Re: Sarge on IBM Thinkcentre A50 - instability
On Tue, 2005-01-04 at 08:24 -0600, Kent West wrote: I'd leave the machine in console mode (no X) for that period of time, to see if the machine gets sluggish. If not, you know it's somehow related to running X. I have left it on console mode, and the problem reappeared. It therefore seems to be unrelated to X. The machine became very sluggish. The clock lost time. Repeated backspace (key held down) wouldn't work on the keyboard. Issuing the reboot command evetually succeeded in rebooting the machine, but it took fifteen minutes or so. George Karaolides -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Sarge on IBM Thinkcentre A50 - instability
Greetings, I am having problems running Debian Sarge on IBM ThinkCentre A50 PC's. I have made an installation of the Desktop Environment from tasksel using Debian Installer RC2. At some point the machine starts to behave in an errratic manner: the system clock jumps forward and back, and the machine responds in a sluggish and erratic manner. It is sometimes possible to run top under these conditions, and it is then seen that there appears to be no process consuming undue amounts of CPU or memory. The machine becomes unuseable and will usually not even reboot; only a hard reset or power cycle will bring it back, but after a while the problem reappears. I have tried two ThinkCentre A50 models, 8084-7GG and 8085-78G and had the same problems on both. On the 8085-78G which has an AGP slot besides the on-board Intel 865 graphics chipset, I also tried using an nVidia GeForce FX5200 graphics card, but had the same problem. I have tried disabling all the power management features without result. I have tried kernel images 2.4.27-1-686 and 2.6.8-1-686, and had the same problem with both. I am now at a loss as to what to try next (besides a different OS). Anyone replying please CC me as I am not subscribed to this list. -- Best regards, George Karaolides System Administrator OTEnet Telecom 20 Ayias Paraskevis St. CY-2002 Strovolos, Nicosia, Cyprus tel: +357 22 69 fax: +357 22 455686 www.otenet-telecom.com Confidentiality notice and disclaimer applies: http://www.otenet-telecom.com/Disclaimer.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sarge on IBM Thinkcentre A50 - instability
On Tue, 2005-01-04 at 08:24 -0600, Kent West wrote: but after a while the problem reappears. How long a while? Minutes? Hours? Days? Hours. I'd leave the machine in console mode (no X) for that period of time, to see if the machine gets sluggish. If not, you know it's somehow related to running X. Then I'd try a different window manager/environment. Good thinking, I'll do that and report. Thanks, George Karaolides -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
LVM on SW RAID for sarge - success
/ /mnt Mount the RAID1 for /boot and copy over /boot: mount /dev/md0 /mnt/boot cp -avx /boot /mnt Mount the /home, /usr and /var LV's and copy over filesystems: mount /dev/vg0/home /mnt/home mount /dev/vg0/var /mnt/usr mount /dev/vg0/usr /mnt/var cp -avx /home /usr /var /mnt Then mount the /var/log and /usr/local LV's: mount /dev/vg0/usrlocal /mnt/usr/local mount /dev/vg0/varlog /mnt/var/log Copy over /usr/local and /var/log: cp -avx /usr/local /mnt/usr cp -avx /var/log /mnt/var The system is now copied over to LVM-on-RAID. It should be made bootable. Before this can be done, /dev and /proc should be available when chrooting into the target system: mount -t devfs devfs /mnt/dev mount -t proc proc /mnt/proc Edit /mnt/etc/fstab to reflect the filesystem sructure, replacing /dev/sda* with /dev/vg0/* (not forgetting /dev/md0 for /boot) Make a new LVM-on-RAID capable initrd image in the target system: mkinitrd is clever enough to figure out what's needed from the new fstab file. chroot /mnt mkinitrd -o /boot/name-of-existing-initrd-image Edit /mnt/boot/grub/menu.lst and change the kernel boot option for the root filesystem from /dev/sda* to /dev/vg0/root. The line in my file looked like: kernel /vmlinuz-2.4.26-1-686-smp root=/dev/vg0/root devfs=mount noapic ro single (I gave it noapic because of APIC problems with my motherboard killing the ethernet card, you probably don't care about that) Then run grub in the target system: chroot /mnt grub On the grub command line, tell grub to use one of the partitions in the RAID1 for /boot to read data, e.g. second disk, first partition: root (hd1,0) Then tell it to write an MBR on the second disk: setup (hd1) And also on the third disk: setup (hd2) quit It should now be possible to boot into the initial single-disk system when booting from the first disk, and into the degraded LVM-on-RAID system when booting from the second or third disks; use the SCSI BIOS to select. Once it is verified that the LVM-on-RAID system boots and functions correctly, the first disk can be repartitioned like the second and third ones and the partitions incorporated into the RAID arrays: mdadm --add /dev/md0 /dev/sda1 mdadm --add /dev/md1 /dev/sda5 Make the system also bootable from the first disk: grub root (hd1,0) setup (hd0) quit Treat yourself to your favourite beverage while watching the array reconstruction :) watch -n3 cat /proc/mdstat Enjoy redundant, high-preformance, instantly-online-resizeable filesystems: lvextend -L+how-much-more-space-do-you-want[MG] /dev/vg0/filesystem; resize-reiserfs /dev/vg0/filesystem Snapshots: Note that snapshots of journalled filesystems like reiserfs do not work unless a kernel patch is applied and the kernel recompiled. It is possible to generate the patch by downloading the Debian package of the kernel source, and the Debian source package for lvm10. Anyone interested in snapshots drop me a line and I'll tell you how to do it. -- Best regards, George Karaolides System Administrator OTEnet Telecom 20 Ayias Paraskevis St. CY-2002 Strovolos, Nicosia, Cyprus tel: +357 22 69 fax: +357 22 455686 www.otenet-telecom.com Confidentiality notice and disclaimer applies: http://www.otenet-telecom.com/Disclaimer.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
networking problem
Hi people, I have a networking problem on several machines running woody and sarge all with the same hardware: Abit BP-6 dual celeron motherboard w. celeron 466 processors Intel etherexpress pro 100 ethernet cards All machines run Linux kernel 2.4.25 Problem: networking ceases to funcion and only a machine reboot will restore function. Stopping networking, unloading the eepro100 module, reloading and restarting does not work. Problem seems to appear under conditions of high load. Attempted to cure problem by turning off sleep mode, as suggested by the output of the eepro100-diag utility. No difference. Relevant /var/log/messages excerpt: Apr 15 11:15:45 stephano kernel: unexpected IRQ trap at vector 7d Apr 15 11:16:20 stephano kernel: NETDEV WATCHDOG: eth0: transmit timed out Apr 15 11:16:20 stephano kernel: eth0: Transmit timed out: status f048 0c00 at 11611702/11611762 command 000ca000. Apr 15 11:17:14 stephano kernel: NETDEV WATCHDOG: eth0: transmit timed out Apr 15 11:17:14 stephano kernel: eth0: Transmit timed out: status f048 0c00 at 11611762/11611822 command 0001a000. Please cc. any replies to me. Thanks. -- Best regards, George Karaolides System Administrator OTEnet Telecom 20 Ayias Paraskevis St. CY-2002 Strovolos, Nicosia, Cyprus tel: +357 22 69 fax: +357 22 455686 www.otenet-telecom.com Confidentiality notice and disclaimer applies: http://www.otenet-telecom.com/Disclaimer.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Scaling Debian - SSI Linux
Greetings all, Thanks to Debian and this list, I have grown in the past four years from Linux hobbyist to smalltime sysadmin with a brokerage firm, to sysadmin at an ISP. One of the best things about my last job has been that the bosses placed their confidence in me and Debian. Backed by the quality of the software and the support of the lists I have been able to meet the challenges. Now the company I work for has grown to be a voice and data services provider poised to compete at a national level. And decisions have to be made about the server OS and hardware platform that will meet the challenge. We currently run Debian on a bunch of servers, with services distributed randomly among machines. We are set up so that we can restore a service manually on another machine if the machine providing the service goes down, so availability isn't too bad. But we're now thinking about scalability. The bosses are happy to renew their confidence in me and Debian, but their question is can Debian scale to meet the expansion, or should we make the decision to go to heavy metal UNIX? And I have the same question. I naturally want to stick to Debian, which has rocked so far. But I don't think we can continue to run JBOS (Just A Bunch Of Servers) for much longer. We need to go to an available and scalable solution. I've been thinking of Single System Image Linux (SSI Linux) for some time. This will have the advantage of leveraging the currently available dozen or so servers, avoiding the purchase of heavy metal. It will also allow addition of resources as demand requires, rather than initial overprovisioning as is done with heavy metal. Has anyone on the list gone to production with Debian and SSI Linux? Does anyone on the list have any other suggestions re. Debian scalability and availability? -- Best regards, | George KaraolidesOTEnet Telecommunications Ltd., | | System Administrator 2nd Floor, 20 Agias Paraskevis St., | | tel: +357 22 45 65 00 Strovolos, Nicosia CY 2002, | | email: [EMAIL PROTECTED] Republic of Cyprus. | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Woody+initrd+raid1+boot = :-(
Hi, Before we go into your message, which kernel version are you using? On Tue, 18 Jun 2002, Thomas -Balu- Walter wrote: # Better use the first BIOS ID for the disk we'll boot from; # we'll use the BIOS setup to boot from this disk=/dev/hdc bios=0x80 What exactly does that do? When you install an MBR on a disk using lilo, lilo will use the BIOS ID currently lilo thinks is currently assigned to the disk. But when you select the disk to boot from using the BIOS setup, you will change the order of the BIOS disks so that the one you boot from has the first ID (0x80); that is how the BIOS selects which disk to boot from first. So if you are installing an MBR to a disk you will boot from by selecting it with the BIOS setup, you need to tell lilo to use the first BIOS ID (0x80). snip initrd stuff I did not even have to change the partition type to autodetect, since raidstart and such are available in initrd. Maybe, but it's always a good idea... I am stuck and booting while simulating a broken secondary IDE controller. Have you got it all working normally before simulating failures? Can you boot into the RAID from either disk, using the BIOS setup to select? If I take off the master cable (powered off of course) I can boot into the raid - and get the following messages: md: could not lock [dev 03:03] sero-size? Marking faulty md: could not import [dev 03:03], trying to mount raid nevertheless md: autorun ... former device [dev 03:03] is unavailable, removing from array ... So it just ignores hda and boots into the remaining mirrored hdc. But after rebuilding the raidset and removing the cable to the secondary controller i get: ide0 on 0x1f0-0x1f7, 0x 3f6 on IRQ14 md: could not lock [dev 16:03] sero-size? Marking faulty md: could not import [dev 16:03] md: autostart [dev 16:03] failed /dev/md1: invalid argument ... Kernel panic, because of missing root. To fix I've tried to make the partitions raid-autodetect That could have been a reason, but since it didn't work... created a new initrd (because the old one still had failed-disk in it) Er, no, the initrd contains kernel modules and possibly user-space utilities but there's no way the initrd can actually know about the state of the RAID arrays so that couldn't have been it. and finally added disk=/dev/hda bios=0x80 disk=/dev/hdc bios=0x80 to lilo.conf. This also couldn't have been it; the system boots from the installed MBR, locates the kernel image and initrd and attempts to boot it, and then attempts to mount the correct root device. That is the end of lilo's jurisdiction; the fact that /dev/md0 starts when /dev/hda is down but doesn't start with /dev/hdc is down must be something to do with RAID: the array itself (likely), the way the system starts RAID (less likely) or the kernel RAID stuff. with no luck - any last hint before it finally works as expected? Balu | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Woody+initrd+raid1+boot = :-(
Hi Eduard, On Wed, 19 Jun 2002, Eduard Bloch wrote: The current md driver is broken-by-design in the autodetection issue. Broken-by-design? So there's a reason for this? You have to compile the driver into the kernel and mark the partition as type 0xFD to make it autodetecteable. Since this is what I've always done since going Debian a year ago, I haven't noticed this. IMHO this is the simplest solution. When using modules, you will be forced to reconstruct the array from a userspace utility. I suggest, you install mdadm Thanks for the pointer to mdadm. When the first disk fails, you need to know before the second disk fails and the array is irretrievably lost... mdadm has monitoring and email notification; good stuff. add mdadm to /etc/mkinitrd/exe, then edit /usr/share/initrd-tools/linuxrc and add something like /sbin/mdadm -R -A /dev/md0 /dev/hdx /dev/hdy in the first lines. There should be an easier way of modifying the start-script of initrd, I am just writting a wishlist bugreport about that. And all this just so you can compile the RAID stuff as modules rather than in the kernel? Sheesh... By the way, have a look at my Unofficial Kernel 2.4 Root-on-RAID and Root-on-LVM-on-RAID HOWTO: http://karaolides.com/computing/HOWTO/lvmraid/ It's probably full of mistakes/omissions/inaccuracies, but since I found that the current official HOWTO's are a bit out of date, I decided to document my own experiences with the recent versions of lilo and add the LVM stuff as well. I have also successfully built a root-on-LVM (no RAID) woody machine (a laptop, actually) using your LVM-and-RAID woody extdisk (thanks!) and the bf2.4 woody floppies. It booted just fine with the stock 2.4.18 kernel that comes with the bf2.4 floppies, so I intend to document that too in the HOWTO. I'd also like to draw your attention to a post of mine from yesterday re. woody on LVM which wasn't answered; I'd really appreciate some input: http://lists.debian.org/debian-user/2002/debian-user-200206/msg02783.html Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Woody+initrd+raid1+boot = :-(
Hi, The problem is that the existing HOWTO's on root-on-RAID refer to lilo versions earlier than 21.8. Debian woody comes with 22.2, which is easier to configure for root-on-RAID, once you find out how... You should have, in /mnt/etc/lilo.conf (better leave lilo.conf on your original installation alone): --- Begin lilo.conf extract --- # Better use the first BIOS ID for the disk we'll boot from; # we'll use the BIOS setup to boot from this disk=/dev/hdc bios=0x80 # /boot is on /dev/md0 boot=/dev/md0 # Write an MBR only to /dev/hdc for now raid-extra-boot=/dev/hdc # The root fs is on /dev/md1 root=/dev/md1 --- End lilo.conf extract --- NOne of all that messing around with partition geometry specs that you needed to trick the earlier lilo into booting from a RAID-1 /boot. Once you've done this, then do chroot /mnt /sbin/lilo You don't need initrd if the RAID stuff is compiled into your kernel. If the RAID stuff is in the form of modules, then to make the initrd for your new system, just do chroot /mnt mkinitrd -o /boot/initrd to create an initrd image /boot/initrd (or your choice of filename). mkinitrd is good at detecting what modules to include; I haven't yet had an instance where I needed to tell it what to include. If you're using initrd as above, you will obviously need to include a line initrd=/boot/initrd in /mnt/etc/lilo.conf. Reboot your machine, go into BIOS setup and set your BIOS to boot from the secondary master (/dev/hdc). With any luck, you will boot into the root-on-RAID system with the degraded arrays. Check that everything is OK. Then you can add the partitions on /dev/hda to the arrays. Make sure the partition type is set to 0xFD (Linux RAID Autodetect), and edit your /etc/raidtab to change the failed-disk entries to raid-disk. You can then raidhotadd the partitions: raidhotadd /dev/md0 /dev/hda1 raidhotadd /dev/md1 /dev/hda3 Now you have a full RAID system. It's a good idea to also duplicate the MBR's to make sure that if one disk fails, you can boot from the other by changing the BIOS setup. This is another instance where the new lilo is much easier than the old one to set up. Just make the following changes to /etc/lilo.conf: --- Begin lilo.conf extract --- # Better use the first BIOS ID for all disks; # we'll use the BIOS setup to choose which one to boot from. disk=/dev/hda bios=0x80 disk=/dev/hdc bios=0x80 # /boot is on /dev/md0 boot=/dev/md0 # Write an MBR to all disks raid-extra-boot=/dev/hda,/dev/hdc --- End lilo.conf extract --- Now run /sbin/lilo. Done; you have a fully redundant disk system, down to the MBR's. You might want to look at my Unofficial Kernel 2.4 Root-on-RAID and Root-on-LVM-on-RAID HOWTO at: http://www.linustech.com.cy/linux/HOWTO/lvmraid/ or at http://karaolides.com/computing/HOWTO/lvmraid/ Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Woody lvm10
Hi, I've built a couple of woody machines with the root fs on lvm-on-software-RAID. I used lvm10 version 1.1rc1-2 for this without mishap. I have read the following article, xhich states that this version of lvm10 has various issues: http://www.debianplanet.org/article.php?sid=684mode=threadorder=0 Apparently, the issues have been addressed by including version 1:1.0.4-4 (an earlier version) in woody. Reading the bug reports on lvm10, however, I see that some people have had their systems rendered unusable when going from 1.1rc1-2 to 1:1.0.4-4: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=148909repeatmerged=yes Does anyone have any suggestions as to how to make sure the system isn't broken when changing from 1.1rc1-2 to 1:1.0.4-4? A suggestion in the replies to the bug reports is to do vgexport in 1.1rc1-2 and vgimport in 1:1.0.4-4, but it didn't seem to work for the person who posted the bug. Suggestions, anyone? Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
DNS resolution problem
