate with a PC-Card (16-bit) NIC. Also wonder if laptops of that era have USB 2.0 so you could swap on a thumb drive. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Block read resulted in short read when dumping root fs, but all else is well
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Marcin Owsiany wrote: I get the following error when dumping the root partition: [...] | DUMP: Dumping volume 1 on /srv/backup/2006-07-26_Wed_level_3/_001 | DUMP: Volume 1 started with block 1 at: Wed Jul 26 06:25:08 2006 | DUMP: dumping (Pass III) [directories] | /dev/mapper/lv-pv: Attempt to read block from filesystem resulted in short read while reading inode #2 However my filesystem seems OK. I haven't run fsck on it yet, but the fact that the next dump (on the subsequent day) runs just fine, suggests that this is not an error in the filesystem. I suspect hardware. Something's barely working, and when temperature, vibration, and unlucky bit patterns line up against it, it fails. People think of hardware as either working or not, but the reality is noise and poor signal integrity. Hardware needs operating margin to be reliable, and typical PC motherboards don't have much. Then people do silly things and make it worse. The Ultra DMA 133 interface was designed for an eighteen inch cable. That's why retail- packaged drives come with them, so you won't buy a longer one from a clueless/unscrupulous retailer and the manufacturer won't get trouble calls about it. If you use a twenty four inch cable you may get lucky for a while. I had a system that got disk errors during backup. Replaced drives and controllers, no change and they all work fine in other motherboards. Turns out the motherboard is defective, gets DMA timeouts and other errors with any PCI ATA controller in any slot. Moved the drives to the second motherboard ATA channel to get by till I can replace the motherboard; this box doesn't really need a CD drive anyway. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian on i486
In article [EMAIL PROTECTED], Leonid Grinberg wrote: The machine has both a floppy drive and a CD drive, but the BIOS does not recognize the CD drive as a drive, and does not allow me to boot off of it, let alone USB. That leaves only floppies. the machine is unable to recognize the ethernet card, and I can't get much further than that The machine already runs some old distro of Linux, Any ideas? Save yourself a big hassle. Move the hard drive to a newer machine and do the base install there. Then edit etc/fstab as needed and move it back. Make a raw GRUB floppy with the new machine. Use it to boot the old one the first time, then install syslinux or grub or lilo or whatever in place. That should get you around any broken BIOS. I do laptop installs that way. That's *much* easier than fiddling around with Debian install floppies and non-booting CDs and unsupported Ethernet cards. If the newer machine has a Debian system on it already, and it fits on the old drive, you don't even need to go through the base install. Just make a file system on the old drive. You can cp -a a Debian system from one drive to another. (Remember not to copy /proc/ and /tmp/, and clean out /var/cache/apt/archives first.) Then edit etc/fstab, boot the copy with your raw GRUB floppy, and uninstall everything you don't need. While you're at it, the newer machine is faster, so make a kernel for the old machine there. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Make an updates disk?
[This message has also been posted to linux.debian.user.] I've got three friends who installed 3.1r2 from CD. They're on dial-up. I'd like to make an update CD for them from the files in my /var/cache/apt/archives. It would be nice if they could use apt-cdrom add to bring it in. (Otherwise I'm just going to unload a tarfile in their /var/cache/apt/archives) The instructions for debian-cd seem to require that I maintain a complete Debian mirror in order to make this CD. Is that really necessary? Is there any way to make a CD that will work with apt-cdrom without reproducing the entire Debian distribution? Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: CD writer wear out?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Manon Metten wrote: Art Edwards wrote: I have had several CD burners die. They read just fine, but they stop writing after a very finite number (say tens) of CD's.r My cd writer of white label died after just two years of heavy use. At last it still was able to read but made a lot of errors during write. I have a new cd writer fitted now along with a cd/dvd writer, both of label LG (it's hard to find another one here nowadays). Both seem to work fine for the moment. I make a lot of Knoppix disks because the Green Party here gives them away. I get CD drives used or salvage or surplus, all kinds and ages. Over time I collected nine drives that stopped working: some wrote bad disks, others got read errors. Took the cases off and cleaned the lenses carefully with aqueous isopropyl rubbing alcohol on cotton swabs. Eight of the nine work perfectly now. Try it. The key is to remove the extra alcohol before it dries and leaves a residue. Use a wet swab and then a dry one. Note, even a perfectly good brand new CD-RW drive will write bad CDs if you burn at full speed. The maximum writing speed on modern CD blanks is very optimistic. Try burning at half the automatically detected speed, or 16x, whichever is slower. You will get a much higher yield. Your disks will be readable in marginal drives that can not read disks burned at full speed. Also, don't waste your time with CD-RW media. I have tried several brands and none erases well. Second burn yield is under 50%. Third burn is near zero. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
MP3s from an analog signal ?
[This message has also been posted to linux.debian.user.] A friend of mine uses a user friendly Windoze box to record (with permission) radio shows in background, on a predetermined schedule. It's the last thing he still needs Windoze for. We'd like to replace that setup with a cron-driven shell script on his Sarge box. According to the Linux Sound HOWTO, which seems unmaintained, I can get raw samples from /dev/dsp0 with dd. This works on about half the sound cards/motherboards I've tried. I suspect the rest have some kind of mixer issue, or maybe the codecs just aren't wired right on those motherboards. But once we have that raw samples file, we don't know how to make a usable MP3 out of it. Searching with Google, I find a lot of people asking the same question, and people barking one-word answers at them, that don't lead anywhere. Is there anything in Debian that can take a raw samples file and turn it into an MP3? Has some other form of documentation replaced the collection of HOWTOs at TLDP.org? Thanks. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: subscribing to d-u on a thread basis
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], kamaraju kusumanchi wrote: Currently on web based forums, one can request the forum software to send an email when a particular threads gets updated. These threads could be anything that you opted in (for example, threads started by a person, threads to which a person replied, threads in which a person is interested in). Is there any such feature available for d-u mailing list also? The advantage is that one need not read all the messages but only those which are of interest to him. If that is not possible, can someone recommend a debian forum which provides such a feature? I've been reading debian-user as a newsgroup, linux.debian.user. My netnews provider (newsguy.com) carries it, and I see it's in google-groups now. I don't know how anyone could deal with it as an email stream. (I subscribed by email to get posting privileges, and discard the incoming stream with a procmail recipe. Seems a waste, but the gated-at.bofh.it gateway seems to be read-only.) If you just want to follow a thread now and then, use google groups. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Question
In article [EMAIL PROTECTED], Denis wrote: [three lines of some (c 0x7f) alphabet, deleted] I specialize in selling PCs. I would like to start installing Linux Debian to customers' PCs. Would you please let me know what packages i can charge customers for. Denis. I think the GNU General Public License (GPL) is clear enough on that question. You can charge them any fee they'll pay for doing them the service of copying and installing the software. You may charge them for any work you do on the software. You may charge them for tutoring and training them. The GPL was designed in hopes that you could make a living wage doing that. You are not allowed to charge any money for the software itself at all. It is not yours to sell. The software itself is free, and everyone is free to download it and figure it out on her own. You must tell your customer those terms. If she resells it, she has to tell her customers. That's as simply as I can put it. Have a really wonderful Yule Holiday, no matter what you call it. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
etch install partitioner fails, no CD
I burned a Dec 10 weekly Etch 1 of 15 CD to try the installer. Seems to work fine until we try the auto-partitioner. I had already prepared my 6 GB test drive with ext3 on hda1 and swap on hda2. Autopartitioner blew that away and replaced it with an LVM partition. Then it reported failure. I booted knoppix and prepared the drive again. This time I said no to autopartition and it failed, explaining there is nothing mounted as /target. So I got a shell and made /target and mounted /dev/ide/host0/bus0/target0/lun0/part1 there. Next it did not know where to look for install media. Since there was no helpful message about where the CD is supposed to appear, I installed the base system and GRUB over the network. At every step, the install menu wants to run that broken autopartitioner again. After rebooting, the base install works, except the apt-configure needs to be told the CD is /dev/hdc. I notice /etc/fstab is unconfigured here. During package selection, I have to repeatedly mount the CD; apt is not doing it automatically. Tasksel hangs and I give up on it. Apparently it doesn't know what to do when the things behind it fail. Apparently it traps signals so I have to kill it from the shell. Will use apt-get as usual. Don't give this release to a newbie, they'll be frustrated and go back to Windoze. This is about as generic a PC as you can find today. P4, APIC, SiS chip set, ATA drives, one per cable. Looks like hardware detection is hosed in the Dec 10 Etch snapshot, resulting in no /etc/fstab. If I knew which package that was I'd file a bug. Otherwise, the new base-config is a little nicer. Thanks, folks! Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
suspend? Re: [OT] good laptops
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Joe Mc Cool wrote: Yesterday I took delivery of an X30 (£250). I took it home after work, booted off a Knoppix installation CD and a couple of hours later it had installed itself fine. Video, sound, everything working fine. Could you describe what you did to get suspend-to-disk (or even suspend-to-RAM) working? I blew away a perfectly good Sarge on a Dell Latitude to try Knoppix' installer. Haven't been able to get it to suspend to disk using APM or ACPI. (Sarge used APM.) A laptop that won't suspend to disk and wake up properly is IMHO broken. I've been supporting Knoppix in my PC repair business, and my current advice to laptop owners is don't do it. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
make an updates CD?
My friend has installed a sarge workstation from the 14 CDs. With limited dial-up access he can't upgrade it. I would like to make a disk 15 for him from the contents of my /var/cache/apt/archives. Is there a script to gather a set of .deb files and produce an .iso which apt-cdrom would recognize? Do I have to go all the way through debian-cd and set up a local Debian mirror to make one supplement disk? TIA Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: make an updates CD?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Hodgins Family wrote: Good morning: Follow this link http://ubuntuforums.org/showthread.php?t=7455page=1 My friend has installed a sarge workstation from the 14 CDs. With limited dial-up access he can't upgrade it. I would like to make a disk 15 for him from the contents of my /var/cache/apt/archives. Is there a script to gather a set of .deb files and produce an .iso which apt-cdrom would recognize? Do I have to go all the way through debian-cd and set up a local Debian mirror to make one supplement disk? This isn't EXACTLY what you are looking for but it could give you some ideas. This is an Ubuntu page, I now, but the commands shouldn't be so very different from Debian (if at all!) showthread.php?t=7455 gives me some people wondering the same thing, and generating a Packages file, and creating a depository *directory*. (There's also some stuff about Synaptic, whatever that is.) That's neater than just copying all the deb files from my apt/archives to his, but it still doesn't get me something I could put on an isofs so that apt-cdrom would understand it. Make a CD which apt-cdrom will accept is the part I still haven't found. Thanks. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: New Linux worm crawls the web
[This message has also been posted to linux.debian.user,comp.infosystems.www.servers.unix.] In article [EMAIL PROTECTED], Hugo Vanwoerkom wrote: Paul Johnson wrote: Hugo Vanwoerkom wrote: Mike McCarty wrote: http://www.securityfocus.com/brief/38?ref=rss How to detect whether infection has occurred? Don't go overboard yet. Might want to read Steve Lamb's comment about this just upthread. Like Joey says, Debian Sarge with security updates avoids the problem. Yet... it would still be nice to know how to tell that there was no infection. It's misleading to call these things Linux worms. The worm attacks PHP applications. You can update Sarge every day. If one of your users is running PHP Nuke or Mambo or phpBB or Squirrel Mail, you have directories where the Web server can create executable files and run them. If your users don't maintain their PHP apps, they can have holes that let the worm create files in /tmp or /var/tmp/. If you install in the default places, the worm knows where your Mambo modules directory is. Sure, the worm wants to pull in a rootkit, and maybe Sarge with security updates will prevent the root escalation. That depends on the rootkit, and the worm. But even if it only gets UID 33 (www-data), it can pull in and run PHP code. Your box can become a spammer bot or an attack bot that way, and you can help propagate the worm to other hosts where the rootkit might succeed. I think it's a major security bug for /tmp and /var/tmp to be mounted with exec privileges. It's a major security problem for the Web server user to be able to create and run executables anywhere. I hope the Debian maintainers are going to fix it, because the PHP application community never will. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dialup modem recommendation for debian sagre?
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], John Hasler wrote: Bruno Buys writes: I moved recently, and there´s no easy/cheap broadband in the new address. So, what do I need to run a dialup modem on my debian sarge? Hook up your modem, run pppconfig, and follow instructions. Guess the best model I can find would be a hard pci us-robotics or similar... Be very careful there. There are many kinds of hardware PCI modems. Almost all of them contain some signal processing hardware, which is what distinguishes them from the typical software modem. But hardly any of them contain hardware which emulates the traditional 16550 serial port or the defacto standard Hayes AT command set behind it. Almost all hardware modems for sale on Ebay will not work with Debian out of the box. There are many 3Com/USR hardware PCI modem types. Only one of those actually emulates 16550 and Hayes ATDT and it appears at a strange serial port location. Knoppix won't find it. Sarge won't find it. You will have to research and fiddle around and try nonfree drivers and it is a huge pain in the ass. 16550+Hayes emulation just never caught on in the PCI modem market, and it's a damn shame. (I suspect it was because of Hayes' stupid software patent on their plusplusplus delay feature. 16550 emulation isn't expensive any more, and helps performance, but the Hayes royalty was expensive. Yet another unexpected harmful effect of software patents.) The Ebay vendors do not know this. Most of them do not care either. They know nothing but Microsoft, and in that world all modems come with a CD which automatically installs an appropriate driver. They will sell you a hardware modem that works with Linux! and throw in a CD with some random driver they downloaded from linmodems.org but never tested, and you will never get it working. No. The best you could do is buy an old USR external modem on eBay. (there are winmodems available, but I´m avoiding these). Are they debian friendly? Some winmodems can be used with Linux but they are not at all friendly. Winmodem is a 3Com trademark. There are many non-3Com hardware modems on Ebay that are useless crap because nobody has written the part of the Linux driver to do the AT command set. Web sites like linmodems.org and start.at/modem are years out of date, and unmaintained, and full of wrong recommendations and instructions which DO NOT WORK in the general case. They are also full of anectdotes (it works for me!!!) from people who just happened to get lucky. I assure you, for every it works for me!!! cluebie there are ten frustrated would be Linux users who gave up on Linux and went back to MS-Windows because they could not get their garbage PCI hardware modem working. Save yourself hours of frustration. PCI modems are an MS-Windows-only thing. Stay away. Buy an external modem for your serial port. This goes double for Card-bus. Card-bus is the PCI-like PCMCIA slot in most laptops. You will probably never get a Cardbus modem working. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: automating lynx
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Joe Mc Cool wrote: Please, I am trying to get create a cron job that will go fetch a doc file from a particular web site. I had hoped to run: lynx -cmd_log my_script http://www.siteiwant.com navigate (entering passwords etc) to the file I want and fetch it. Then I would: lynx -cmd_script my_script http://www.siteiwant.com But after logging in lynx presents me with a blank screen, not the Try it as lynx -cmd_log=my_script http://www.siteiwant.com You're missing the equals between -cmd_log and its parameter. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: IDE controller card recommendation
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Roberto C. Sanchez wrote: I have a stack of old IDE hard drives that I want to use in a RAID configuration. The server's case has plenty of room for drives and a number of PCI expansion slots. I would like to get two PCI IDE controller cards to connect all the drives with. The server has an IDE CD-ROM, leaving 3 open slots in the on-board controllers. If I get two more cards (eah with two connectors), I should be able to connect 8 more hard drives. I would like to use inexpensive cards (~$15-$30) since I will be setting up software RAID and this is an older server with older drives. I have experience with certain Promise cards under Linux (and I am not happy with that experience).Personally, I am just looking to find out what cards people are using and how happy/unhappy they are with said cards. I've been buying the Promise ATA-66 cards at the Computer Recycling Center in Sunnyvale for a dollar or two. Linux works fine. BIOS boots. So if you're really on the cheap, call them. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Name resolver problem after upgrade from woody to sarge
