Suse 10.0 OSS boot freezes after "boot.local done"
Hi everybody, I try and try and try to install Suse 10.0 OSS. Boot hangs during first reboot after showing the last two messages: System Boot Control: The system has been set up System Boot Control: Running /etc/init.d./boot.local done I've been looking around for hours and hours in google and forums, finding a lot of similar problems but no solution that worked. I did the first install using a downloaded Boot-CD. Then I downloaded all 5 CD's, burned them slowly, and try to installs with them now. But ftp-/CD-install make no difference. Here's what I tried: -boot option "acpi=off" (I always used this, because Suse 9.0 needed it, too) -tried to start with "failsave" option - no difference -tried additional boot-parameter "BrokenModules=yenta_socket" as proposed in another forum - no difference - made a completely new text-only installation (without x), same result... - again a new text-only installation, but now everything on same partition (except swap) - same result My computer: Motherboard ASUS P4PE/GBL/SATA/1394/GD-UAY with Intel 845 PE Chipset Pentium P4 2.4 Ghz 2x512 MB Memory CD-/DVD-ROM-Drive CD-Burner Zip-Drive Promise Ultra Controller 3 HD's, one one motherboards controller, 2 on promise, all as Master devices Adaptec SCSI card NVIDIA Gforce4 TI graphic card boot-line in Suse 9 additionally contains "desktop hdd=ide-scsi hddlun=0", this is not in the boot-line (menu-list) of Suse 10. I've added this, too, but: no change, freezes on same place. I tried to come to another console before the freeze occurs, but ALT-Fx didn't show another console (nor did CTRL-ALT-Fx), I tried F1 to F12... If you need to know more, tell me what (and if it isn't obvious, tell me how to find out - I can still start Suse 9 and find out...). Oh: I'm everything else than a computer-crack, so maybe tell me how to obtain any information you need, please. I would very much appreciate to receive help! Thanks. Daniel
Hi, Daniel Bauer schrieb:
I try and try and try to install Suse 10.0 OSS. Boot hangs during first reboot after showing the last two messages:
System Boot Control: The system has been set up System Boot Control: Running /etc/init.d./boot.local done [...] Here's what I tried:
-boot option "acpi=off" (I always used this, because Suse 9.0 needed it, too)
Please retry without acpi=off and if that fails, use the boot option splash=none to be able to scroll back (with Shift-PageUp) and read the kernel messages. How does the machine freeze? Does it still respond to CapsLock? How long did you wait? Do any of the keyboard lights blink during the freeze? Can you hit Ctrl-Alt-Del? Regards, Carl-Daniel -- http://www.hailfinger.org/
Please retry without acpi=off and if that fails, use the boot option splash=none to be able to scroll back (with Shift-PageUp) and read the kernel messages.
How does the machine freeze? Does it still respond to CapsLock? How long did you wait? Do any of the keyboard lights blink during the freeze? Can you hit Ctrl-Alt-Del?
removing acpi=off didn't change anything, freezes on same point. The machine freezes after the said last message, cursor is blinking on the next line for a few seconds, then stops blinking. No keyboard lights blink during freeze. After this the keyboard doesn't respond at all, no reaction on CapsLock or Ctrl-Alt-Del, so I also wasn't able to scroll back to read all the kernel messages. I once left it in "frozen state" for approx. 2 hours, having a cup of coffre then. Usually I don't wait at all. I just try do hit something on the keyboard and then push the reset button of the PC...
boot option splash=none to be able to scroll back (with Shift-PageUp) and read the kernel messages.
I cant scroll thru the kernel messages, but I can read the /var/log/boot.msg file using the other OS. I don't know much about that, but when comparing to the file on Suse9 I saw, that Suse10 doesn't mount any disk (as far as I understand). The part of boot.msg of Suse10 looks like this: Mounting local file systems... proc on /proc type proc (rw) sysfs on /sys type sysfs (rw) tmpfs on /dev/shm type tmpfs (rw) devpts on /dev/pts type devpts (rw,mode=0620,gid=5) nothing was mounted done but in Suse 9 the partition-dev's follow the devpts-line. Could this point in direction of the problem?
On Mon, 19 Dec 2005, Daniel Bauer wrote:
boot option splash=none to be able to scroll back (with Shift-PageUp) and read the kernel messages.
I cant scroll thru the kernel messages, but I can read the /var/log/boot.msg file using the other OS. I don't know much about that, but when comparing to the file on Suse9 I saw, that Suse10 doesn't mount any disk (as far as I understand). The part of boot.msg of Suse10 looks like this:
Mounting local file systems... proc on /proc type proc (rw) sysfs on /sys type sysfs (rw) tmpfs on /dev/shm type tmpfs (rw) devpts on /dev/pts type devpts (rw,mode=0620,gid=5) nothing was mounted done
Things are a bit difficult when you cannot get more information. I guess, that the system stops working in boot.udev, which comes after boot.localfs at least on my system. What you can do, is boot the system with init=/bin/sh then run the different boot scripts in /etc/init.d, e.g. run /etc/init.d/boot.localfs start /etc/init.d/boot.udev start and see if it crashes there. Maybe even something like echo "foobar" > /dev/shm/foobar instead of the boot.udev kills the machine, then it would be a shm problem... Sorry, I have no idea what the cause is, and can only give some hints about how to do a bit of debugging. Berthold -- ------------------------------------------------------------------ Berthold Gunreben SUSE Linux GmbH -- Dokumentation mailto:bg@suse.de Maxfeldstr. 5 http://www.suse.de/ D-90409 Nuernberg, Germany
Mounting local file systems... proc on /proc type proc (rw) sysfs on /sys type sysfs (rw) tmpfs on /dev/shm type tmpfs (rw) devpts on /dev/pts type devpts (rw,mode=0620,gid=5) nothing was mounted done
Things are a bit difficult when you cannot get more information. I guess, that the system stops working in boot.udev, which comes after boot.localfs at least on my system. What you can do, is
boot the system with init=/bin/sh then run the different boot scripts in /etc/init.d, e.g. run /etc/init.d/boot.localfs start /etc/init.d/boot.udev start
and see if it crashes there. Maybe even something like echo "foobar" > /dev/shm/foobar
instead of the boot.udev kills the machine, then it would be a shm problem...
hmmm, sorry, I didn't understand too much ;-) So I booted with the normal bootline from the menu list but additionally inserted "init=/bin/sh" right after "acpi=off". I had a shell prompt. I typed "/etc/init.d/boot.localfs start" had the messages as above (mounting... nothing was mounted...) typed in "/etc/init.d/boot.udev start" had a shell prompt again - at least it didn't freeze... I typed "yast" just to see if it starts - and it did (didn't know what else I could type...). I just quit Yast without doing anything and shutdown. I didn't understand what you mean with the foobar thing... But now at least I have a shell prompt - if you can tell me what to do, I hope I can deliver the infos you need... b.t.w. I have no idea what "shm" is.
Hello, Am Montag, 19. Dezember 2005 19:56 schrieb Daniel Bauer:
boot the system with init=/bin/sh then run the different boot scripts in /etc/init.d, e.g. run /etc/init.d/boot.localfs start /etc/init.d/boot.udev start [...] hmmm, sorry, I didn't understand too much ;-) So I booted with the normal bootline from the menu list but additionally inserted "init=/bin/sh" right after "acpi=off". I had a shell prompt. I typed "/etc/init.d/boot.localfs start" had the messages as above (mounting... nothing was mounted...) typed in "/etc/init.d/boot.udev start" had a shell prompt again - at least it didn't freeze...
That means that your problem comes from one of the init scripts you didn't test ;-)
I didn't understand what you mean with the foobar thing... But now at least I have a shell prompt - if you can tell me what to do, I hope I can deliver the infos you need... b.t.w. I have no idea what "shm" is.
shm is a tmpfs (a filesystem living in RAM, not hard disk) that is mounted to /dev/shm. The echo-command Berthold mentioned just tries to create a little file in this filesystem. If this also doesn't freeze your system, try the other init scripts: cd /etc/init.d/boot.d/ for file in S* ; do echo "--- Starting $file... ---" ./$file start echo "done - press enter" read done Just wait some seconds after each script, then press enter to start the next one. (You could also type each script name manually, but the above script is easier ;-) Regards, Christian Boltz --
Bitkollisionen finden v.a. in Kneipen und Festzelten statt, würde ich mal annehmen. [H. Bengen] Und zu viele davon führen zu einem stomach overflow? [L. Barth]
Good morning, again...
had a shell prompt again - at least it didn't freeze...
That means that your problem comes from one of the init scripts you didn't test ;-)
Oh, thank you, I didn't know what to test...
shm is a tmpfs (a filesystem living in RAM, not hard disk) that is mounted to /dev/shm. The echo-command Berthold mentioned just tries to create a little file in this filesystem.
I see! ok, tried this "echo 'foo...", nothing happened (got next prompt), so this seems to be ok.
If this also doesn't freeze your system, try the other init scripts:
cd /etc/init.d/boot.d/ for file in S* ; do echo "--- Starting $file... ---" ./$file start echo "done - press enter" read done
Just wait some seconds after each script, then press enter to start the next one. (You could also type each script name manually, but the above script is easier ;-)
Thank you, Christian for detailed descrition that even learners like I can understand ;-) I very much appreciate this. So, here's the result: S01boot.proc S02boot.shm S03boot.rootfsck S04boot.udev seemed to run like they should, some messages, but not looking like errors, then S05boot.colplug freezed "Tons" of messages run over the screen. I just wrote down the last few that I could see: hw-random: RNG not detected ACPI-0768: ***warning: Thread 865 could not acquire Mutes [NULL] AE_BAD_PARAMETER shpchp: acpi_shpchprm: get_device PCI ROOT HID fail=0x1001 phci1394: $Rev: 1299 $ Ben Collins <(email-addr of bcollins)> PCI: Found IRQ 9 for device 0000:02:03.0 PCI: Sharing IRQ 9 for device 0000:02:0a.0 phci1394: fw_host0: OHCI-1394 1.0 (PCI) IRQ=[9] MMIO=[ed000000 - ed0007ff] Max-Packet=[2048] tg3.c: v3.37 (August 25, 2005) PCI: found IRQ 10 for device 0000:02:05.0 Then it freezes. ------------------------------ The "tg3.c"-line makes me remember another experiment I've tried earlier: I tried to update with Yast (staring it from the boot CD). After collecting the software list Yast said, it found some hardware and gave the message: load device driver TG3 Then it freezed. ----------------------------- Ok, I hope these "tests" will bring me closer to a solution. What should I do next? Thanks for any help!
On Tue, 20 Dec 2005, Daniel Bauer wrote:
"Tons" of messages run over the screen. I just wrote down the last few that I could see:
hw-random: RNG not detected ACPI-0768: ***warning: Thread 865 could not acquire Mutes [NULL] AE_BAD_PARAMETER shpchp: acpi_shpchprm: get_device PCI ROOT HID fail=0x1001 phci1394: $Rev: 1299 $ Ben Collins <(email-addr of bcollins)> PCI: Found IRQ 9 for device 0000:02:03.0 PCI: Sharing IRQ 9 for device 0000:02:0a.0 phci1394: fw_host0: OHCI-1394 1.0 (PCI) IRQ=[9] MMIO=[ed000000 - ed0007ff] Max-Packet=[2048] tg3.c: v3.37 (August 25, 2005) PCI: found IRQ 10 for device 0000:02:05.0
Then it freezes. ------------------------------ The "tg3.c"-line makes me remember another experiment I've tried earlier: I tried to update with Yast (staring it from the boot CD). After collecting the software list Yast said, it found some hardware and gave the message:
load device driver TG3
Then it freezed. ----------------------------- Ok, I hope these "tests" will bring me closer to a solution. What should I do next? Thanks for any help!
The tg3 driver is a driver for a broadcom network card. Possibly one that is onboard. To continue for now, you may disable that hardware in your bios, or, if it is an extra PCI card, just remove the card for now. There is also a parameter tg3_debug, but I don't know the values for this parameter right now. It would not solve your problem anyway... Berthold -- ------------------------------------------------------------------ Berthold Gunreben SUSE Linux GmbH -- Dokumentation mailto:bg@suse.de Maxfeldstr. 5 http://www.suse.de/ D-90409 Nuernberg, Germany
The tg3 driver is a driver for a broadcom network card. Possibly one that is onboard. To continue for now, you may disable that hardware in your bios, or, if it is an extra PCI card, just remove the card for now.
There is also a parameter tg3_debug, but I don't know the values for this parameter right now. It would not solve your problem anyway...
Berthold
Wow, thank you! That was it. In fact there is a broadcom network card on the motherboard (plus an extra davicom pci card). I disabled the onboard one in the bios, and booted - Yast took over for the second part of install - and it runs! Yippie! So now I'm going to install the X-things, hope it works :-) Again: thanks a lot for fast and effective help! Daniel
participants (4)
-
Berthold Gunreben -
Carl-Daniel Hailfinger -
Christian Boltz -
Daniel Bauer