Hi, I have a really strange DNS resolution problem. I have set up and configured a Debian woody box as a gateway and firewall for an internal network connected to the Internet via ADSL. I use the Debian ipmasq package for this. The ISP assigns an IP address to the ethernet interface connected to the ADSL network via DHCP. A private IP address is assigned. The ISP presumably runs a NAT gateway to the public Internet. The ISP also uses DHCP to give you the IP address of the nameserver you're expected to use (they block UDP port 53 to every other address so you can only use their nameserver). This is also a private address (192.168.4.5). Now here's the problem, and it's a really freaky one. All hosts behind the Debian box can access the Net just fine, using nameserver 192.168.4.5 to resolve names. But the Debian box itself can't resolve names. And now's the really freaky part. Using dig to query nameserver 192.168.4.5 from the Debian box as in: dig @194.42.142.4 www.google.com works just fine! But putting 194.42.142.4 in /etc/resolv.conf and doing dig www.google.com doesn't work! How can a nameserver which does accept to do recursive resolution for me when I query it directly refuse to do so when I put it resolv.conf? Any ideas? Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: DNS resolution problem
Sorry to be replying to my own post, but there was a typo in my email. I meant to say: --- Begin erratum --- dig @192.168.4.5 www.google.com works just fine! But putting 192.168.4.5 in /etc/resolv.conf and doing dig www.google.com doesn't work! --- End erratum --- This makes the point of the message: a nameserver which will do er3cursive resolution for me when queried using dig will not do so when placed in /etc/resolv.conf. Sorry for the typo and best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | -Original Message- From: George Karaolides [mailto:[EMAIL PROTECTED] Sent: Wednesday, June 05, 2002 10:51 AM To: debian-user@lists.debian.org Subject: DNS resolution problem Hi, I have a really strange DNS resolution problem. I have set up and configured a Debian woody box as a gateway and firewall for an internal network connected to the Internet via ADSL. I use the Debian ipmasq package for this. The ISP assigns an IP address to the ethernet interface connected to the ADSL network via DHCP. A private IP address is assigned. The ISP presumably runs a NAT gateway to the public Internet. The ISP also uses DHCP to give you the IP address of the nameserver you're expected to use (they block UDP port 53 to every other address so you can only use their nameserver). This is also a private address (192.168.4.5). Now here's the problem, and it's a really freaky one. All hosts behind the Debian box can access the Net just fine, using nameserver 192.168.4.5 to resolve names. But the Debian box itself can't resolve names. And now's the really freaky part. Using dig to query nameserver 192.168.4.5 from the Debian box as in: How can a nameserver which does accept to do recursive resolution for me when I query it directly refuse to do so when I put it resolv.conf? Any ideas? Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
DNS resolution problem - SOLVED (not quite)
Hi, Further to my previous message, the DNS resolution problem has not been quite solved but has been found to be due to an outside cause. I have checked using iptraf, and the nameserver I am querying on 192.168.4.5 replies from a different (public) IP address! dig will buy this, but the resolver libraries (obviously) won't. I will be taking this up with the people (hominids?) who run this ISP. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Permission denied?
Hi all, I am running a woody server with the root fs (reiserfs) on software RAID1. Kernel is 2.2.20 compiled with Debian make-kpkg from the Debian package of the kernel source. The server suddenly stopped working and I was unable to log in from the console. Attempting to reboot failed with No init found. I booted the woody bf2.4 install floppies (that have reiserfs capability) and used Eduard Bloch's RAID-and-LVM extdisk floppy to restart the arrays. I successfully mounted the filesystems under /target BUT /lib and /etc were missing from the root filesystem. Doing ls /target returned: /target/etc: Permission denied /target/lib: Permission denied normal listing followed I copied the root filesystem onto a single disk partition on a fresh disk using dd, and ran reiserfsck --rebuild-tree on that. That got me /etc back, but what appears to be /lib is in /lost+found. There are files and directories in the top level named with numbers. The good news is that only the files and directories in the top level are named with numbers; it appears that files in the lower levels have retained their names. Any suggestions as to - How not to have this happen again? - How to get the names in the top level of /lib back? Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 89 87 28 31 Evagorou Avenue, Office 32, | | web: www.linustech.com.cy Nicosia CY 1066, | | email: [EMAIL PROTECTED] Republic of Cyprus. | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
ipmasq logging
Hi, I'm running woody with kernel 2.2.20 and using the ipmasq Debian package to set up packet filtering. I find that on woody, ipmasq (well really, ipchains as ipmasq calls ipchains with the -l option) logs to the console as well as /var/log/messages which is a big pain; it makes the console unusable. Any ideas as to how to make it log only to the screen? | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 89 87 28 31 Evagorou Avenue, Office 32, | | web: www.linustech.com.cy Nicosia CY 1066, | | email: [EMAIL PROTECTED] Republic of Cyprus. | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Woody install floppies
Hi all, I have had a problem with the i386 woody floppies (3.0.22-2002-04-03, flavour bf2.4)? The problem is that when rebooting for the first time after installation of the base system, the base system configuration environmenst starts (as normal) but gets stuck in a loop, presenting the same three configuration screens (configure keymap, enable md5 passwords, and enable shadow passworsd) no matter what answer is given in each screen. Has anyone else had this? Any pointers? Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 89 87 28 31 Evagorou Avenue, Office 32, | | web: www.linustech.com.cy Nicosia CY 1066, | | email: [EMAIL PROTECTED] Republic of Cyprus. | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Woody install floppies - SOLVED
On Thu, 9 May 2002, Colin Watson wrote: Check the list archives for debian-user for the last couple of days. Lots of people have reported it, and somebody posted a workaround pending the fixed base-config getting into woody. Thanks for the pointer. It led me to a solution: http://wiki.debian.net/DebianWiki/DebianWiki/WoodyNetinstBaseconfigLoop The solution given there is a user-friendly script to move base-config_1.33.17_all.deb out of the way (in /target/var/cache/apt/archives) while downloading, and install base-config_1.33.18_all.deb instead. If you're installing from the 'net on a slow connection, there's plenty of time to do this manually... I did try to search the debian-user archives and the bug-tracking system but couldn't find references; says a lot about my searching skills, I guess. I've also used Eduard Bloch's raid and LVM extdisk: http://people.debian.org/~blade/install/ So now I have an installation of woody with the root fs on LVM. I'll try root-on-LVM-on-software-RAID next... Thanks again, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 89 87 28 31 Evagorou Avenue, Office 32, | | web: www.linustech.com.cy Nicosia CY 1066, | | email: [EMAIL PROTECTED] Republic of Cyprus. | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Woody net base install from local mirror
Hi all, With Debian potato, it was fairly trivial to put the base2_2.tgz file in a webserver under a directory $DOCUMENTROOT/debian/dists/potato/main/disks_i386 and do a net install of the base system from a local mirror. Now with woody, the install system seems to look for debian packages where my local partial mirror (created with apt-move) doesn't have them. My local mirror works for the main installation after the base system is installed, but not for the base system. Does anyone know how I can set up a local partial Debian mirror so the base system can be installed from it? Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 89 87 28 31 Evagorou Avenue, Office 32, | | web: www.linustech.com.cy Nicosia CY 1066, | | email: [EMAIL PROTECTED] Republic of Cyprus. | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Moving /var to another drive
On 12 Mar 2002, Shri Shrikumar wrote: On Tue, 2002-03-12 at 15:42, Franηois Chenais wrote: And what happens if the /var/log and /var/run dirs that can change during the tar ? Franηois Go into single user mode telinit 1 and then tar. I actually use cp -a which seems to preserve all the required attributes. Point taken guys, thanks. So go into single-user mode before doing the tar. Of course there are about a million ways of transferring filesystems (cp, tar, cpio, dd, rsync, dump...). That's the joy of UNIX, but you can get fancy above and beyond the call of duty. One guy I learnt a lot of UNIX from told me to do, for example to transfer /var (command line incomplete and probably full of mistakes as I can't really remember all the correct minutiae): find /var --print | cpio -pdum (plus other cpio syntax) Needless to say, that had me running for the tar manpage in no time at all... ;) In my humble opinion, the choice of transfer method in a case as simple as that posed by Andrew must be down to personal taste/habit/familiarity/convenience rather than any technological consideration. Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Re: kernel-2.4.17
On Tue, 12 Mar 2002, Gerard Robin wrote: About lilo : I use lilo to boot on HDD1 on which win98 and potato 2.2r2 are installed. And on HDD2 potato - woody is installed.(I experiment ) I don't know if lilo can accommodate the three systems ? Yes, in several ways. It's possible to have lilo installed in the MBR of hdd1 and having three boot choices: - win98 on hdd1 - potato on hdd1 - woody on hdd2 However, I personally don't like including boot entries for a system on a separate disk in the MBR of another disk. If I have systems on different disks, I prefer to choose between them using the BIOS setup. This gives you the choice of removing either disk and not affecting the other system, if your disks are IDE. The disk will also work on another machine (provided the rest of the system is compatible) if installed in the same position in the IDE bus. This is because the IDE disk devices always point to the same disk: /dev/hdaMaster disk, primary bus /dev/hdbSlave disk. primary bus /dev/hdcMaster disk, secondary bus /dev/hddSlave disk, secondary bus If you're on SCSI it's another ballgame as if, for example, hdd1 has SCSI id 0 and hdd2 has SCSI ID 1, then hdd1 is /dev/sda and hdd2 is /dev/sdb, but if you remove hdd1 without installing another disk with SCSI id 0, hdd2 will become /dev/sda and you'll need to change the entries in /etc/fstab as well as lilo.conf. Most recent BIOSes give you a decent flexibility of setting up the boot order. With lilo correctly configured in the MBR of both disks, you will have the following choices: - BIOS set to boot from hdd1 first: - win98 - potato - BIOS set to boot from hdd2 first: - woody So leave hdd1 alone. Boot into your woody system on hdd2 and edit /etc/lilo.conf. Suppose your woody disk is the slave disk on the secondary IDE BUS (/dev/hdd). Include the following entries: --- begin lilo.conf excerpt --- # Inform lilo about the disk on which we'll be installing an MBR disk=/dev/hdd # The disk we'll be installing to will have the first # BIOS disk id (0x80) when we try to boot from it, # because we'll change the BIOS disk order with the BIOS setup bios=0x80 # Install an MBR on this disk boot=/dev/hdd --- end lilo.conf excerpt --- Save the file and do /sbin/lilo That should install an MBR on hdd2 which will be bootable when hdd2 is set as the first biit disk in the BIOS setup. Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Re: two ethernet cards without routing
On Wed, 13 Mar 2002 [EMAIL PROTECTED] wrote: Hi folk's: I've installed two ethernet cards to my debian server and they work fine. I want to give some services to one IP and other services to the other, so: How can I deactivate the routing option betwen cards? There have to be some kind of routing now because I can connect to my apache typing the two IPs even I've just one cable connected to eth0. Hi, There are some machines which, when fitted with two ethernet cards, give you access to both interfaces when only one cable is connected, but work correctly (each interface to the appropriate network) when both cables are connected. I had exactly this behaviour with an Intel server based on a Nightshade motherboard. I was completely nonplussed until someone on this list told me about it. If this is your problem, it should disappear when your cables are connected to both networks. In the meantime, check if IP forwarding is turned on in the kernel: cat /proc/sys/net/ipv4/ip_forward 1 is on, 0 is off. If it's 1, turn it off by: echo 0 /proc/sys/net/ipv4/ip_forward Check if it's turned on by sysctl at system startup. Read /etc/sysctl.conf and look for a line like: net/ipv4/ip_forward=1 If there is one, comment it out. Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Re: Moving /var to another drive
Hi Andrew, You need to be root on your machine to do this. Make two partitions on the new drive using cfdisk /dev/drive Then make a filesystem on each partition. Stick to the tried-and-tested ext2, or go for one of the new journalling ones like ext3 or reiserfs, if you're running a kernel recent enough to support them or can compile one that does. E.g. mke2fs /dev/partition Then mount the first partition e.g. under /mnt mount /dev/partition /mnt And transfer the data using tar: tar cplf - -C / var | tar xvf - -C /mnt Unmount the partition umount /mnt Mount the other one and do the same thing for /usr: mount /dev/partition tar cplf - -C / usr | tar xvf - -C /mnt umount /mnt Make sure you type the tar commands exactly as above, you don't want to go dropping the entire copnmtents of /var or /usr (or even / if you v=get it wrong) in the wrong place! Edit /etc/fstab and either add lines for /usr and /var, or change existing ones, to mount your new filesystems instead of the old ones. Here's an example: --- fstab excerpt --- /dev/sdb6 /usr ext2 rw 0 2 /dev/sdb7 /var ext2 rw 0 2 --- end fstab excerpt --- Then go to single user mode: telinit S Unmount /usr and /var: umount /usr umount /var Mount all partitions using your new fstab: mount -a Check that everything is mounted where it should be: mount And go back to your usual runlevel (usually 2 on Debian): telinit 2 That's it. Note that if /usr and /var were originally part of your root filesystem, the data will still be there but the new filesystem will be mounted on the top level directory so you won't see it. After you've successfully transferred /usr and /var out of the root filesystem, you can reclaim the space by going to single user mode, unmounting /usr and /var, and doing the following: rm -rvf /var rm -rvf /usr mkdir /var mkdir /usr Be VERY careful that /usr and /var are NOT mounted when you do this! Also note, a space or a slash in the wrong place in either of the above commands can wreck your system completely! Good luck, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. | On Tue, 12 Mar 2002, Andrew Stephen wrote: Hi My /var and /usr partitions have just run out of space and I was wondering what is the best way to copy them to a new drive that has just been installed. Any suggestions would be greatly appreciated. Regards, Andrew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Greek keyboard on Woody
Hi, Any pointers/suggestions on how to set up a Greek keyboard on woody? Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Re: Greek keyboard on Woody
On Fri, 8 Mar 2002, Pietro Cagnoni wrote: George Karaolides wrote: Hi, Any pointers/suggestions on how to set up a Greek keyboard on woody? Best regards, there's an hellenic-howto on www.linuxdoc.org. pietro. Hi Pietro, Thanks for the pointer, but I was aware of the Hellenic HOWTO. It is, however, desperately out of date: August 1997. Also, the solutions suggested therein are mostly non-standard. For example, there are dedicated software packages for the Greek keyboard, but I am aware that X comes with its own standard mechanism for installing alternative keyboard maps. I am sure that the Debian community, with its predilection for standards adjerence and up-to-date-ness, has built much better Greek support into woody's internationalisation than the non-standard, out-of-date hacks suggested in the Hellenic HOWTO. The stuff must be in the distribution somewhere; I just can't find where. Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Re: Greek keyboard on Woody
Hi, I found some stuff under www.linux.gr about setting up the Greek keyboard under X. IN the end, setting up the Greek keyboard on Woody (X 4.1.0) was as simple as making these changes to the InputDevice section: --- Section InputDevice Identifier Generic Keyboard Driver keyboard Option CoreKeyboard Option XkbRules xfree86 Option XkbModel pc105 Option XkbLayout el Option XkbOptionsgrp:shift_toggle EndSection -- So now I have a Greek keyboard that switches between Latin and Greek characters when I hit both Shift keys. But this isn't the end of the story. There are (at least) two different widely used Greek keyboard layouts: One based on the US keyboard layout and one based on the UK keyboard layout. The settings above give me the Greek keyboard based on the US layout. But being in Cyprus, where our local currency is the Cyprus Pound, the most useful keyboards here are the Greek keyboards based on the UK keyboard layout (Shift-3 gives the pound sign instead of #). Windoze supports both tyoes of keyboard. What about X? Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Re: Greek keyboard on Woody - SOLVED
Hi all, I now have a Greek keyboard based on the UK keyboard layout, as needed for Cyprus. I got that by copying the file /etc/X11/xkb/symbols/el to a file el_CY in the same directory, and hand-editing the differences between UK and US keyboards into it as obtained from the file named gb in the same directory. All include entries in the same file referring to el must also be changed to el_CY. I then changed the XkbLayout entry in /etc/X11/XF86Config to el_CY. My keyboard map still needs a little debugging but I think I'm almost there. Thanks, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Re: Sid packages in Woody
On 5 Mar 2002, Dave Carrigan wrote: Add both unstable and testing to your /etc/apt/sources.list. In your /etc/apt/apt.conf file, put APT::Default-Release testing; To install a sid package, do apt-get install packagename/unstable Apt will do the rest. Does all this nice stuff work for potato as well, or only for woody? Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Informix on Debian
Hi all, Anyone successfully installed Informix on Debian? I'm trying to install version 9.21.UC2-1 and not having too much success. Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Apache on Debian - high loads?
Hi all, Does anyone have experience of running Apache on Debian successfully with high loads? I've been asked about the possibility of running a webserver with up to five thousand concurrent users. I would, of course, prefer to do it with Apache on Debian. Would Debian be up to the task? Any pointers on what I have to look out for? Best regards, | George Karaolides Linustech Advanced Solutions, | | tel: +357 22 55 61 29 86 Ifigenias Street, 3rd Floor, | | web: www.linustech.com.cy Strovolos, Nicosia CY 2003, | | email: [EMAIL PROTECTED] Republic of Cyprus. |
Re: Building custom kernel based off stock debian kernel
Hi Nick, On Fri, 8 Feb 2002, Nick Jennings wrote: Thanks for your reccomended reading. However, just so you know, I have been building kernels for years. Actually THAT was alot easier than learning how to do all this the debian way I think. :) And building a kernel package seems to have been streamlined to avoid problems. It's even harder to build a userspace package I think. :) I was, too, but the Debian way got me completely hooked on its ease of use and I've never looked back. As you'll see, once you've got the source, re-compiling and installing can be done with just two commands. Building a kernel the Debian way: (if you don't have the source for the version you want): # apt-get install kernel-source-version # cd /usr/src/linux # bzcat kernel-source-version | tar xvf - If you want to base the configuration on a kernel you're currently running: # cp -v /boot/config-old-version /usr/src/kernel-source-version/.config Then, to build the kernel once you've got everything: # cd /usr/src/kernel-source-version # make menuconfig # make-kpkg --revision versionextraversion --append_to_version \ extraversion kernel_image NOTE: append_to_version is for woody; it used to be called flavor in potato. This is very handy. If you use a different string for extraversion every time you compile, the modules will be installed in appropriate separate directories under /lib, even if you're re-compiling from the same version of the source tree. This avoids over-writing the modules for one kernel with those for another. Same as editing the top-level Makefile and changing $EXTRAVERSION when you do things the old way. To install: # cd /usr/src # dpkg -i kernel-image-versionextraversion_versionextraversion_arch.deb Your running kernel is installed as the backup lilo opion LinuxOLD and your new kernel as default option Linux. So you can re-compile with just a make-kpkg command and install with just a dpkg command, with your running kernel as a backup lilo option. Two-command kernel recompilation and installation... Can't be bad, eh? Try it and see. Best regards, George Karaolides
Re: partition types
On Wed, 6 Feb 2002 [EMAIL PROTECTED] wrote: Dumb question is: can I change my root partition to type extended and create two logical partitions (one holding the original root stuff, the other for Linux from Scratch things) within it? Without destroying my beloved Debian system? What are the other three partitions you mention doing? Can't one of them be resized instead? Messing around with the root partition is possible, but not easy. The problem is that you need to reboot after re-partitioning the disk, and before making filesystems on the new partitions. If you touch the root partition during re-partitioning, it may be that your Debian system becomes unbootable. A kind of Catch-22 situation. You will anyway need to backup all your data before attempting anything so drastic, and most times you'll end up restoring from backup, too. There is a fairly non-intrusive way to do this by borrowing a second empty hard disk, copying the system to it, making it bootable, checking everything is there and only then re-partitioning the original disk, copying the system back, and making that bootable. Once the original disk has booted again and you've checked everything is there, the borrowed disk can be returned. I understand that hard disks don't generally lie around waiting for people to borrow them for copying systems over, but if you do happen to have access to an extra disk I'd be happy to talk you through the process. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: direct ISDN connection
On Tue, 5 Feb 2002, Vegh Karoly wrote: Feb 05 18:57:28 Call to tei 69 from onCONNECT (Data) Feb 05 18:57:28 Call to tei 69 from onINTERFACE isdn0 called by 9905 Feb 05 18:57:36 Call to tei 69 from onNormal call clearing (Private network serving local user) Feb 05 18:57:36 Call to tei 69 from onHANGUP ( 0:00:08) what is this normal call clearing? The log messages seem a bit different to what I'm used to. I'm using the Debian package of isdnutils, version 3.0-20 on a potato system. What are you using? If I'm not mistaken, normal call clearing is good news; it means the incoming call is accepted. Then the calling computer seems to hang up for some reason (If the receiving computer was hanging up the message would have been local hangup). I used to occasionally get this in automatic dial mode when both computers try to dial at the same time. Try setting the dialmode to manual on both computers: isdnctrl dialmode interface-name manual Then try dialling out on one computer: isdnctrl dial interface-name and post info. about what happens. | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: journalling filesystems comparison (was ReiserFS, ext3 (was praise to the debian gods))
My additional observations on the following journalling filesystems ext3, ReiserFS, XFS, as I use them at home as well as at the office... (at home, I have a mixed scheme of filesystems employed in partitions) Thank you for sharing your observations. I'm now trying woody with root fs on reiserfs on LVM on software RAID-1. The woody reiserfs install disks were handy for this. Haven't done much more than successfully install and boot the system, though; tests are still in the future. Just how much better is XFS than reiserfs at handling large files? As you observed, this is important for database systems. However, many databases can use raw devices rather than file systems, and with LVM you have the flexibility to change the disk space allocated to a logical volume device without re-partitioning disks. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: direct ISDN connection
On Tue, 5 Feb 2002, Vegh Karoly wrote: I'd like to set up an ISDN connection, in case the net-connection would be unreachable. I mean direct, because the two computers are far away from eachother, and we really need a fallback connection. I would like to be able to dial in to the other ISDN computer, but found only HOWTOs explaining how to set up connection with ISPs. What do i need for this, how to set up a dial in ISDN server? Documentation URLs are also welcome, or which HOWTO shall I read? Hi, You have two choices: 1. Raw IP over ISDN (rawip). It looks very much like an ethernet interface. 2. Synchronous PPP over ISDN (syncppp). You usually use this to connect to your ISP over ISDN, so most HOWTO's only cover setting up the calling side of this. You can also set up the answering side. Which one you choose depends on your needs. If you are to network two computers which are both under your control, then I suggest you use rawip for the folllowing reasons: - Simpler configuration. - Faster connection without the PPP authentication overhead. In fact, the computers will feel almost as if they're on the same LAN. - Ample security is provided by the isdnctrl phone-in and secure features; you only need allow incoming connections from the one phone number on the other side. - Easier and more efficient channel bundling (slave channels). The slave channels for rawip only dial out when the data rate on the master channel exceeds a presettable level for a presettable time interval. For syncppp, they are always up unless you use the ibod software to control them, and this provides much less control. IMHO you only need to set up syncppp dial-in if you're to allow users to connect with user-based authentication from sites not under your control. Here we use rawip for WAN connections (three sites) and syncppp for the connection to our ISP. If you like, I can tar up my scripts and send them to you. You can use them to set up any number of rawip and syncppp interfaces. They may look a bit complex, but they're quite flexible. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: direct ISDN connection
On Tue, 5 Feb 2002, Vegh Karoly wrote: also they do see eachother, they connect, but they also sofort do a HANGUP, and this is what i dont like... what/how shall i fix it? TO tell the truth i dont know what should happen, but i await at least that a ping goes through... what should be the next step? Please do: isdnctrl list interface-name on both computers, and post the output. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: direct ISDN connection
On Tue, 5 Feb 2002, Vegh Karoly wrote: Current setup of interface 'isdn0': EAZ/MSN:9905 snip Can't see anything wrong with these, at first glance at least. Send me the output of ifconfig interface-name on both machines. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: Woody root on RAID and 2.4.17
Hi John, You came up trumps. Your pointers got me running woody with root on software RAID. Thanks! STOP PRESS: I've now got woody with root fs on LVM on software RAID-1. The ultimate server weapon... Watch this space. I'd like to write all this up in an informal HOWTO which I'll put up on my site. I would like to give you credit for your RAID-lilo pointers. Would you like me to put up your email address in the credits or not? If you'd like me to include any other details, please forward them. I did it differently to how you suggest, in a way that leaves lilo.conf, fstab, /boot and the MBR on the single disk we start from completely untouched and may even be a bit easier to do. The key is editing the lilo.conf copied to the RAID only, and doing chroot [target] /sbin/lilo to install an MBR to a disk other than the one with the initial single-disk install. Starting with an initial install to a single disk, it goes a bit like this: /dev/md[boot]: RAID-1 array for /boot fs /dev/md[root]: RAID-[1|4|5] array for root fs (I've only tried RAID-1 but I'm confident RAID-4 and -5 will work too). /dev/[h|s]d[raid-mbr]: Disk to whose MBR we will install lilo for the RAID system; must not be the disk to which we make the initial single-disk install. - We install to a single disk. It makes things simpler if this disk is partitioned the same way as the others will be. Set the BIOS to boot from this disk and install. - Obtain or compile a RAID kernel and reboot with it. - Install raidtools-2. - If you're going to be using 2.4.x with devfs, install devfsd. - Write an /etc/raidtab with the partitions on the disk we've just installed to marked as failed-disk. Make sure these entries are not the first ones in each array. - Create the RAID arrays: raidstart /dev/md[number] - Make filesystems on them: mk[e2fs|reiserfs] /dev/md[number] - Mount the RAID array which will be used as the root fs under /mnt mount /dev/md[root] /mnt - Copy the root filesystem to it: tar cplf - -C / . | tar xvpf - -C /mnt - - Mount the other raid arrays under /mnt: /mnt/boot, /mnt/usr etc. Instead of issuing all those mount commands, you might as well put these mounts in /etc/fstab and do: mount -a In that way, they get mounted automatically in case you need to reboot into the single-disk system and work on them. - Copy the other filesystems to the corresponding arrays: tar cplf - -C / [boot|tmp|usr|other] | tar xvpf - -C /mnt - Edit /mnt/etc/lilo.conf and include: --- begin relevant lilo.conf excerpt --- # The disk we will install the RAID MBR to. # We should leave the MBR on the initial install # disk alone until we've successfully booted from RAID, # so choose another disk. disk=/dev/[h|s]d[raid-mbr] # We will use the BIOS to switch this disk to the # first BIOS id when we want to boot from it. bios=0x80 # The RAID-1 array for the /boot filesystem boot=/dev/md[boot] # Write the RAID MBR to the above disk only raid-also-boot=/dev/[h|s]d[raid-mbr] # Root filesystem device root=/dev/md2 --- end relevant lilo.conf excerpt --- - Install lilo to the disk we've chosen to hold the RAID MBR: chroot /mnt /sbin/lilo - Edit /mnt/etc/fstab to mount the RAID filesystems instead of the single-disk ones. - Reboot, go into setup and change the BIOS settings to boot from the disk we've installed the RAID MBR to. - Boot into the RAID system. If something goes wrong, use the BIOS setup to get the system to boot the disk with the initial installation and fix it. - Once booted into the raid system, edit lilo.conf and change it so the MBR is installed to all disks in the array. More details later when I've verified I can use the BIOS setup to boot from any one of the disks. - Edit /etc/raidtab and change the failed-disk entries to raid-disk. Add the partitions on the disk with the single-disk installation to the arrays with raidhotadd. - Run lilo and reboot to check that you can now boot from any of the disks. - Celebrate appropriately. Thanks again and best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: Woody root on RAID and 2.4.17
Hi, /dev/md0 - /boot /dev/md1 - swap /dev/md2 - root # Specifies the boot device. This is where Lilo installs its boot # block. It can be either a partition, or the raw device, in which # case it installs in the MBR, and will overwrite the current MBR. # boot=/dev/md0 raid-extra-boot = /dev/hdi,/dev/hde # Specifies the device that should be mounted as root. (`/') # root=/dev/md2 Thanks John. Now bear with me for a while... I've got this setup: /boot on /dev/md1 (/dev/sda1 and /dev/sdb1) root on /dev/md2 (/dev/sda2 and /dev/sdb2) I've got a normal single-disk installation on /dev/sdb. I've marked all the partitions on /dev/sda as failed-disk in /etc/raidtab. So the RAID is running in degraded mode, until I manage to boot into it. Then I'll change the entries to raid-disk and raidhotadd the partitions. This is fairly standard procedure, I understand; you install a single-disk system, start RAID arrays which include partitions on the disk with the system on it as failed-disk, clone the system into the arrays, make the RAID system bootable, boot into it, mark the failed partitions as raid and raidhotadd them. I have done this successfully a number of times with kernel 2.2 and the older lilo. Now to keep my single-disk system bootable, I guess from what you write that I need to put this in /etc/lilo.conf: --- # device with /boot filesystem boot=/dev/md1 # MBR's to which lilo will be installed; omit /dev/sdb # until the RAID system is bootable raid-extra-boot=/dev/sda # device with root filesystem root=/dev/md2 --- And then, once the RAID system is bootable, install the MBR to /dev/sdb as well: # MBR's to which lilo will be installed raid-extra-boot=/dev/sda, /dev/sdb Am I getting this right? I wouldn't like to trash the single-disk install too... Best regards, George
Woody root on RAID and 2.4.17
Hi, Has anyone successfully built a woody system with root fs on software RAID and kernel 2.4 ? I am running several installations of potato with root fs on RAID and kernel 2.2.19 with the raid-a1 patch, but my attempts to build one with woody and kernel version 2.4.17 have failed so far. I've searched the web quith thoroughly but have failed to locate an indication that to build root fs on RAID with 2.4, a different strategy must be followed to the one used with the 2.2 kernels and the new RAID patches. When booting, lilo fails with either Lnothing or L 01 01 01 01 01... I have tried both lba32 and linear in lilo.conf without result. I am following the same method as I used successfully with potato and kernel 2.2.19: - Root fs on /dev/md2 (RAID-1) (sda2+sdb2) - /boot in /dev/md1 (RAID-1) (sda2+sdb2) Included in lilo.conf: # The root fs device disk=/dev/md2 # Geometry obtained by fdisk -ul cylinders=8704 heads=64 sectors=32 # We will use the SCSI bios to choose which MBR to boot from; # in which case it will be assigned BIOS ID 0x80 bios=0x80 # The /boot device, as a partition of the root device; this is # a trick to tell lilo where to look for the /boot data. partition=/dev/md1 # The sector the /boot device starts from # (first sector of one of the partitions in the array # /dev/md1, i.e. /dev/sda1 or /dev/sdb1; life is easier # if they are both the same). start=32 # The MBR we will actually be installing to. We can change this # to /dev/sdb and rul lilo again so we can boot from either disk. boot=/dev/sda # The root device root=/dev/md2 remainder of lilo.conf is fairly standard Any ideas? | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: Woody root on RAID and 2.4.17
Hi Alvin, Thanks for your prompt reply. On Sat, 2 Feb 2002, Alvin Oga wrote: a. try to boot debian off of one single scsi disk first It works fine off one disk. I have installed on one disk and will be transferring to RAID as per Boot+Root+RAID+Lilo-HOWTO. - copy the initrd.tgz file and the kernel onto floppy I'm not using an initrd; the SCSI driver is compiled in the kernel, as is the RAID support. Used to work fine on potato with patched 2.2.19... - you will need it to read the scsi disks... and cant read teh disk till you have the scsi driver ( catch-22 problem ) - make a boot floppy that boots debian on one scsi disk first... ( at least i would wanna know that i can boot a regular ( scsi disk first from itself... and boot the scsi off a floppy The single disk boots off of the single disk MBR, so it's not a SCSI driver issue. b. make sure that your partition type is fd ( raid autodetect ) It is; the arrays on the other disks are started automatically when I boot into the single disk. when you are ready to build your sda1 and sdb1 root raid ... - i assume you have a typo below for your /boot raid partition definitions Sorry: root fs: /dev/md2 (sda2+sdb2) /boot: /dev/md1 (sda1 +sdb1) c. make sure lilo.conf has a line for initrd=initrd.gz Not using initrd... and use lba32 nt linear Tried it both ways, no joy... i'd also remove the chs info from lilo.conf OK, will try that next. - patches arent needed for raid0/raid1 for 2.2 or 2.4 kernels... and probably required for raid5 on most 2.2 kernels I'm using 2.4.17 without any patches. Thanks, G. On Sat, 2 Feb 2002, George Karaolides wrote: Hi, Has anyone successfully built a woody system with root fs on software RAID and kernel 2.4 ? I am running several installations of potato with root fs on RAID and kernel 2.2.19 with the raid-a1 patch, but my attempts to build one with woody and kernel version 2.4.17 have failed so far. I've searched the web quith thoroughly but have failed to locate an indication that to build root fs on RAID with 2.4, a different strategy must be followed to the one used with the 2.2 kernels and the new RAID patches. When booting, lilo fails with either Lnothing or L 01 01 01 01 01... I have tried both lba32 and linear in lilo.conf without result. I am following the same method as I used successfully with potato and kernel 2.2.19: - Root fs on /dev/md2 (RAID-1) (sda2+sdb2) - /boot in /dev/md1 (RAID-1) (sda2+sdb2) Included in lilo.conf: # The root fs device disk=/dev/md2 # Geometry obtained by fdisk -ul cylinders=8704 heads=64 sectors=32 # We will use the SCSI bios to choose which MBR to boot from; # in which case it will be assigned BIOS ID 0x80 bios=0x80 # The /boot device, as a partition of the root device; this is # a trick to tell lilo where to look for the /boot data. partition=/dev/md1 # The sector the /boot device starts from # (first sector of one of the partitions in the array # /dev/md1, i.e. /dev/sda1 or /dev/sdb1; life is easier # if they are both the same). start=32 # The MBR we will actually be installing to. We can change this # to /dev/sdb and rul lilo again so we can boot from either disk. boot=/dev/sda # The root device root=/dev/md2 remainder of lilo.conf is fairly standard Any ideas?
Re: Woody root on LVM on RAID
Hi, On Thu, 31 Jan 2002, Calyth wrote: Although I don't have the experience, I'd advise against using a soft-RAID root partition because if you don't have the right kernel for recovery you could be in pretty big trouble. I've been running potato with the root fs on RAID-5 on several servers. You could be in trouble if you had no rescue floppy. But I found it was a cinch to replace the standard kernel image on the potato rescue floppy with an image of the RAID kernel I'm using (2.2.19 with the RAID-2 patch). I checked that I could boot from it, mount all my filesystems and generally do anything you could do with the standard potato rescue floppy, short of installing from scratch... You're better off using a separate drive as a root partition for LVM or soft-RAID. ...Hence, I don't agree with your statement. RAID is for providing resilience, not data protection; tape backups (which, on a server, you should be doing anyway) protect data. If you're running the root fs on a single device and it fails, you'll have to restore the system from backup or rebuild from scratch. After that, restoring the data from backup is a cinch, so you might as well not run RAID at all. Bitter experience speaks, not I. A server we had here, set up with root fs on one disk and data on RAID, failed when the disk with the root fs went kaput. Not only did my predecessor as admin (I was then his assistant) go through ground-up rebuild hell, with all the users on his case, he also had a lot of explaining to do to some very un-impressed bosses, who were asking why they'd paid money for an array of disks when a single disk failure could take the server down for a complete OS rebuild, and why we were using RAID to protect data which was on tape backup anyway, instead of providing resilience. I learnt the lesson well. My servers are now redundant right down to the MBR's; you can use the SCSI BIOS to switch boot disks, and boot from any of the five disks in the array. The machine continues to run, and is able to reboot, after failure of any one disk. I have tested this by simply withdrawing a disk (hot-swap SCSI). You may have to use the BIOS to switch boot disks if the failed disk happens to be the one the machine was configured to boot from, but that's all. All this with potato and kernel 2.2.19. With woody and kernel 2.4.17, I have fond hopes of combining the redundancy achieved with Linux Software RAID with the flexibility of LVM to build a real enterprise-level server. Best regards, George Karaolides
Woody root on LVM on RAID
Hi, Has anyone out there attempted building a Debian woody system with root fs on LVM on software RAID? And while we're at it, has anyone done this with the root fs being one of the exotic new ones (reiserfs/XFS/JFS/ext3)? Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: DSL router and networking - Help!
Hi Philip, On Fri, 11 Jan 2002, Phillip Deackes wrote: snip info. about service to be purchased I use Samba on the Linux box so that the Windows machine can print to the Linux printers, and squidso that the Windows machine can access web sites. In my Linux PC I have two network cards. One card is connected to thecable modem, the other to a small 4 port Netgear hub. When I get the SpeedTouch Pro I intend removing the second card on the Linux box since the STP can be plugged into my hub. What I would like to do is to continue using the networking I have already setup for file sharing and printing (though there will obviously be some reconfiguring with there being a single nic in the Linux box. IMHO the best way to do this is to keep the two NIC's in your PC. Leave one connected to your hub and connect the other one to the STP. Then use IP masquerading. This is made easy in Debian by using the ipmasq Debian package. apt-get install ipmasq Provided there is a default route configured through the ethernet interface connected to the STP, and the IP address, netmask and broadcast for the ethernet interface connected to your hub are correct for your local net, ipmasq will set up firewalling and masquerading automatically for your net. Any machine connected to your hub and configured to use your Debian box as the default gateway will have access to the Internet. You will have the option to use squid on the Linux box as a web proxy, or access the web directly. Use the IP address of the ethernet interface on the Linux box which is connected to the hub as a default gateway on all machines connected to the hub. All machines will be able to access all Internet services, not just the Web. Services provided to your local net by the linux box should not be affected by any of this. You will also need to set up name services so that machines on your internal net can resolve names on the Internet. You can either use your ISP's nameserver(s) or run your own local nameserver on the Linux box. The latter solution would give a performance advantage because query results are cached, avoiding the net traffic caused by repeat queries, but you can leave this upgrade for later. The quick setup guide for the STP uses DHCP as a default which would be great apart from the fact I can't see how I would be able to network the two machines if the IP addresses keep changing. On your Linux box, set up DHCP only for the interface connected to the STP and leave the interafce connected to the hub configured with a static address. Edit /etc/network/interfaces and check the entry for the ethernet interface connected to the hub, assuning it's eth0: iface eth1 inet static address 192.168.1.1 (for example) netmask 255.255.255.0 network 192.168.1.0 broadcast 192.168.1.255 Add (or change) the entry for the interface connected to the STP as follows, assuming it's eth1: iface eth1 inet dhcp Of course, this assumes that your ISP will be assigning your IP address and routing info. via DHCP, which might not be the case: Looking through the manual, I see sections for all sorts of protocols. Eclipse Internet tell me to use RFC2364 PPPoATM VC Encapsulation Multimode AutoModulation. The nearest I can find in the STP manual is either PPoA-to-PPTO relaying or PPP IP routing. Which should I use? Can't help you there, sorry. Take this up with the people who sold you the STP and with your ISP. Once you sort out how to connect the STP to your ISP, we can sort out how to get the Debian IP layer to play with them. I would imagine I would need to allocate a static IP address to each machine (ideally what I already use - 192.168.1.1 and 192.168.1.0) this would surely mean that the networking would just carry on working. Don't use the address 192.168.1.0 as a host address. It's the first address on the network 192.168.1.0/24 (netmask 255.255.255.0). Being the first address, it's reserved to refer to the whole network. Use e.g. 192.168.1.2 instead. If the address of your private net is 192.168.0.0/16 instad of /24 (netmask 255.255.0.0 instead of 255.255.255.0), then the first address is 192.168.0.0 and technically 192.168.1.0 is useable as a host address. However, accepted practice is that addresses on 24-bit boundaries such as 192.168.0.0, 192.168.1.0, 192.168.2.0 etc. should be reserved in case the network 192.168.0.0/16 needs to be divided (subnetted) into smaller nets e.g. 192.168.0.0/24, 192.168.1.0/24 etc. Best regards, good luck and don't hesitate to come back with any more questions, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: DSL router and networking - Help!
Hi Philip, On Fri, 11 Jan 2002, Phillip Deackes wrote: snip background info. I use Samba on the Linux box so that the Windows machine can print to the Linux printers, and squidso that the Windows machine can access web sites. In my Linux PC I have two network cards. One card is connected to thecable modem, the other to a small 4 port Netgear hub. When I get the SpeedTouch Pro I intend removing the second card on the Linux box since the STP can be plugged into my hub. What I would like to do is to continue using the networking I have already setup for file sharing and printing (though there will obviously be some reconfiguring with there being a single nic in the Linux box. IMHO the best way to do this is to keep the two NIC's in your PC. Leave one connected to your hub and connect the other one to the STP. Then use IP masquerading. This is made easy in Debian by using the ipmasq Debian package. apt-get install ipmasq Provided there is a default route configured through the ethernet interface connected to the STP, and the IP address, netmask and broadcast for the ethernet interface connected to your hub are correct for your local net, ipmasq will set up firewalling and masquerading automatically for your net. Any machine connected to your hub and configured to use your Debian box as the default gateway will have access to the Internet. You will have the option to use squid on the Linux box as a web proxy, or access the web directly. Use the IP address of the ethernet interface on the Linux box which is connected to the hub as a default gateway on all machines connected to the hub. All machines will be able to access all Internet services, not just the Web. Services provided to your local net by the linux box should not be affected by any of this. You will also need to set up name services so that machines on your internal net can resolve names on the Internet. You can either use your ISP's nameserver(s) or run your own local nameserver on the Linux box. The latter solution would give a performance advantage because query results are cached, avoiding the net traffic caused by repeat queries. You can start off using your ISP's nameserver and leave this upgrade for later. The quick setup guide for the STP uses DHCP as a default which would be great apart from the fact I can't see how I would be able to network the two machines if the IP addresses keep changing. On your Linux box, set up DHCP only for the interface connected to the STP and leave the interafce connected to the hub configured with a static address. Edit /etc/network/interfaces and check the entry for the ethernet interface connected to the hub, assuning it's eth0: iface eth1 inet static address 192.168.1.1 (for example) netmask 255.255.255.0 network 192.168.1.0 broadcast 192.168.1.255 Add (or change) the entry for the interface connected to the STP as follows, assuming it's eth1: iface eth1 inet dhcp Of course, this assumes that your ISP will be assigning your IP address and routing info. via DHCP, which might not be the case: Looking through the manual, I see sections for all sorts of protocols. Eclipse Internet tell me to use RFC2364 PPPoATM VC Encapsulation Multimode AutoModulation. The nearest I can find in the STP manual is either PPoA-to-PPTO relaying or PPP IP routing. Which should I use? Can't help you there, sorry. Take this up with the people who sold you the STP and with your ISP. Once you sort out how to connect the STP to your ISP, we can sort out how to get the Debian IP layer to play with theirs. I would imagine I would need to allocate a static IP address to each machine (ideally what I already use - 192.168.1.1 and 192.168.1.0) this would surely mean that the networking would just carry on working. Don't use the address 192.168.1.0 as a host address. It's the first address on the network 192.168.1.0/24 (netmask 255.255.255.0). Being the first address, it's reserved to refer to the whole network. Use e.g. 192.168.1.2 instead. If the address of your private net is 192.168.0.0/16 instad of /24 (netmask 255.255.0.0 instead of 255.255.255.0), then the first address is 192.168.0.0 and technically 192.168.1.0 is useable as a host address. However, accepted practice is that addresses on 24-bit boundaries such as 192.168.0.0, 192.168.1.0, 192.168.2.0 etc. should be reserved in case the network 192.168.0.0/16 needs to be divided (subnetted) into smaller nets e.g. 192.168.0.0/24, 192.168.1.0/24 etc. Best regards, good luck and don't hesitate to come back with any more questions. | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Tape devices
Hi, Can somebody tell me what the various tape devices are? I know that /dev/st0 is the first SCSI tape drive and /dev/nst0 is the non-rewinding version of this, but what are /dev/(st|nst)0a, -l and -m? mymachine:~# ls -l /dev/*st0* crw-rw1 root tape 9, 128 Aug 3 10:11 /dev/nst0 crw-rw1 root tape 9, 224 Aug 3 10:11 /dev/nst0a crw-rw1 root tape 9, 160 Aug 3 10:11 /dev/nst0l crw-rw1 root tape 9, 192 Aug 3 10:11 /dev/nst0m crw-rw1 root tape 9, 0 Aug 3 10:11 /dev/st0 crw-rw1 root tape 9, 96 Aug 3 10:11 /dev/st0a crw-rw1 root tape 9, 32 Aug 3 10:11 /dev/st0l crw-rw1 root tape 9, 64 Aug 3 10:11 /dev/st0m Thanks, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
RAID repartitioning
Hi, I'm trying to repartition disks in a RAID array on potato, and have largely succeeded except for making a new RAID array to use the space I freed up on each disk. The machine runs kernel 2.2.19 with the new RAID patch compiled from the Debian packages of both kernel 2.2.19 and the raid patch. The disks are five identical scsi disks, /dev/sd[a-e]. The last partition on each disk, /dev/sd[a-e]10, forms a RAID-5 array /dev/md10. The filesystem on this is mounted under /home. It was my intention to split /dev/md10 up into two smaller arrays because I was using a very small part of /dev/md10 for /home and was running out of space in /var. I backed up the /home filesystem data in /dev/md10, unmounted /home and then did raidstop /dev/md10. I re-partitioned the disks, reducing the size of /dev/sd[a-e]10 on each. I then did # mkraid /dev/md10 # mke2fs -b 4096 -R stride=8 /dev/md10 # mount -a # tar -x restore data in /home from archive This went just fine, I now have my home filesystem in a smaller array as I intended (I was using a very small part of it). While reducing the size of /dev/sd[a-e]10 on each disk, I also made a new partition /dev/sd[a-e]11 on each disk. The intention was to make these partitions into a new, additional RAID5 array /dev/md11. Trying to do this, I got the following error: # mkraid /dev/md11 handling MD device /dev/md11 analyzing super-block couldn't open device /dev/sda11 -- Device not configured mkraid: aborted, see the syslog and /proc/mdstat for potential clues. So I can't make my new partitions into a RAID-5 array. I haven't found any clues in either /proc/mdstat or the log files. Any suggestions? | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: RAID repartitioning - SOLVED
Hi guys, In the end my RAID-5 repartitioning problem was quite simple... From the cfdisk manpage: W Write partition table to disk (must enter an upper case W). Since this might destroy data on the disk, you must either confirm or deny the write by entering `yes' or `no'. If you enter `yes', cfdisk will write the partition table to disk and the tell the kernel to re-read the partition table from the disk. The re-reading of the partition table works is most cases, but I have seen it fail. Don't panic. It will be correct after you reboot the - system. In all cases, I still recommend rebooting -the system--just to be safe.^^^ ^^ Having been totally spoilt by Debian, I have completely got out of the habit of rebooting any machine if it can be avoided. In this case I thought it could, but it couldn't. I needed to reboot after repartitioning (using cfdisk) and before mkraid. Having done this, I was able to mkraid, mke2fs and transfer all my data. Happiness and bliss. Thanks for the suggestions; I will definitely look into the software I was pointed to. Have a nice weekend, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: Missing log file
Hi, On Tue, 8 Jan 2002, Nick Hastings wrote: Hi, * George Karaolides [EMAIL PROTECTED] [020108 09:09]: Hi all, Does anyone know what, if anything, besides foul play could cause one of the old, compressed log files to disappear? I'm missing a whole day's /var/log/messages.n.gz from this past week. Old logs are deleted periodically. Look at the man pages for logrotate and cron. Nick. Yes, but normally it's the oldest one that gets deleted. I'm missing one from the middle of the series. How could that come about? Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Missing log file
Hi all, Does anyone know what, if anything, besides foul play could cause one of the old, compressed log files to disappear? I'm missing a whole day's /var/log/messages.n.gz from this past week. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: Securing bind..
that the source is available so you can make and distribute patches to it. the rest of the unix world got sick of hunting for obscure patches to programs ages ago - hunt for the original source, then hunt for half a dozen patches to fix various problems or add required features and hope that the patches will actually apply cleanly without any conflict. that was the way we had to do things back in the bad old days. it sucked then, and it still sucks now. Yes it does suck if you *have* to apply patches. But you only need them if you need features not included in the core djbdns distribution. I humbly submit that for the needs of the original poster, and for the majority of Debian users who need a nameserver, no patches are needed. Again, let's take this elsewhere. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: Securing bind..
Hi Martin, On Thu, 3 Jan 2002, martin f krafft wrote: snip djbdns data example i find this horrible. BIND zonefiles at least allow for usage of tabs to organize your zone into tabular data. Everyone has their favourite wokrkign techniques. You like your tabular BIND zone files. I like my line-based djbdns data files. Fine. I know there are management tools that automate synchronisation of forward and reverse mappings in BIND zone files, but why should the reverse-mapping information be in a file separate from the forward information? why does the DNS protocol even allow this? keeping them in sync is really what you should do, but there are cases, where they may need to be out of sync. for instance, domain.com and mail.domain.com for small domains usually get the same A record on my servers, if what i delegate to mydomain.com is a single IP address. that's not a CNAME, that's an A record. but what does the IP map to? *not* mail.domain.com. i find it very handy to have explicit control over the mappings. As I've written to Craig, you can do all this with djbdns. But where you can keep the foward and reverse data on the same server, djbdns allows you to make a single entry that takes care of both the A nd PTR record. I like the convenience of that, and the reduction of the potential for mistakes. priorities or preferences. i find BIND does it's job. it might not be optimal, but it's free, and it works. No arguments there. BIND is free. BIND works. But I suggested that with djbdns, you get the security by default that you have to configure into BIND, and that for a setup as needed by the original poster, djbdns could be easier to administer. djbdns might even be better, smaller, faster, more secure. but my nameservers are not going to be running a piece of software because of either of two reasons: (a) the author is braindead in many aspects; and (b) it's non-free. I use djbdns. It works. I can administer it. I didn't need to read a HOWTO about how to run it securely. It's packaged in Debian, and the above benefits are available to Debian users. Issues to do with its license and the brain-deadedness or otherwise of its author are off-topic so I agree with you; this discussion is best continued elsewhere. we'd like to package it, not require the end-user to download the -source package, which includes voodoo magic to compile djbdns without the user ever knowing what gcc is. the problem is that this is an aweful job for the maintainer, and that it requires some -dev libraries on the target system, as well as the entire gcc family, which might not be needed otherwise... This too finds me in perfect agreement. I would love it if djb would GPL his software. let's take this discussion somewhere else. As stated above, agreed. Best regards, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: Securing bind..
Hi, On Tue, 1 Jan 2002, Craig Sanders wrote: someday soon, someone's going to take the good ideas from djbdns, combine it with the good stuff from bind (including backwards compatibility with bind config zonefile formats), add a few useful new ideas (e.g. an RXFR protocol that embedded the rsync protocol directly) to produce a fast, secure DNS daemon, and release it with a GPL-compatible license. this will blow both bind djbdns out of the water. Roll on the day! When such a godsend appears, I'll grab it with both hands, provided of course that besides reverse compatibility with BIND config. files, it also gives you a new, simpler config. scheme. In the meantime, I submit that djbdns is OK unless you really, really want to stick to the BIND zonefile format, though I seem to recall hearing it can be made to read BIND zone files. IMHO once you're used to it, the djbdns data file format is actually quite nice. I've worked on both BIND and djbdns and I find the latter easier to use. For example, the following ten entries would require four entries in named.conf and four zone files: # Nameserver for my network addesses... .my-network-in-reverse-order.in-addr.arpa:my-nameserver-ip-address:TTL # ... and for addresses in my other network... .my-other-network-in-reverse-order.in-addr.arpa:my-nameserver-ip-address:TTL # ... and for names in my domain... .my-domain:my-nameserver-ip-address:TTL # ... and for names in my other domain... .my-other-domain:my-nameserver-ip-address:TTL # Now to get on with mapping names to addresses, and vice versa =i-want-this-to-map-both-reverse-and-forward.my-domain:ipaddress1:TTL +i-only-want-this-to-map-forward-cos-its-an-alias.my-domain:ipaddess1:TTL =another-bothways-map.my-domain:ipaddress2:TTL +and-this-too-has-an-alias.my-domain:ipaddress2:TTL =a-bothways-map.my-other-domain:ipaddress3:TTL +an-alias.my-other-domain:ipaddress3:TTL Surely that's not all bad? You don't have to worry about keeping A and PTR records in sync. In fact, you don't have to worry about subdomains of in-addr.arpa at all, beyond making sure that your clients are querying the right servers (which you'd do anyway), that any delegations to your server are done correctly by your parent (ditto) and that your data includes a nameserver entry for the appropriate in-addr.arpa subdomains (again ditto). I know there are management tools that automate synchronisation of forward and reverse mappings in BIND zone files, but why should the reverse-mapping information be in a file separate from the forward information? Once the three conditions above are met, why should we need to administer the forward and reverse mappings separately? BTW these are not rhetorical questions; I'd love to hear input on this. Myself, I haven't thought about any subdomain of in-addr.arpa since installing djbdns, besides the three points mentioned above. In the end, it's all a question of priorities. If compatibility with existing config. and zone files is an issue, then djbdns may well be a non-starter, my recall that there's a way to get it to read BIND zone files notwithstanding. If managing a DNS name space painlessly, securely and reliably is, then it could well be. It was for me. For all the arguments against djb's attitude re. development and licensing, it must be acknowledged that his keeping tight control of the software has prevented it from suffering from feature bloat. And since it's open-source and you can distribute patches to it, there's no shortage of patches to get it to do what you want. Heppy new year, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: Securing bind..
On Mon, 31 Dec 2001, Petre Daniel wrote: thank you all very much. you're right.if one doesn't have anything useful to say i'll recommand him to let others help.. thx guys. Hi Petre, Maybe the initial suggestion to use djbdns wasn't worded in the best possible way, but might I suggest that you consider the possibilities djbdns, which could perhaps make your life a bit simpler? Your requirement was for a DNS server which does name resolution for clients on your internal net and serves out names for a few .ro domains to the Internet at large. So here are my points: - djbdns comes with the security features you require out of the box. - Yours does not sound like a large and complex setup with years of BIND configuration development behind it, so you have little to lose if you do try djbdns. - In my experience, it is very easy to set up a system like the one you describe using djbdns. - It also sounds like you're busy enough, so In My Humble Opinion(tm) you will save time administering djbdns. Sources of djbdns debs: potato: Gerrit Pape has some unofficial debs at the following apt source: deb ftp://ftp.innominate.org/pub/pape/Debian potato unofficial innominate apt-get install djbdns woody or sid: There are official installer packages: apt-get install djbdns-installer If you do decide to try it out, then I am at your disposal to help with any setup queries you may have. Best regards and a happy new year, | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus |
Re: What's a debian kid look like?
Hi, Single white male, 33, systems admin at a small financial services company here in the island Republic of Cyprus in the Eastern Medirerranean. About to complete the process of moving most services and networking at my workplace to Debian GNU/Linux, with office users on Windoze accessing a Samba server. Would be over the moon with joy if I could get some office application under Linux to read M$ Word and Excel files written in Greek characters, because then I could flush M$ completely out of the company and have done with viruses, crashes and users installing silly software. I am in the happy position of being able to take most decisions in my area (OS/network) myself, and have the support of the company's financial controller who is tech-aware and now also a Debian fan. I bet that makes most of the sysadmins on the list envious... Absolutely delighted with Debian, and love the mailing lists too. Nothing easier than keeping a Debian machine current with all the security updates and stability patches. I keep everything on stable, with the exception of my personal workstation which runs testing. My bosses also love the stability and performance improvement of our systems since the switch to Debian. Hated computers until doing my B.Sc. in Physics at Imperial College in London in '91-'95, where I found a cluster of DEC Alphas running mostly GNU software in the second year lab. Heve been a GNU addict since. Eventually Linux saved me from the drudgery of earning my living by teaching Physics to rude teenagers who hate learning in general and learning Physics in particular. Have written some half-decent ISDN setup scripts and ipchains firewalling scripts, but generally can't program to save my life. I have no interest at all in computer games. Own four Alfa Romeos from '96 to '79, one in very nice condition, one being restored to even nicer condition, one long-term restoration project (read disaster area) and one complete but immobile, with a questionmark hanging over its future. Have a big record (LP, not CD) collection, mostly classical. Latest read: Cliff Stoll's The Cuckoo's Egg (am I the last person on the list to read this?). Mess about with antique/high end audio: turntables, valve (I think you call them tubes in the States) amplifiers, reel-to-reel tape decks (big fan of Revoxes), electrostatic loudspeakers and the like. Best regards, | | | George Karaolides 8, Costakis Pantelides St., | | tel: +357 99 68 08 86 Strovolos, | | email: [EMAIL PROTECTED] Nicosia CY 2057, | | web: www.karaolides.com Republic of Cyprus | | | | Please note Cyprus telephone area codes have changed; see above. | On 19 Dec 2001, Matt wrote: Forget support and distro discussion for a second.. Who are we here? I'm curious about what the Debian demographic is. I know this is sort of an impossible thing to answer, but it'd be interesting to explore in vague terms who uses Debian, at least compared to those who don't. Crushing realities like, Am I typical? Does using Debian separate me from the rest? Controversial questions like, Will I ever meet more than 1 or 2 female Debian users who are active on mailing lists? Am I a Microsoft-branded pinko 'cause I like open-source? IT? Artsy? Young? Old? Bitter? Socialist or CrewCut? Utopian or realist? I'll try to break the ice a bit: white male, 22, Canada, FineArts ex-pat turned CompSci near-grad, downtempo and raregroove on the weekends, csound, php and java development on the weekdays. Left-leaning (Harpers, not New Republic), piano-playing, community-oriented, social. Curious why so many people use RedHat. (Just an experiment) cheers, Matt -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Missing disk image file on ftp.debian.org
Hi, A disk image file seems to be missing from ftp.debian.org. Specifically, in the directory /debian/dists/potato/main/disks-i386/2.2.26-2001-06-14/images-1.44 the file base-1.bin seems to be missing. This is the correct directory for the latest stable disks, right? Why is the first base-system disk image missing? Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Errors compiling PHP4 on potato
On Thu, 6 Dec 2001, Brian Clark wrote: * George Karaolides [EMAIL PROTECTED] [Dec 06. 2001 09:09]: I'm trying to compile PHP4 on Debian potato. I'm using the Debian source package php4_4.0.3pl1-0potato1. You should consider using a recent snapshot from snaps.php.net I'm getting the following errors, which seem to be related to mysql. This is despite the fact that I've commented out the mysql configuration options in debian/rules (I don't need mysql support). If you're positive you don't want mysql support, try configuring with: --disable-mysql I couldn't find out how to pass configuration options to the build process; no matter what I put in debian/rules, the build process seemed to have its own ideas about what the configure options were. In the end, I had to go with compiling from original sources. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Character-mode clock
Yes, thanks David, 'watch -n1 date' worked fine for me. George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus On Wed, 5 Dec 2001, David Z Maze wrote: George Karaolides [EMAIL PROTECTED] writes: GK Is there a clock application that can be used to show system time in a GK character-mode terminal? There seem to be any number of them that can be GK used under X, but I can't find any that can be used in character mode. For something quick, cheap, and easy, try 'watch -n1 date'. My memory of using screen on a VT320 plugged into a serial port suggests that screen would also display the current time on the hardware modeline if it could be convinced that such existed, but this is kind of a fuzzy memory, and I'm not sure if you could (or would want to) emulate the modeline in an xterm/Linux virtual console. -- David Maze [EMAIL PROTECTED] http://people.debian.org/~dmaze/ Theoretical politics is interesting. Politicking should be illegal. -- Abra Mitchell -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Character-mode clock
On Wed, 5 Dec 2001, Alan Shutko wrote: George Karaolides [EMAIL PROTECTED] writes: Is there a clock application that can be used to show system time in a character-mode terminal? M-x display-time. You _are_ running everything in Emacs, right? 8^) I'm actually trying to un-learn emacs and get used to vi... ;) But anyway, M-x display-time only shows minutes. Could it be made to show seconds? Unless you want a clock that will just take over the terminal? A clock that would take over the terminal, like the 'watch -n1 date' solution suggested by other posters, is OK. Having the time display in a modeline or other appropriate area while I'm using the xterm would be an added bonus. What I have is a lot of machines I'm administering via ssh, and what I need is a way to check the system clocks that isn't a snapshot like 'date'. George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Character-mode clock
On Thu, 6 Dec 2001 [EMAIL PROTECTED] wrote: Sounds like what you want is `vcstime` ...it's part of console-tools, so I'd be betting it's on your system already. Check in /etc/console-tools/config for the option to turn it on on system boot. I think vcstime only runs in the console, not in a terminal like an xterm. I need to see the clock when remotely logged in. If you want a copy of my `timescreen` script, let me know and I'll clean it up a bit and get a release out ;) Depends on what it does. If it can display the time while I'm logged in remotely and using the terminal (doesn't take over the terminal completely like 'watch -n1 time'), then OK. Otherwise just running 'watch -n1 time' does the job for me. Thanks, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Errors compiling PHP4 on potato
Hi, I'm trying to compile PHP4 on Debian potato. I'm using the Debian source package php4_4.0.3pl1-0potato1. I'm getting the following errors, which seem to be related to mysql. This is despite the fact that I've commented out the mysql configuration options in debian/rules (I don't need mysql support). Help, anyone? --- begin errors --- In file included from /root/php/php4-4.0.3pl1/ext/mysql/libmysql/libmysql.c:14: /root/php/php4-4.0.3pl1/ext/mysql/libmysql/m_string.h:165: parse error before `__extension__' /root/php/php4-4.0.3pl1/ext/mysql/libmysql/m_string.h:165: parse error before `' make[5]: *** [libmysql.lo] Error 1 make[5]: Leaving directory `/root/php/php4-4.0.3pl1/apache-build/ext/mysql/libmysql' make[4]: *** [all-recursive] Error 1 make[4]: Leaving directory `/root/php/php4-4.0.3pl1/apache-build/ext/mysql/libmysql' make[3]: *** [all-recursive] Error 1 make[3]: Leaving directory `/root/php/php4-4.0.3pl1/apache-build/ext/mysql' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/root/php/php4-4.0.3pl1/apache-build/ext' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/root/php/php4-4.0.3pl1/apache-build' make: *** [build-apache-stamp] Error 2 --- end errors --- George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Errors compiling PHP4 on potato
Hi, On Thu, 6 Dec 2001, Brian Clark wrote: * George Karaolides [EMAIL PROTECTED] [Dec 06. 2001 09:09]: I'm trying to compile PHP4 on Debian potato. I'm using the Debian source package php4_4.0.3pl1-0potato1. You should consider using a recent snapshot from snaps.php.net OK, I'll try that next. I'm getting the following errors, which seem to be related to mysql. This is despite the fact that I've commented out the mysql configuration options in debian/rules (I don't need mysql support). If you're positive you don't want mysql support, try configuring with: --disable-mysql No, that didn't work with the Debian source package... :( Thanks and best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Manager for software in /usr/local
Hi, I recall recently someone on this list mentioning a software, maybe a Debian package, that manages installation of software under the /usr/local hierarchy. I seem to recall that it was said that this manager installs each piece of software in a separate subdirectory, and makes symlinks in the appropriate places. Does anyone remember what this was? Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Manager for software in /usr/local
On Thu, 6 Dec 2001, Alan Shutko wrote: George Karaolides [EMAIL PROTECTED] writes: I recall recently someone on this list mentioning a software, maybe a Debian package, that manages installation of software under the /usr/local hierarchy. stow, maybe? Yes, that's the one, thanks!
Character-mode clock
Hi, Is there a clock application that can be used to show system time in a character-mode terminal? There seem to be any number of them that can be used under X, but I can't find any that can be used in character mode. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Thoughts on RTFM
Have a look at RUTE (Rute User's Tutorial and Exposition) at http://rute.sourceforge.net It's available as a book, but being GPL'd it's also available for free download. Being GPL'd also means it's open to your contributions. It'll tell you everything you always wanted to know but everyone assumed you knew already. Debian package of rute, anyone? Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus On Thu, 29 Nov 2001, cmasters wrote: Greetings all, My recent difficulties with printer and mail setup have led to to the point where I simply must express my thoughts on RTFM. 1. I'm beginning that this should be the credo to linux use. Which would be fine if reading the fine manual didn't imply previous knowledge of the OS or app that one is reading on 2. Here's an appropriate analogy: I've decided to learn Hebrew, after some research, I find an instructor (eg a mailing list) and am given directions to download a complete list of standard vocabulary, syntax, sentence structure, and other tools necessary to learn this new language. I begin to read these documents, only to discover within the first page, that it has been written from the point of someone ~raised~ with Hebrew. Their are shortcuts, strange citations, symbols and references to structures that only make sense once Hebrew has ~been learned~. It becomes an excercise in circular reasoning. In order to learn A, one must read B, which extensively references C, which directs one to re-read A. If in doubt return to C ... or was that B. Do you see. No-one can possible be expected to ~read for comprehension~ a manual that has ~not~ been written with a ~complete~ newbie in mind. In a previous response to email regarding my ongoing email problems, I made reference to the Mutt manual. Section 6.3 Configuration Variable is chock full of all kinds of variable that can be invoked through command line or the config file. The author(s) forgot to include any ~in context~ examples or to identify which of these variables requires the #$% ~set~ prefix. Ah well ... I guess I have the time to configure, test, oops ... remove the set, re-save, test again until I have it running as I wish. Yep ... allowing the user to have complete and total control over the way their system runs is a grand idea. I actually fully support it (hence my refusal to cave-in and use M-crap). But somewhere along the line, bells and whistles should have gone off that all the configuring in the world WILL NOT HAPPEN without clear, concise, and comprehensible instructions/directions. I vow that once I've overcome the current mail hurdle ~I~ will document my trials and errors and post them to the web in a format that a complete and utter newbie will be able to understand. I've got to assume that no one is getting rich off the constant mistakes and misconfigurations that many people suffer from. Nah that would be too much like BG. Last words (for now) ... in order for Open Source to have continued and increasing validity, it ~has~ to mean more than just change at will. It must include if I've written the application, I will include ~clear~ instructions. All for now, C. Masters -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Compiling PHP4 on potato
Hi, I'm trying to compile PHP4 on potato. I know very little about handling Debian source packages; in fact nothing beyond what I could find in the Debian FAQ (pointers to TFM's I could R very welcome). I need to compile PHP with Sybase support, but how do I pass configuration options (e.g. --with-sybase=/opt/sybase)? Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Fritz!PCI2
On Sun, 25 Nov 2001, Ilja Vladimirov wrote: Hi! I've got a Fritz!PCI v2.0 -ISDN-card and just don't manage it to get this card installed. Some days ago I got the new Kernel (2.2.14). I'm new in Linux and'll be very thankful if someone could help me! thx Ilja Hi, Refer to http://www.isdn4linux.de, and read the latest mailing list archives. Look for Kai Germaschewski's kernel patches for the Fritz!PCI version 2. Better yet, subscribe to the mailing list; send a message with: subsribe isdn4linux your-email-address in the message *body* to: [EMAIL PROTECTED] Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Fritz!Card PCI ISDN adapter and Woody..
I use the Frits!PCI on potato, not woody, and only install isdnutils to get the user-space utilities; I rolled my own startup scripts etc. But from what I can figure out from your message it doesn't look like a module problem. It looks more like a problem with setting up the network interfaces, which has to do with the /etc/init.d/isdnutils script and its associated config. files, not the module. So I'd suggest you start at looking at configuration rather than modules. Sorry not to be more helpful, but that's the best I can do, given that I've no experience of the Debian isdnutils scripts. If this is an initial install and you get really stuck, I could scoot you a tarball of my setup scripts. You can use them both for synchronous ppp (usually for connecting to ISP's) and for raw IP over ISDN (usually for linking bits of your WAN). No guarantees of course, but they work quite well for providing Internet access to about two dozen users and linking three sites in the WAN I administer for my employers, all on ISDN. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus On Wed, 21 Nov 2001, Kilian wrote: Hi All I am trying to get the AVM Fritz!Card PCI ISDN adapter to work with my Woody system. I've built a custom 2.4.14 Kernel with the HiSax module and have installed the fcpci driver from the AVM website. So far it is working, the system recognizes the card during boot: | CAPI-driver Rev 1.21.6.8: loaded | capifs: Rev 1.14.6.8 | capi20: started up with major 68 | kcapi: capi20 attached | capi20: Rev 1.44.6.15: started up with major 68 (middleware+capifs) | fcpci: AVM FRITZ!Card PCI driver, revision 0.1 | fcpci: Loading... | fcpci: Driver 'fcpci' attached to stack | kcapi: driver fcpci attached | fcpci: Auto-attaching... | PCI: Found IRQ 5 for device 02:0d.0 | PCI: Sharing IRQ 5 with 02:09.0 | fcpci: Stack version 3.09-10 | kcapi: Controller 1: fritz-pci attached | kcapi: card 1 fritz-pci ready. | fcpci: Loaded. | kcapi: notify up contr 1 | capi: controller 1 up and lsmod shows the following: | Module Size Used by | fcpci 538304 1 | capi 18688 0 | capifs 3424 1 [capi] | kernelcapi 29216 2 [fcpci capi] | capiutil 22272 0 [kernelcapi] | hisax 156960 0 (unused) However, caling /etc/init.d/isdnutils start brings me the following errors: | Starting ISDN services : interfaces/dev/isdnctrl: No such device | ippp0 failed. | /dev/isdnctrl: No such device | isdn0 failed. | iprofd/dev/isdninfo: No such device | isdnlog/dev/isdnctrl: No such device I am quite desperate because I just can't figure out where the problem is.. Any hints and RTFMs are very welcome. --Kilian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to NOT start X-Windows @ Boot?
On Wed, 14 Nov 2001, Colin Watson wrote: On Wed, Nov 14, 2001 at 07:38:03AM -0500, Rafe B. wrote: I need to upgrade my video card. At present, my Debian system launches into X-windows all on its own... even the login prompt is in 'X'. mv /etc/rc2.d/S99xdm /etc/rc2.d/s99xdm ... or disable it in some other way. You might be using gdm, kdm, or wdm instead of xdm. On my machine, I configure runlevel 2 to be everything but X, and runlevel 5 to be everything including X. This can be done using update-rc.d: update-rc.d xdm start 01 5 . stop 99 0 1 2 3 4 6 . I can change runlevels simply by issuing [EMAIL PROTECTED]:# telinit 2 (or telinit 5, of course). Then I can set the default runlevel by editing /etc/inittab: --- begin excerpt from /etc/inittab --- # /etc/inittab: init(8) configuration. # $Id: inittab,v 1.8 1998/05/10 10:37:50 miquels Exp $ # The default runlevel. id:2:initdefault: --- end excerpt from /etc/inittab --- Change to id:5:initdefault to boot into runlevel 5 instead of 2, and vice versa. This allows me to work in text-only or full graphical mode as required. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: xscreensaver on woody
On 11 Nov 2001, Michael Heldebrant wrote: /etc/alternatives/x-session-manager does not exist. Should it, and if yes, what should it point to? It should only point to something if you have an x-session-manager installed. If you don't there's no point in having it exist. What's an x-session-manager? Do I need one? Have you also tried making your .xsession: xscreensaver exec icewm (or whatever the command is, I don't know) I'll try that next. Well that didn't work either... I'm at a complete loss. George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: xscreensaver on woody
On 8 Nov 2001, Michael Heldebrant wrote: On Thu, 2001-11-08 at 11:41, George Karaolides wrote: On 7 Nov 2001, Michael Heldebrant wrote: On Wed, 2001-11-07 at 16:15, George Karaolides wrote: Hi, I'm running woody, using the icewm window manager. I can start xscreensaver from the command line, and I need to get it to start when I log in (from wdm). I've tried to put xscreensaver -no-splash in my ~/.xsession file without success. Does the xscreensaver line come before or after the icewm line in .xsession? Please show relevant config files (.xsession) and output (.xsession-errors) if this is not the case. --mike This is .xsession-errors; the last line is the result of my trying to lock the screen from the icewm menu entry ( I saw it happen using tail -f). icewm: Bad option: TaskBarShowPPPStatus icewm: Bad option: IgnoreNoFocusHint icewm: Bad option: ShowXButton icewm: Bad option: WindowListFontName icewm: Warning: Could not load font ''. icewm: Warning: Could not load font ''. xscreensaver-command: no screensaver is running on display :0.0 The only line in my .xsession file reads: xscreensaver I have tried giving the absolute path, and renaming the file to .Xsession, without result. Does /etc/X11/Xsession.options have allow-user-xsession? Yes. Which X version are you running? 4.1.0-8 What do /etc/alternatives/x-session-manager and /etc/alternatives/x-window-manager point to? /etc/alternatives/x-window-manager - /usr/bin/X11/icewm /etc/alternatives/x-session-manager does not exist. Should it, and if yes, what should it point to? Maybe wdm is the missing link here, I don't use them so I don't know if it re-reads your .xsession once it logs you in. Perhaps the man page for it has something to say. Have you also tried making your .xsession: xscreensaver exec icewm (or whatever the command is, I don't know) I'll try that next. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: xscreensaver on woody
On 7 Nov 2001, Michael Heldebrant wrote: On Wed, 2001-11-07 at 16:15, George Karaolides wrote: Hi, I'm running woody, using the icewm window manager. I can start xscreensaver from the command line, and I need to get it to start when I log in (from wdm). I've tried to put xscreensaver -no-splash in my ~/.xsession file without success. Does the xscreensaver line come before or after the icewm line in .xsession? Please show relevant config files (.xsession) and output (.xsession-errors) if this is not the case. --mike This is .xsession-errors; the last line is the result of my trying to lock the screen from the icewm menu entry ( I saw it happen using tail -f). icewm: Bad option: TaskBarShowPPPStatus icewm: Bad option: IgnoreNoFocusHint icewm: Bad option: ShowXButton icewm: Bad option: WindowListFontName icewm: Warning: Could not load font ''. icewm: Warning: Could not load font ''. xscreensaver-command: no screensaver is running on display :0.0 The only line in my .xsession file reads: xscreensaver I have tried giving the absolute path, and renaming the file to .Xsession, without result. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
xscreensaver on woody
Hi, I'm running woody, using the icewm window manager. I can start xscreensaver from the command line, and I need to get it to start when I log in (from wdm). I've tried to put xscreensaver -no-splash in my ~/.xsession file without success. Any ideas? George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Weird File Permissions
Hi, s is the setuid and/or setgid permission: setuid in the user field, setgid in the group field. On files, setuid/setgid allow the group/user ID of the process started when invoking an executable file to be set to the group/user ID owning the file, respectively. Setting the setgid bit on a directory influences the group ownership of files and directories created under it. On Linux, the default behaviour is for new directories to be created with the default group of the creating user (System V convention). Setting the setgid bit on the directory forces new directories created under it to have the same group as the parent directory; this is the BSD convention. In your case, all new files and directories created under /home will be owned by group staff. I'm not sure what setting the setuid bit on a directory does; it seems to have much the same effect as setting the setgid bit, but there must be some subtle difference I'm not aware of. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus On Wed, 7 Nov 2001, Sunny Dubey wrote: hey, what does it mean to have an S or an s when doing ls -l ?? ([EMAIL PROTECTED])(/)$ ls -l | grep home drwxrwsr-x8 root staff1024 Oct 15 12:02 home thanks for any info =) Sunny Dubey -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: copying/cloning root partition
Hi, Perhaps, after copying everything, you need to run lilo, as suggested by the error message you got. On the regular desktop system, after making your final copy, mount your finished disk (e.g. under /mnt) and run lilo in the new system: [EMAIL PROTECTED]:~# mount /dev/hdc1 /mnt [EMAIL PROTECTED]:~# chroot /mnt /sbin/lilo Then take the machine out of the desktop machine, put it in the PC104 and try and see if it boots then. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus On Mon, 5 Nov 2001, Stephen A. Witt wrote: I'm trying to figure out how to clone the root partition of a Debian Linux installation. The installation is pretty small and hosted on a small, PC-104 based system that has only a root partition and runs without swap. I would like to be able to clone its root partition onto other disks but am running into a problem. I installed the disk from the PC-104 system in another regular desktop machine and tried to copy the image of the PC-104 system using: # dd bs=512 if=/dev/hdc1 of=disk.image I then tried to copy this new image onto another disk of the same type as running in the PC-104 system using (obviously installed in the desktop machine): # dd bs=512 if=disk.image of=/dev/hdc1 The partition image gets copied fine, but when I go to boot up this copied disk (after being re-installed in one of the PC-104 systems) it doesn't finish booting. I see LIL- when booting the PC-104 system. This, according to the LILO docs indicates that the descriptor table is corrupt, caused by a geometry mismatch or by moving /boot/map without running /sbin/lilo. I'm confused by this as it seems that the low level copy with 'dd' should copy everything. Any ideas?? Thanks... -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Console keymaps for woody
Hi, I'm trying to install the Greek keymap on woody, but not having much success. dpkg-reconfigure console-common returns the following, after exiting the interactive screens: Looking for keymap to install: gr findfile(): timeout waiting for undead child(ren) ? Trying the same with a British keymap returns no errors, but I don't get the actual keymap. Any pointers? George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
woody console-tools
Hi, I tried to install the console-tools on woody, but configuration failed. I tried to fix it and got the following errors: phobos:~# apt-get -f install Reading Package Lists... Done Building Dependency Tree... Done 0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded. 1 packages not fully installed or removed. Need to get 0B of archives. After unpacking 0B will be used. Setting up console-tools (0.2.3-23.2) ... Looking for keymap to install: gr findfile(): timeout waiting for undead child(ren) ? dpkg: error processing console-tools (--configure): subprocess post-installation script returned error exit status 70 Errors were encountered while processing: console-tools E: Sub-process /usr/bin/dpkg returned an error code (1) I try to reconfigure it, in hopes that choosing something other than the Greek keymap will help, but no joy: phobos:~# dpkg-reconfigure console-tools /usr/sbin/dpkg-reconfigure: console-tools is not fully installed Any pointers? George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: 2 partitions - same dir?
Hi, There are several approaches to your problem. It is indeed possible to make one *filesystem* span two partitions, by making the two partitions into what's called a linear or append multiple device, but this is not the simplest solution, and in my opinion it offers little in return for the effort of setting it up. Briefly, this would entail recompiling the kernel with the option Block Devices - Multiple Devices Support - Linear (append) mode enabled. If you're using a 2.2 kernel, it could be a good idea to install the patch for new-style RAID before doing this, but you should be OK with using the standard kernel source. Then you'd have to back up all the data in the existing partition, make a multiple device out of this and the new partition, make an ext2 filesystem on the multiple device, mount the filesystem, and restore your data into the new filesystem. And then you would be running the risk of having your filesystem trashed if *either* partition goes bad. Hardly worth it, in my humble opinion. Instead, the simplest and safest solution for distributing a *directory* over two partitions is: split up the subdirectories of the directory between the two partitions, by making mount points under the directory. For example, if this is your personal machine and you're the only user, make an ext2 filesystem in the new partition, make a mount point under your home directory, and just mount the new filesystem there. Move (copy first; then check; then delete) all your biggest subdirectories into it to free up the existing space in your home directory. Make the appropriate entry in /etc/fstab so that it's mounted at boot time and you're set. All files you save in subdirectories of the mount point will be in the new partition, all others in the old one, and they'll all be under your home directory. If this is a multi-user machine, work similarly with a new mount point under /home. Keep some users' home directories in /home and some in /home/newmountpoint; these latter users will need the entries for their home directories changed in /etc/passwd, to add the new level. If you want all home directories to be at the same level, unmount /home, make two mount points under /home for the two filesystems, and mount each partition on one. Entries for user home directories in /etc/passwd will then have to be changed for all users. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus On Mon, 5 Nov 2001, Rory O'Connor wrote: I've got a drive that's completely full of files, mounted at /home/dir. I added a second drive so I could continue storing files of the same type, and mounted it at /home/dir2. Is there a way to make it so that it appears these files are all in one dir? I don't think it's possible to mount two partitions at the same point...but perhaps with dynamic links? thanks, rory -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Sangoma S514 Frame Relay card
Hi, Has anyone on the list successfully used a Sangoma S514 Frame Relay card with Debian, potato to be more specific? Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: XF86Setup for woody?
On 31 Oct 2001, Brian Nelson wrote: I don't know about a card database, but you want the nv driver. Thanks, got that from xfree86.org. I seem to be on the right track now, thanks to everyone who helped. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: XF86Setup for woody?
On Tue, 30 Oct 2001, D. wrote: Best regards, George if when you did the dist-upgrade you got xfree4.1.x do a apt-get install xserver-xfree86 Then do a dpkg-reconfigure xserver-xfree86 that should bring up a set of configuration screens for XFree86-4 HTH Don Thanks, Have done as above. The configuration screens ask me to choose a driver. Where can I find info on which driver to choose for my card? (Nvidia TNT2 w.32MB RAM). xf86config and xf86setup used to show a card database so you could choose, but not here. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: XF86Setup for woody?
On Tue, 30 Oct 2001, D. wrote: Best regards, George if when you did the dist-upgrade you got xfree4.1.x do a apt-get install xserver-xfree86 Then do a dpkg-reconfigure xserver-xfree86 that should bring up a set of configuration screens for XFree86-4 HTH Don On Wed, 31 Oct 2001, George Karaolides wrote: Thanks, Have done as above. The configuration screens ask me to choose a driver. Where can I find info on which driver to choose for my card? (Nvidia TNT2 w.32MB RAM). xf86config and xf86setup used to show a card database so you could choose, but not here. Sorry to be replying to my own post, but info on xfree86.org pointed me to the nv driver. I now seem to be on the right track, thanks for the pointer, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
XF86Setup for woody?
Hi, I'm having my first go at woody on a spare machine. I have installed X, but not configured it yet. Is there a way to run XF86Setup on woody? apt-get install xf86setup returns an error: phobos:~# apt-get install xf86setup Reading Package Lists... Done Building Dependency Tree... Done Package xf86setup has no available version, but exists in the database. This typically means that the package was mentioned in a dependency and never uploaded, has been obsoleted or is not available with the contents of sources.list However the following packages replace it: xbase-clients E: Package xf86setup has no installation candidate I seem to have xbase-clients already installed: phobos:~# apt-get install xbase-clients Reading Package Lists... Done Building Dependency Tree... Done Sorry, xbase-clients is already the newest version. 0 packages upgraded, 0 newly installed, 0 to remove and 13 not upgraded. I have had no joy so far from xf86config. Pointers, anyone? Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Microsoft bullies again
On Sun, 28 Oct 2001, Greg Madden wrote: More wasted bandwidth, No, it didn't deny intrance it redirected you to a site that suggests you download a MS product., while suggesting MS was standards compliant whatever you were using was not. FWIW, Konq Opera passed the W3C validator, Netscape didn't., don't use IE. - -- Greg Madden Now if only there were a way for W3C to forbid you from calling your site a www-site unless all the content was W3C standards-compliant, the situation would have been quite different; you would have to call it msie.whatever instead of www. It would have also been nice if the W3C could force you to include a warning on your site if it includes non-compliant HTML content, and a warning about sugegsting that people download non-compliant browsers: msie.msn.com: This site uses HTML in a way not compliant with the standards set by its designers. Standards-compliant browsers may not display the contents of this site in the manner which the site owners intended. For this purpose, the site owners suggest you downgrade to a non-compliant browser, which may not function with standards-compliant material and is available only for two families of proprietary operating systems; click on the link below to download. If only it could be done, the community that invented the standards Microsoft are perverting and abusing could give them a healthy taste of their own medicine. George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Some routing advice (connecting through SSH)
On 26 Oct 2001, Adam Warner wrote: On Fri, 2001-10-26 at 03:07, George Karaolides wrote: Now to determine some more facts about the network geometry. I assume that machine R at your institution has one interface connected to the Internet, with a public IP address, and one on the institution's LAN with a private IP address. Just one public IP address. But after Code Red they unilaterally firewalled all incoming connections, even to the Dept's web servers! (something I had to alert people about). I'm not serving public content on this machine. OK, so machine R has one public IP address, routed through your institution's gateway/firewall. It's well firewalled locally (iptables). I'm pretty sure no one will be able to connect from anywhere else (I'm employing IP address checking, port blocking and of course password protection). Ping is global but that's because I believe people should be able to check if a machine connected to a public IP address is functioning. Your security sounds OK, bit do look at some kernel settings in /proc. For example, enabling syncookies is a good idea, and disabling replies to broadcast pings: echo 1 /proc/sys/net/ipv4/tcp_syncookies echo 1 /proc/sys/net/ipv4/icmp_echo_ignore_broadcasts Also, that the services you want to access are also on the institution's LAN. I think access to services is determined by network card mac address. I think the following would work: 1. Set up an IP tunnel between machines H and R. Now I haven't done this before but I know it can be done. Look for IP:tunneling (CONFIG_NET_IPIP) in the kernel configuration options, under Networking options. Quoting from the help on this: This particular tunneling driver implements encapsulation of IP within IP, which sounds kind of pointless, but can be useful if you want to make your (or some other) machine appear on a different network than it physically is...check out http://anchor.cs.binghamton.edu/~mobileip/LJ/index.html; which kind of sounds like what you need. As I said, I haven't tried this before, but I am virtually sure that you use this to set up a network interface representing the entrance of the tunnel. 2. Set up the routing table on machine H to route all traffic destined for your institution's network IP address space (get that from your friendly admin, if you haven't got it already) to use the tunnel interface. 3. On machine R, enable IP masquerading, with the tunnel interface as the internal interface and the machine's actual publicly available interface as the world interface. This should be the basis for your solution. The routing on machine H will make it access the machines at your institution through the tunnel and machine R, not the Internet. Masquerading on R will make those machines think they are being accessed by R instead of H, which is what you want. They will reply to R, and the demasquerading will then forward everything back to H. Linux networking magic at its best. I am also virtually sure you can build this to work for all machines in your private LAN at home, with machine H as gateway. Though I have no hands-on experience of this, I will, of course, try and help out with any questions of yours which might arise if you do try it, to the best of my ability. Do let me know how you get on! Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Some routing advice (connecting through SSH)
Hi, I need a bit more info. to start thinking about your question. How do you connect from H to R? Through the Internet? If so, does your Internet connection (that you use to connect H to the Internet) have a static IP address or a dynamic one? Is is permanent (DSL, Frame Relay, ATM etc.) or dialup (PSTN, ISDN etc.)? Another point: It seems to me as if you'll be trying to use a server hosted at an ISP as a masquerading gateway for your home LAN. You may well be able to do it, and the ISP may well not catch you at it, for a while anyway, but if the ISP sees this as violating the service agreement they have with whoever they're hosting the server for (you?) and they do find out, you are in for trouble. The traffic won't be difficult to spot; servers are expected to be serving out a lot of stuff, not to have a lot of incoming traffic. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus On 25 Oct 2001, Adam Warner wrote: Hi all, I just need to understand where I should look and how I should approach this challenge. I want to route some traffic though a remote computer (R) to my home computer (H). In particular I want to have the ability to surf the Web as if I was sitting at computer R. Right now I can already do that using the text browser Lynx after connecting via SSH. R is a somewhat puny 133MHz Pentium with 72MB of RAM and ~100MB of free disk space. It is running Debian GNU/Linux with a 2.4.13 kernel (that took a rather long time to compile). X is not installed (the display card is also not compatible, but I imagine that wouldn't matter with a remote connection). I can SSH from H to R. All other ports to R are blocked. So to connect to another port on R, R itself would have to open the connection to H. Instead of using X or VNC I would like to somehow use IP Masquerading for just some chosen traffic (it would be the most efficient solution). However I can't see how it would work yet. So thanks for any preliminary help. Regards, Adam -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Some routing advice (connecting through SSH)
Hi, On 26 Oct 2001, Adam Warner wrote: On Fri, 2001-10-26 at 01:04, George Karaolides wrote: Hi, I need a bit more info. to start thinking about your question. I'm still up, barely :-) How do you connect from H to R? Through the Internet? Yes. If so, does your Internet connection (that you use to connect H to the Internet) have a static IP address or a dynamic one? Static, permanent connection. Good, that makes things somewhat easier. Another point: It seems to me as if you'll be trying to use a server hosted at an ISP as a masquerading gateway for your home LAN. Hey, hold on a minute! You're _way_ off base. What? Server hosted at an ISP? No! You may well be able to do it, and the ISP may well not catch you at it, for a while anyway, but if the ISP sees this as violating the service agreement they have with whoever they're hosting the server for (you?) and they do find out, you are in for trouble. For what? I won't be in trouble for anything. Please don't jump to conclusions. If I was trying to do something naughty I wouldn't be using my real name (and please note that I use my real name in _all_ my communications). Sorry if I offended you, but I said It sounds like... and If... I didn't think you were delliberately trying to swindle an ISP. I had your best interests in mind. Thanks anyway for being concerned. I only want to implement a very limited system for only specific browser traffic so I can perform searches on databases accessible at my Institution from my home computer (just like having a remote desktop, but as I say I'd like it to be more efficient). And I discussed it with my Department's computer support (his suggestion was to tunnel X through SSH). Even though I've never done that before I'd like to try and do it more efficiently (and also work more productively--If you go to save a document on a remote desktop it has only been saved to that computer. Then I'd need to use something like sftp to download it to my computer. Much better if I could operate as if I was at that machine). The traffic won't be difficult to spot; servers are expected to be serving out a lot of stuff, not to have a lot of incoming traffic. Did I say there would be a lot of traffic or did I specifically say for just some chosen traffic? My Institution will be able to see all traffic going to my R computer. I would be very foolish to route personal traffic through my Institution's network. Sorry again, I didn't get exactly what you were trying to do from your first message. Now to determine some more facts about the network geometry. I assume that machine R at your institution has one interface connected to the Internet, with a public IP address, and one on the institution's LAN with a private IP address. Also, that the services you want to access are also on the institution's LAN and also have private IP addresses. Is that correct? Best regards (it's still afternoon here), George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
motd on potato
Hi, Is there a way to stop Debian potato from adding the output of `uname -r` to the beginning of /etc/motd? Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: motd on potato
On Mon, 22 Oct 2001, Paul 'Baloo' Johnson wrote: On Mon, 22 Oct 2001, George Karaolides wrote: Is there a way to stop Debian potato from adding the output of `uname -r` to the beginning of /etc/motd? Go edit /etc/default/rcS and make sure it says EDITMOTD=no at some point. There's some other nifty options there, too. Nice one, thanks! George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: ip_always_defrag
On Tue, 16 Oct 2001, John Patton wrote: I'm not sure about what ip_always_defrag is meant to return when called like that (probably some info that is useful to the developers), but you can set it in the following way: echo 1 /proc/sys/net/ipv4/ip_always_defrag Thanks for the reply. That is the way I did set it, using echo 1 ... . But when I set any other kernel setting, for example /proc/sys/net/ipv4/ip_forward, I can confirm its setting using cat ... ; it returns 1 or 0 according to whether it has been set or not. I can't confirm whether /proc/sys/net/ipv4 has been set because cat ... returns a varying integer like 76 or -25 instead of 1 or 0; the value changes every few seconds. That is my problem. This is different to all the other kernel settings. I need to know for sure whether this setting is enabled or not for security reasons. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Non-Debian kernel source
Vittorio == vdemart Vittorio writes: Vittorio Hi, I'm using potato with Bunk's 2.4.9 kernel suitably Vittorio compiled for my hardware. Vittorio Now I wonder if - according to your very revered opinion Vittorio - it would be reproachable, disgusting to install a Vittorio 2.4.12 (or better the forthcoming 2.4.13) kernel Vittorio downloaded directly from www.kernel.org and compiled in Vittorio the ordinary way make dep clean bzImage . Vittorio I mean: what really is against it? On 17 Oct 2001, Brian Button wrote: Nothing at all, other than risk. There have been a few problems with 2.4.11 and 2.4.12 as they first came out, but its always been easy to go back to a previous kernel. I download the kernel source bz2 from kernel.org and follow the instructions in /usr/share/doc/debian/FAQ about building kernels with make-kpkg. I've also done it manually as you say, with make dep clean bzImage modules modules_install, but make-kpkg is *much* easier. This route will build a .deb for you that will automatically install your kernel, update maps, set symlinks, etc. It will even run /sbin/lilo for you. Just edit /etc/lilo.conf before you install this new package to add a stanza for your new kernel, and everything else happens for you. I second that. All you need to do is run make-kpkg --**check manpage for appropriate options*** kernel_source to make a debianised source tree. Then do make config make-kpkg clean make-kpkg --**check manpage for appropriate options*** kernel_image One point: use --flavour (on stable) or --add_to_version (on testing/unstable) to add an extra string to your kernel version, the date of your compilation for example. In this way, you can compile several different builds of the same kernel version without over-writing the modules, which are stored in /lib/modules/kernel-version-with-any-extra-string. In this way you could build and boot several builds of the same kernel version. E.g. if yesterday you did make-kpkg --add_to_version -2001.10.16 and today you did make-kpkg --add-to-version -2001.10.17 the modules would be stored in the following two directories: /lib/modules/2.4.12-2001.10.16 /lib/modules/2.4.12-2001.10.17 whereas if you didn't, they'd be stored in /lib/modules/2.4.12 and today's modules would overwrite yesterday's. In case you want to do this using the source directly, then edit the Makefile in the top level of the kernel source tree and add your extra string to the variable EXTRAVERSION. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
ip_always_defrag
Hi, I have a strange situation on my main gateway (firewall + masquerade using the ipmasq package on Debian 2.2r3 potato). The command cat /proc/sys/net/ipv4/ip_always_defrag returns an integer which seems to vary all the time. I would like to set this (to 1) for security reasons, but seem to be unable to do so because of this. Any ideas, anyone? George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Blocking streaming audio and video
Hi, This may well be a bit off-topic, but how do I block users on the LAN from accessing streaming audio and/or video data from the Internet and filling up the Internet line? I'm running Debian GNU-Linux 2.2r3 (potato) on the gateway. I use the ipmasq Debian package to set up masquerading and firewall rules between our internal nets and our ISDN dialup Internet connection. Thanks and best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
/sbin/ipchains: host/network `domain.name' not found
Hi, I'm running Debian GNU-linux 2.2r3 (potato) and trying to set up firewall rules with sources and destinations specified as domain names rather than IP addresses. The problem is that ipchains returns an error: /sbin/ipchains: host/network `domain.name' not found Try `/sbin/ipchains -h' or '/sbin/ipchains --help' for more information. The host this is running on can resolve the domain names OK. Anyone know what could be wrong? Thanks and best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: Blocking streaming audio and video
Hi, Sorry to be posting a followup to my own question, but I think I have a solution for this. I found some ipchains rules at: http://nw-hoosier.dyndns.org/rlohman/linux/firewall/es-block.html and put these in post-processing rules files for ipmasq, separately for internal and external interfaces. It remains to be seen how effective these are. Best regards, This may well be a bit off-topic, but how do I block users on the LAN from accessing streaming audio and/or video data from the Internet and filling up the Internet line? I'm running Debian GNU-Linux 2.2r3 (potato) on the gateway. I use the ipmasq Debian package to set up masquerading and firewall rules between our internal nets and our ISDN dialup Internet connection. Thanks and best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: /sbin/ipchains: host/network `domain.name' not found - SOLVED
Sorry to be posting a followup to my own question, but I solved this. I was using ipmasq to set up rules which needed to resolve external domain names before actually allowing external traffic... A bit stupid, but there it is. Putting my rules in post-processing rules files: /etc/ipmasq/P30internal.rul /etc/immasq/P90external.rul which are executed after the internal and external input files in which I initially put them: /etc/ipmasq/I30internal.rul /etc/ipmasq/I90external.rul solved it. I'm running Debian GNU-linux 2.2r3 (potato) and trying to set up firewall rules with sources and destinations specified as domain names rather than IP addresses. The problem is that ipchains returns an error: /sbin/ipchains: host/network `domain.name' not found Try `/sbin/ipchains -h' or '/sbin/ipchains --help' for more information. The host this is running on can resolve the domain names OK. Anyone know what could be wrong? Thanks and best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Re: how to download debian
Hi, re. your following: On Wed, 3 Oct 2001, Suresh Kumar R wrote: Hi, I tried downloading debian using psudo image kit. Many servers I tried didnt allow downloading. It just stops connecting. Somesites allows to download upto a few Mbs and then halts. Reconnection doesnt help Where can we download it from? Thanks I think it's a much better idea to download the distribution package by package, rather than a big image. Especially if you don't have a *very* fast internet connection. You will have the big advantage that you can keep your mirror up to date without downloading another huge image. All this is automated in Debian and not tedious at all. Debian has tools which allow you to download the distribution on an as needed basis, and organise the packages in a local partial mirror of the Debian site. You can keep extending and updating your mirror as you download and install more packages. To start with, all you need is the installation boot (rescue), root and driver floppuy images; if you choose the compact dist. that's just three floppies. During installation, tell the installer to install the kernel and drivers from floppy, and use the floppies you made. Then choose to install the base system from the network and the installer will download the base system for you, which is only a dozen megs or so. You can then put this on a local webserver so that any further machines you build will download it from your local net instead of the Internet. After that, tell the installer to download the core packages from the net. They will be cached on disk and you can use apt-move to move these to a directory which will become your local partial Debian mirror. If you try this and need more help, post again in this thread and I'll do my best. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Debian package security
Hi all, It seems to me that this message belongs here rather than on debian-security or debian-security-announce, but please correct me if I'm wrong. :) I am running Debian potato on a number of machines. I have downloaded potato from the net, and used apt-move to make a local mirror on one of my local servers. I add any new packages I download to this mirror. I would like to check the packages on my mirror for security, against the packages in the official Debian site or an official mirror thereof. I'm sure that a distribution as security-conscious as Debian must provide tools for this. Any pointers would be greatly appreciated. Best regards, George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
Two ethernet cards, one IP?!?
Hi all, I have an Intel Nightshade server motherboard installed in an Intel server case, running Debian 2.2r3 (potato). Its on-board Intel EtherExpress Pro 100 ethernet card works fine on its own. When I install an Intel EtherExpress Pro 100 PCI card and configure two ethernet interfaces, the PCI card answers to both IP addresses! ifconfig returns separate hardware addresses for the two interfaces, yet when I disconnect the cable to the on-board card, both interfaces continue to function through the PCI card! prompt:~# ifconfig eth0 Link encap:Ethernet HWaddr 00:D0:B7:7F:52:0C inet addr:192.168.2.9 Bcast:192.168.2.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:14753 errors:0 dropped:0 overruns:0 frame:0 TX packets:36664 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:100 Interrupt:10 Base address:0x8000 eth1 Link encap:Ethernet HWaddr 00:90:27:73:C3:99 inet addr:192.168.2.8 Bcast:192.168.2.255 Mask:255.255.255.0 UP BROADCAST MULTICAST MTU:1500 Metric:1 RX packets:9 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:100 Interrupt:11 Base address:0xa000 loLink encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:3924 Metric:1 RX packets:6 errors:0 dropped:0 overruns:0 frame:0 TX packets:6 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 I'm running kernel version 2.2.19 with the eepro100 driver compiled into the kernel, not as a module, but this does not seem to be the cause of the problem; other machines running with this code compiled into the kernel have no problem with two EEpro100's. This is certainly freaky... George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus
RE: Two ethernet cards, one IP?!?
On Tue, 25 Sep 2001, Theo Zourzouvillys wrote: I have an Intel Nightshade server motherboard installed in an Intel server case, running Debian 2.2r3 (potato). Its on-board Intel EtherExpress Pro 100 ethernet card works fine on its own. When I install an Intel EtherExpress Pro 100 PCI card and configure two ethernet interfaces, the PCI card answers to both IP addresses! ifconfig returns separate hardware addresses for the two interfaces, yet when I disconnect the cable to the on-board card, both interfaces continue to function through the PCI card! *snip* I'm running kernel version 2.2.19 with the eepro100 driver compiled into the kernel, not as a module, but this does not seem to be the cause of the problem; other machines running with this code compiled into the kernel have no problem with two EEpro100's. This is certainly freaky... I have had *exactly* the same problem on a number of Intel ISP1100's, with dual on board eepro100's, on both 2.2 and 2.4. I've never looked into it, as in the setup I was using; it was more of an advantage than a problem, made my life a lot easier :) I would be interested to know why though, as it may be a problem one day. Theo Thanks for the reply. In my situation, this is a serious disadvantage; I will be running this server hosted at an ISP. One ethernet card will be the world interface, the other will be the interface to our company WAN. I certainly need to distinguish between them. I hope this isn't an incurable trait of the Intel server motherboards. I certainly don't look forward to having to convince the bosses to scrap two perfectly healthy server boards, with on-board SCSI too, because of such a silly quirk. George Karaolides 8, Costakis Pantelides St., tel: +35 79 68 08 86 Strovolos, email: [EMAIL PROTECTED] Nicosia CY 2057, web: www.karaolides.com Republic of Cyprus