December 2006 - Grml - lists.mur.at

sarge + kernel 2.6.18-grml loads only few modules
by Erich Minderlein 20 Dec '06

20 Dec '06

Hi after having successfully installed kernel 2.6.18-grml and a sarge base configuration via the grml-skript I have stopped working on this notebook. Now I take up on this again : It loads only few modules, so no network, no periphery: just the 6 shells and the ipv6 module is loaded. /sys/module is rather empty : about 15 modules are listed , for HDD, CDROM and video, while at the funkentzutzler there are 74 modules listed. the line "populating /dev" does not show up in dmesg when doing the boot. It seems that some skripts must be added to my installation to get the kernel scan its periphery. a shortened diff ps(grml0.8,left) ps(kernel2.6.18-grml+sarge,right) is added: PID TTY STAT COMMAND PID TTY STAT COMMAND 1 ? Ss init [2] | 1 ? Ss init [2] 2 ? S [migration/0] 2 ? S [migration/0] 3 ? SN [ksoftirqd/0] 3 ? SN [ksoftirqd/0] 4 ? S [watchdog/0] 4 ? S [watchdog/0] 5 ? S< [events/0] 5 ? S< [events/0] 6 ? S< [khelper] 6 ? S< [khelper] 7 ? S< [kthread] 7 ? S< [kthread] 9 ? S< [kblockd/0] | 10 ? S< [kblockd/0] 10 ? S< [kacpid] | 11 ? S< [kacpid] 94 ? S< [kseriod] | 96 ? S< [cqueue/0] 148 ? S [pdflush] | 97 ? S< [kseriod] 149 ? S [pdflush] | 157 ? S [pdflush] 150 ? S [kswapd0] | 158 ? S [pdflush] 151 ? S< [aio/0] | 159 ? S< [kswapd0] 152 ? S< [jfsIO] | 160 ? S< [aio/0] 153 ? S< [jfsCommit] | 161 ? S< [jfsIO] 154 ? S< [jfsSync] | 162 ? S< [jfsCommit] 155 ? S< [xfslogd/0] | 163 ? S< [jfsSync] 156 ? S< [xfsdatad/0] | 164 ? S< [xfslogd/0] 745 ? S< [cqueue/0] | 165 ? S< [xfsdatad/0] 757 ? S< [kpsmoused] | 294 ? S< [kpsmoused] 832 ? S< [ata/0] | 384 ? S< [ata/0] 841 ? S< [scsi_eh_0] | 385 ? S< [ata_aux] 842 ? S< [scsi_eh_1] | 394 ? S< [scsi_eh_0] 868 ? S< [kcryptd/0] | 395 ? S< [scsi_eh_1] 869 ? S< [kmirrord] | 416 ? S< [kcryptd/0] 943 ? S< [khubd] | 417 ? S< [kmirrord] 989 ? S< [scsi_eh_2] | 423 ? S< [kjournald] 990 ? S< [usb-storage] | 807 ? Ss /sbin/syslogd 1944 ? S< [loop0] | 810 ? Ss /sbin/klogd 2215 ? S<s udevd --daemon | 842 ? Ss /usr/sbin/exim4 -bd -q30m 3837 ? S [khpsbpkt] | 848 ? Ss /usr/sbin/inetd 3934 ? S [pccardd] | 855 ? Ss /usr/sbin/sshd 3935 ? S [knodemgrd_0] | 868 ? Ss /sbin/mdadm -F -i /var/run/mdadm.p 3993 ? S< [shpchpd] | 871 ? Ss /usr/sbin/atd 4017 ? S< [kgameportd] | 874 ? Ss /usr/sbin/cron 4504 ? Ss /usr/sbin/acpid -c /etc/acpi/event | 882 tty1 Ss -bash 4625 ? SNs pump -i eth0 | 885 tty2 Ss+ /sbin/getty 38400 tty2 4680 ? SNs /sbin/syslog-ng -p /var/run/syslog | 886 tty3 Ss+ /sbin/getty 38400 tty3 4695 ? SNs /usr/sbin/powernowd -q | 887 tty4 Ss+ /sbin/getty 38400 tty4 4719 ? SNs /usr/sbin/gpm -m /dev/psaux -t aut | 888 tty5 Ss+ /sbin/getty 38400 tty5 4759 tty1 Ss /bin/zsh -l | 889 tty6 Ss+ /sbin/getty 38400 tty6 4760 tty2 Ss+ /bin/dash /usr/bin/grml-init-scree | 910 tty1 R+ ps ax 4761 tty3 Ss+ /bin/dash /usr/bin/grml-init-scree < 4762 tty4 Ss+ 0:00 /bin/zsh < 4763 tty5 Ss+ 0:00 /bin/zsh < 4766 tty6 Ss+ 0:00 /bin/zsh < 4767 tty9 Ss+ 0:00 /sbin/getty 38400 tty9 < 4768 tty10 Ss+ 0:00 /bin/zsh --login -c /usr/bin/iptst < 4771 tty11 Ss+ 0:00 /bin/zsh --login -c /usr/bin/htop < 4774 tty12 Ss+ 0:00 /bin/zsh --login -c /usr/bin/multi < 4822 tty10 S+ 0:00 /usr/bin/iptstate < 4824 tty11 S+ 0:33 /usr/bin/htop < 4825 tty12 S+ 0:00 /usr/bin/multitail /var/log/syslog < 4834 tty12 S 0:00 tail -48lf /var/log/syslog < 4841 tty2 S+ 0:00 /usr/bin/screen -c /etc/grml/scree < 4845 ? Ss 0:00 /usr/bin/SCREEN -c /etc/grml/scree < 4848 tty3 S+ 0:00 /usr/bin/screen -c /etc/grml/scree < 4849 ? Ss 0:00 /usr/bin/SCREEN -c /etc/grml/scree < 4869 pts/0 Ss+ 0:00 zsh < 4873 pts/1 Ss+ 0:00 zsh < 5059 ? S< 0:00 [kjournald] < 5078 ? S< 0:00 [kjournald] < 5219 tty1 R+ 0:00 ps ax < >From the process numbers I see that a lot of transient processes did not happen on my install, and usb-storage, cqueued, loop0, udevd do not exist thereafter. best regards -- Erich Minderlein <erminderlein(a)locoware.de> MARC Mannheim Rail Commerce GmbH

2 1

Lilo and the boot-prompt
by keithint38＠commspeed.net 20 Dec '06

20 Dec '06

Greetings friends, I am a Visually impaired user of GRML, and noticed that LILO brings up this boot-prompt thing, that isn't exactly that useful. I was wondering how to either perhaps get rid of the prompt, or at least how to set the timeout option to 0 so that LILO doesn't have a chance to display the prompt, and automatically boot up GRML without waiting for user-interaction? If that isn't possible, then I may switch to GRUB, eaven though I'd rather not. Regards, --Keith

2 1

primitive lilo question
by Martin Yazdzik 19 Dec '06

19 Dec '06

Dear Friends at grml, google lilo information tends to be rather out of date, to say the least - since grml 0.9 works, well, just works, I need to make what was my experimental partition my main OS. /dev/sda1 = windows /dev/sda2 = swap /dev/sda3 = old grml /dev/sda4 = new grml currently, the laptop boots from the old grml, that is lilo is installed to the mbr, but goes to the root at old grml and uses the "other OS" to boot into the new grml. I wish to have the mbr point to the new grml, in other words, use the sda3 partition for either data, or one of my experiments. So, how can I re-isntall lilo to the mbr from the new grml without making a mess? Will lilo -v -v just work? And, yes, I apologise and know that this is not the right place for this question. Best, M

2 2

discover and rtl8111 ethernet
by Markus Wigge 15 Dec '06

15 Dec '06

Hi, I compiled an r1000.ko module for my PXE bootenvironment and added it to the initrd image but I don't understand the discover tool yet, so I just added r1000 to the discovered result as dirty hack... Is there a possibility to discover the hardware automagically? /Markus

2 1

grml-small PXE kernel
by Markus Wigge 15 Dec '06

15 Dec '06

Hi, I managed to add a section to linuxrc of the netboot 0.8 package which downloads the seconds stage via http, but after updating the files to the grml-small 0.3 kernel and modules I noticed it lacks unionfs support?! So can anyone give me a small hint how to compile a customized grml-small kernel I can use for PXE booting? BTW: My intention is to prevent setting up a NFS server in my network... bye, Markus

2 7

hints about grml 0.9 on HD and ipw3945
by Lars Schimmer 14 Dec '06

14 Dec '06

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi! Today I installed grml 0.9 (two times) to HD with grm2hd. If I boot with the CD the autoconf detects eth0 (e1000) and eth1 (ipw3945) from my T60 laptop. After install on HD and reboot the system just recognize eth0. After rmmod ipw3945 /modprobe ipw3945 (and for some test another /etc/init.d/ipw3945d restart) I get a eth1 but with a dhclient eth1 I can?t get a IP. While booting from the live CD I just need to wait a few sec and retry dhclient eth1 and the LED for the WLAN activity is enlighten and blinks. With boot from HD it is most time dark, just a part of second enlightent. Oh, I use grml-network with WPA supplicant options. Anyone has a idea? Is ipw3945 buggy? Yes, it is a clean install, an now the (hopeful) exact steps: boot from CD: boot with standard option grml-network (enter correct SSID, standard options, WPA, enter correct Key), wait, break fist dhclient wait some sec dhclient eth1 *ipw3945 is active, WLAN LED is active and I got net access* HD install: - -grml2hd - -chroot, install grub - -reboot - -apt-get update, apt-get upgrade (NO grml, kenrel oder ipw3945 update!) - -grml-network (same as with CD boot) - -reboot - - after boot no eth1 present - - rmmod ipw3945 - - modprobe ipw3945 - - ifconfig - no eth1 - - ifconfig eth1 - eth1 shows up with correct MAC ID, no IP - - dhclient eth1 - WLAN HD is dark, some flickering and NO IP.. MfG, Lars Schimmer - -- - ------------------------------------------------------------- TU Graz, Institut fÃ¼r ComputerGraphik & WissensVisualisierung Tel: +43 316 873-5405 E-Mail: l.schimmer(a)cgv.tugraz.at Fax: +43 316 873-5402 PGP-Key-ID: 0x4A9B1723 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (MingW32) iD8DBQFFfIewmWhuE0qbFyMRAofEAKCGT6+xhpKhBiQwAafjFwHjtVAHNgCcD9IA Zpy8MVsm8fjbVLpiYZI4RA4= =d7ya -----END PGP SIGNATURE-----

3 3

Unlocking CD drive
by Michael Schierl 13 Dec '06

13 Dec '06

Hi, sometimes I booted grml, started a job there (like using dd to zero a partition) and then have to leave (but let it try continuing the zeroing). So I want to remove the CD (and risk a panic later when a file from it is accessed). But that is no problem since I do not run anything new after the CD is removed, so everything needed is already in cache. For these cases I have a perl script on my USB pendrive: #!/usr/bin/perl -w open(H, "</dev/cdrom") or die("Can't open: $!"); ioctl(H, 0x5329 , 0) or die("Can't ioctl: $!"); close(H) or die("Can't close: $!"); That one works well - if I have my pendrive with me. My question: Is there a way inside grml to do the same thing (without memorizing the whole script or the ioctl number)? Something in /proc or /sys? Or a script? (Using the hardware eject "hole" of the CD drive does not count, since it requires a paper clip ;-) ) TIA, Michael

5 9

noob scripting question
by Martin Yazdzik 12 Dec '06

12 Dec '06

Dear Friends at grml, Please forgive a really dumb question, but I cannot seem to get fluxbox to run the gnome-settings-daemon automatically, yet prefer to use flux as my wm. On the other hand, I must have my own fonts in apps, particularly evolution as my lappy has a uxga screen, which renders, for all practical purposes, apps illegible. So, I thought a little script in .xinitrc would work to launch the daemon, but, since gdm seems to bypass the xinitrc, where should I stick a script, presuming I can write one? This is not to say that gdm does not suck, merely that I need to boot up at the push of one button.... ;) New York is under way too much time stress. Best, Martin

2 1

Making an initrd.img for 2.6.18
by Tom White 09 Dec '06

09 Dec '06

I eventually got one to boot, but I suspect I'm missing something basic. mkinitrd couldn't find sd_mod: # mkinitrd -o initrd.img-2.6.18-grml -r /dev/sdb3 2.6.18-grml /usr/sbin/mkinitrd: add_modules_dep_2_5: modprobe failed FATAL: Module sd_mod not found. yaird worked for me going from 2.6.16 to 2.6.17, but now / isn't in fstab # yaird --output=initrd.img-2.6.18-grml 2.6.18-grml yaird error: mount point not in fstab: . (fatal) So I modified /etc/yaird/Default.cfg < MOUNTDIR "/" "/mnt" --- > MOUNTDIR "/mnt/sdb3" "/mnt" But yaird now didn't like /dev/usb-sdb3 in fstab # yaird --output=initrd.img-2.6.18-grml 2.6.18-grml yaird error: duplicate device name in fstab: /dev/sdb3 (fatal) So I commented it out # Added by GRML #/dev/usb-sdb3 /mnt/usb-sdb3 auto user,noauto,nodev,noatime,nosuid,noexec,rw,uid=1000,gid=1000 0 0 And success # yaird --output=initrd.img-2.6.18-grml 2.6.18-grml -- -- Tom White

2 1

boring
by Martin Yazdzik 08 Dec '06

08 Dec '06

Dear Mika and Friends, What a boring install! Everything worked otob on my dell laptop, and the tz config, which used to cause hair pulling was one simple command and voila. While I used the ati installer, which worked perfectly to get hardware accel, and ndiswrapper, because the bcm4306 driver still only allows "b" speeds, nothing took anytime at all. Now, here is the question from the idiot newbie: Since grml .9 installed easily and cleanly on /sda4, but, obviously, I could not risk the change in the production environment, can I more easily upgrade the existing grml .8 with all the data and config that I need to .9, without a lot of brimborium, or can I somehow install lilo to the mbr from the newly installed grml, and use the old partition for data or somthing, without making a mess that a novce cannot clean up? Is the apparent increased speed, by feel, not measured, due to newer kernel, better fglrx drivers, or what? Best wishes from a grateful user, Martin Chi troppo vuole nulla stringe

4 5