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Mats Bengtsson wrote: Hi, I have just upgraded my machine at home from woody to sarge and now the name resolution doesn't work anymore. To be more precise, programs such as Mozilla or lynx don't manage to lookup any names, but i can use host or dig or nslookup (or even socket.gethostbyname(...) in Python) without My impression is Mozilla only reads /etc/resolv.conf once, when it's launched. If you change resolv.conf, you have to exit mozilla and start a new one. I haven't seen Lynx have that problem. problems. Also, once I have looked up a name using host, for example, then Mozilla can also find it, at least for some minutes of time. In that case, I would look for a problem with the forwarder (dns cache) that you're pointing at with resolv.conf. I have tried looking in different documents and HOWTO:s and Google but haven't found anything really related. Also, I don't know the inner workings of name resolution in Linux, so I may have missed something obvious. Some facts on my setup: - I use an ADSL connection with a combined ADSL modem/router configured as a DNS relay. - ifup seems to detect this setup correctly and writes nameserver 192.168.1.1 in /etc/resolv.conf Is there really a name server at 192.168.1.1? Does it work? Try host kernel.org 192.168.1.1 and see. - Since I only have 2 machines connected to the router, I haven't setup any DNS for my private network and have simply added a line 192.168.1.2 MatsPC in /etc/hosts At the moment, I don't really have any clue on where to search, so any hints are welcome. While you're working on your home network, routers, resolv.conf, ifup/down, etc, use a reliable name server on the public Internet. After everything else is working, *then* worry about a local forwarder. Maybe you don't need one. Don't assume the name server you're told about by DHCP actually works. There are still a lot of name servers around that answer all queries from everywhere. They'll die (or be secured) if somebody publishes software that uses them by default, but short of that the potential for abuse isn't too bad. Pick half a dozen of your favorite ISPs and look up their name servers in whois. Try all the listed name servers. Some will work, some won't. Chances are you'll find at least two that answer queries about unrelated domains as well. A certain large telco on the West coast of the US, that doesn't have any network tech support for its residential DSL customers, has one. Perhaps the last technical person forgot to shut it off when they laid her or him off. The largest cable TV operator in the US, who happens also to be #2 on the Spamhaus.org list, has another. And the #1 spam supporter worldwide (see spamhaus top 10) has two more. Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to create an iso BOOTABLE
[This message has also been posted to linux.debian.user.] In article [EMAIL PROTECTED], Alban Browaeys wrote: Le Sat, 04 Jun 2005 02:49:26 -0700, belahcene abdelkader a écrit : I have the knoppix CD, NOT an iso format, I want to create an iso BOOTABLE from the expanded directory. How to do it please. The original CD is bootable this is the content of the CD: KNOPPIX (directory) ,autorun.bat, autorun.inf, autorun.pif, boot (directory), cdrom.ico ,index.html Building Your Own Live CD http://www.linuxjournal.com/article/7246 it also explains how to tweak which programs you want in your knoppix. The procedure in that article is for Knoppix = 3.7. The file /cdrom/KNOPPIX/KNOPPIX was mounted cloop and most of / was symlinks pointing into it. / itself was ramdisk, created at boot time. Starting with Knoppix 3.8, / is a UNIONFS with some directories from the cloop-mounted KNOPPIX file and others as before, in ramdisk. It's a great article, but does it still work with Knoppix 3.7? I haven't seen any remastered 3.8.2's yet, and haven't had time to do it myself. Things are missing from article/7246, and from the similar discussions in _Knoppix Hacks_ and the knoppix.net forums. After you unpack KNOPPIX into a directory on your ext3 FS, note the size with du -s . Now apt-get {update,upgrade} and make your little tweaks. Notice how much the du -s . size has grown. If you make a new KNOPPIX with create_compressed_fs, it's too big now. Hack #95 Trim the Fat suggests removing packages and running apt-clean, but it is never enough. There's more fat, but where? Can you safely remove /var/cache/apt/pkgcache.bin ? Cameron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: where is it?
Now that two people gave the same advice, I'm going to dissent. The instructions in install.txt are adequate if you have Debian on CDROM. Not if you download to disk yourself. I downloaded what I believed was an accurate copy of ftp.cdrom.com:/pub/linux/debian/bo onto a spare hard drive and mounted it on the machine being installed. But I could never find an answer that would satisfy the ambiguous prompt that wants to know where to find the Debian distribution already mounted. (Is there a missing top level directory? What's its name?) I installed from CD ROM later with no problems. Install.txt glosses over this, I suppose because most people are installing from the CD or over the network. I don't know how to fix it or I would have sent in a patch. For now, we shouldn't be telling newcomers to Linux that install.txt is all the info they'll need to install Debian. The ones who try to download it may be disappointed. Cameron -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Stupid Question: Striping Dos ^M From Texts
tr -d '\r' dosfile unixfile removes all ^Ms, even if they are not at the end of the line where MSDOS seems to put them. tr(1) is small and fast. perl -p -i.bak -e 's/\r$//;' dosfile renames the dosfile dosfile.bak and writes the corrected output in dosfile. The $ anchors the search pattern to the end of the line. Omit the -i.bak if you don't want to save the old file. Cameron http://www.rahul.net/cameron/homepage.shtml -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .