Find the answer to your Linux question:
Results 1 to 5 of 5
This is a rather interesting problem I have ran into.... I have 18 Shuttle-X SS56G computers running a Morphix (Debian derivative) LiveCD package which works very well! Here is the ...
  1. #1
    Just Joined!
    Join Date
    Aug 2005
    Location
    Florida
    Posts
    3

    Interesting Reboot Problem... Anyone???

    This is a rather interesting problem I have ran into....

    I have 18 Shuttle-X SS56G computers running a Morphix (Debian derivative) LiveCD package which works very well!

    Here is the problem.... 2 of the computers will reboot properly when the "shutdown -r now" command is issued but the other 16 computers just "hang" at the reboot point (as far as I can determine).

    I have checked the BIO version of the 2 working machines vs the 16 non-rebooting machines and the BIO revision numbers are the same - as far as I can tell the hardware is the "same" but I am not sure there were not any actual chip-level changes to the design of the motherboards...

    What I am wanting to do is put together a very small program (even could be done in assembly language) that basically will cause the machine to start executing code at the computer's cold-boot entry point (FFFF:0000). I am not a C programmer nor an assembly language programmer by any stretch but if given some direction I think I can put together a program to perform this "reboot" operation....

    I am not worried about "losing" any data as these machines do not have a hard drive in them - everything runs from the CD drive (hince - liveCD name)

    Thoughts, suggestions and constructive comments are welcome!

    All negative comments will be directed the /dev/null

    TIA!!!

    gm...

  2. #2
    Linux Engineer LondoJowo's Avatar
    Join Date
    May 2005
    Location
    Richmond, TX
    Posts
    804
    All negative comments will be directed the /dev/null
    Very unlikely that you'll see any negative comments in here.

    Have you checked to make all have the same ACPI/APM settings in the BIOS?
    Dell Precision T7400 Workstation
    Dual 3.33Ghz Xeon "Harpertown" Core
    16GB PC5300 DDR2 ECC CL5
    BFG GeForce GTX 285 OC 2GB
    X-Fi Platinum
    HP w2408 24" Monitor
    Dual Boot:openSUSE 11.2/Win 7 Ultimate

  3. #3
    Just Joined!
    Join Date
    Aug 2005
    Location
    Florida
    Posts
    3
    Quote Originally Posted by LondoJowo
    All negative comments will be directed the /dev/null
    Very unlikely that you'll see any negative comments in here.

    Have you checked to make all have the same ACPI/APM settings in the BIOS?
    Yes - I have checked teh ACPI/APM settings comparing the working (reboot properly) against the machines that are not rebooting properly - they are set the same... I have even gone as far to generate a matrix of possible settings and gone through that exercise to see if there was a different combination of ACPI/APM settings that would work.... No Joy!

    Thanks for the suggestion!

    gm...

  4. #4
    Linux Engineer LondoJowo's Avatar
    Join Date
    May 2005
    Location
    Richmond, TX
    Posts
    804
    So you've compared all the BIOS settings? How about jumpers on the motherboards? I realize it's a pain in the butt as you've got 2 working and 16 not playing well.
    Dell Precision T7400 Workstation
    Dual 3.33Ghz Xeon "Harpertown" Core
    16GB PC5300 DDR2 ECC CL5
    BFG GeForce GTX 285 OC 2GB
    X-Fi Platinum
    HP w2408 24" Monitor
    Dual Boot:openSUSE 11.2/Win 7 Ultimate

  5. #5
    Just Joined!
    Join Date
    Aug 2005
    Location
    Florida
    Posts
    3
    Quote Originally Posted by LondoJowo
    So you've compared all the BIOS settings? How about jumpers on the motherboards? I realize it's a pain in the butt as you've got 2 working and 16 not playing well.
    Yep - thought of that as well - went so far as to remove a motherboard from one of the working units and one of the non-working units and did a visual comparison - all jumpers are set the same!

    I think I have exhausted most all possibles in terms of the bios and hardware configurations - hince the statement that I think there may have been a modification or modifications at the chip level that would normally not affect bios - but apparently affects rebooting in Debian...

    I have not delved into the inner-most recesses of the kernel code to determine "how" the warm reboot is performed but from what I remember it uses a bios call with "magic" number values to initiate a keyboard controller "reboot" command in much the same way the "control-alt-delete" key sequence is supposed to perform a reboot ... that is the "other" interesting thing - the keyboard "control-alt-delete" seems to work properly... I am beginning to think "something" in the firmware design of the keyboard controller may have been changed and the warm-boot magic number no longer is correct for this machine....

    I am still of the opinion the "best" way to "fix" this instance of a problem would be to put together a program that will cause the system to start code execution at the cold-boot entry point (0xFFFF0000) and have a cron job issue the command on a daily basis...

    The reason I need to reboot the system is to clean up the memory - since this is running as a LiveCD system it tends to "use up" all available memory over time and does not recover it - once all the ram memory has been used up the system hangs until power cycled - I want to automate a reboot automatically to clean up the memory so the hang condition does not occur.... I am remote to all but 2 units so having to make a 250+ mile a day drive really is not my idea of fun!

    gm...

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •