1. Forum
  2. Usenet
  3. LINUX.DEBIAN.PORTS.POWERP

  • Updated installer images 2021-04-16

    From John Paul Adrian Glaubitz@21:1/5 to All on Fri Apr 16 11:50:02 2021
    Hello!

    As promised, new images:

    https://cdimage.debian.org/cdimage/ports/snapshots/2021-04-16/

    ppc64 should install fine on PowerMac G5 now.

    Adrian

    --
    .''`. John Paul Adrian Glaubitz
    : :' : Debian Developer - glaubitz@debian.org
    `. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
    `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Johannes Brakensiek@21:1/5 to David VANTYGHEM on Fri Apr 16 14:10:02 2021
    Hey David,

    On 16 Apr 2021, at 13:56, David VANTYGHEM wrote:

    In english : The boot partition isn't the first primary partition on
    the disk. This is necessary to boot the machine. Please restart and
    use the first primary partition as boot partition.

    If you don't come back to the partition menu to modify this default,
    the partition will be used as this. It means that you will probably
    not be able to start from the hard disk.

    Coming to the menu and modify this default ?

    please read the mails Adrian is writing. That’s not an error message.
    You are supposed to say „Non“ at that point. It is going to work nontheless.

    Johannes

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From David VANTYGHEM@21:1/5 to All on Fri Apr 16 14:10:02 2021
    This is a multi-part message in MIME format.
    Le 16/04/2021 à 14:00, Johannes Brakensiek a écrit :
    Hey David,

    On 16 Apr 2021, at 13:56, David VANTYGHEM wrote:

    In english : The boot partition isn't the first primary partition on
    the disk. This is necessary to boot the machine. Please restart and
    use the first primary partition as boot partition.

    If you don't come back to the partition menu to modify this default,
    the partition will be used as this. It means that you will probably
    not be able to start from the hard disk.

    Coming to the menu and modify this default ?

    please read the mails Adrian is writing. That’s not an error message.
    You are supposed to say „Non“ at that point. It is going to work nontheless.

    Johannes

    Aaahh, OK. It's at this place. I thought it was another problem because
    the error message is different in this new ISO file.

    --
    Passez à Linux : https://infolib.re

    .--.
    |o_o |
    ||_/ |
    // \\ Envoyé depuis mon Linux
    (| |)
    / \_ _/ \
    \___)=(___/


    <html>
    <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
    </head>
    <body>
    <p><br>
    </p>
    <div class="moz-cite-prefix">Le 16/04/2021 à 14:00, Johannes
    Brakensiek a écrit :<br>
    </div>
    <blockquote type="cite"
    cite="mid:B800AD03-94C9-4032-8F50-59550489B805@codingpastor.de">Hey
    David,
    <br>
    <br>
    On 16 Apr 2021, at 13:56, David VANTYGHEM wrote:
    <br>
    <br>
    <blockquote type="cite">In english : The boot partition isn't the
    first primary partition on the disk. This is necessary to boot
    the machine. Please restart and use the first primary partition
    as boot partition.
    <br>
    <br>
    If you don't come back to the partition menu to modify this
    default, the partition will be used as this. It means that you
    will probably not be able to start from the hard disk.
    <br>
    <br>
    Coming to the menu and modify this default ?
    <br>
    </blockquote>
    <br>
    please read the mails Adrian is writing. That’s not an error
    message. You are supposed to say „Non“ at that point. It is going
    to work nontheless.
    <br>
    <br>
    Johannes
    <br>
    <br>
    </blockquote>
    <p>Aaahh, OK. It's at this place. I thought it was another problem
    because the error message is different in this new ISO file.<br>
    </p>
    <pre class="moz-signature" cols="72">--
    Passez à Linux : <a class="moz-txt-link-freetext" href="https://infolib.re">https://infolib.re</a>

    .--.
    |o_o |
    ||_/ |
    // \\ Envoyé depuis mon Linux
    (| |)
    / \_ _/ \
    \___)=(___/</pre>
    <div id="grammalecte_menu_main_button_shadow_host" style="width:
    0px; height: 0px;"></div>
    </body>
    </html>

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Dennis Clarke@21:1/5 to John Paul Adrian Glaubitz on Fri Apr 16 14:10:02 2021
    On 4/16/21 5:40 AM, John Paul Adrian Glaubitz wrote:
    Hello!

    As promised, new images:

    https://cdimage.debian.org/cdimage/ports/snapshots/2021-04-16/

    ppc64 should install fine on PowerMac G5 now.

    Adrian


    I will jump on that right now !


    --
    Dennis Clarke
    RISC-V/SPARC/PPC/ARM/CISC
    UNIX and Linux spoken
    GreyBeard and suspenders optional

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From John Paul Adrian Glaubitz@21:1/5 to David VANTYGHEM on Fri Apr 16 14:20:01 2021
    On 4/16/21 1:56 PM, David VANTYGHEM wrote:
    I tried a new installation with 14/04 ISO file, default installation,
    whole disk and only one partition (no /home partition). I've got this
    error message at the end :

    Well, this is an error with setting the contents of NVRAM. Your emulated machine does not provide an accurate emulation of the NVRAM present in PowerMacs which is why the command fails.

    Not sure at the moment how to fix this on Debian's side.

    Adrian

    --
    .''`. John Paul Adrian Glaubitz
    : :' : Debian Developer - glaubitz@debian.org
    `. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
    `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Dennis Clarke@21:1/5 to John Paul Adrian Glaubitz on Fri Apr 16 15:00:01 2021
    On 4/16/21 5:40 AM, John Paul Adrian Glaubitz wrote:
    Hello!

    As promised, new images:

    https://cdimage.debian.org/cdimage/ports/snapshots/2021-04-16/

    ppc64 should install fine on PowerMac G5 now.

    Adrian


    In partman log I see :


    parted_server: OUT: 1 512-32767 32256 primary unknown
    /dev/sda1 Apple


    parted_server: OUT: 2 32768-256033279 256000512 primary hfs
    /dev/sda2 untitled


    parted_server: OUT: 3 256033280-299045033471 298789000192 primary
    ext4 /dev/sda3 untitled


    parted_server: OUT: 4 299045033472-300069027327 1023993856
    primary linux-swap /dev/sda4 swap


    parted_server: OUT: -1 300069027328-300069052415 25088 primary
    free /dev/sda-1


    parted_server: Partitions printed
    parted_server: OUT:


    This is from the automated use the whole disk partition option.

    In the installer I see :

    Guided partitioning
    Configure the Logical Volume Manager
    Configure encrypted volumes
    Configure iSCSI volumes

    SCSI2 (0,0,0) (sda) - 300.1 GB ATA ST3300622AS
    > #1 32.3 kB Apple
    > #2 256.0 MB f hfs untitled /boot/grub
    > #3 298.8 GB f ext4 untitled /
    > #4 1.0 GB f swap swap swap
    > 25.1 kB FREE SPACE
    SCSI3 (0,0,0) (sdb) - 160.0 GB ATA WDC WD1600JD-41H

    Undo changes to partitions
    Finish partitioning and write changes to disk

    Then if I accept this :

    Your boot partition is not located on the first primary partition of
    your hard disk. This is needed by your machine in order to boot.
    Please go back and use your first primary partition as a boot
    partition.

    etc etc

    In the syslog I see :

    Apr 16 12:42:25 check-missing-firmware: /run/udev/firmware-missing does
    not exist, skipping
    Apr 16 12:42:25 check-missing-firmware: no missing firmware in loaded
    kernel modules
    Apr 16 12:42:26 check-missing-firmware: looking at dmesg again,
    restarting from \[ 651.449729\]
    Apr 16 12:42:26 check-missing-firmware: timestamp found, truncating
    dmesg accordingly
    Apr 16 12:42:26 check-missing-firmware: saving timestamp for a later use:
    Apr 16 12:42:26 check-missing-firmware: /dev/.udev/firmware-missing does
    not exist, skipping
    Apr 16 12:42:26 check-missing-firmware: /run/udev/firmware-missing does
    not exist, skipping
    Apr 16 12:42:26 check-missing-firmware: no missing firmware in loaded
    kernel modules
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 3 (pipe:[7634]) leaked on pvs invocation. Parent PID 16815: /bin/sh
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 4 (/dev/pts/6) leaked on pvs invocation. Parent PID 16815: /bin/sh
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 5 (/dev/pts/6) leaked on pvs invocation. Parent PID 16815: /bin/sh
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 6 (/dev/pts/6) leaked on pvs invocation. Parent PID 16815: /bin/sh
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 3 (pipe:[7634]) leaked on pvs invocation. Parent PID 16962: /bin/sh
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 4 (/dev/pts/6) leaked on pvs invocation. Parent PID 16962: /bin/sh
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 5 (/dev/pts/6) leaked on pvs invocation. Parent PID 16962: /bin/sh
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 6 (/dev/pts/6) leaked on pvs invocation. Parent PID 16962: /bin/sh
    Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 3 (pipe:[7634]) leaked on pvs invocation.
    Apr 16 12:42:26 main-menu[3674]: (process:16811): Parent PID 16992: /bin/sh Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 4 (/dev/pts/6) leaked on pvs invocation.
    Apr 16 12:42:26 main-menu[3674]: (process:16811): Parent PID 16992: /bin/sh Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 5 (/dev/pts/6) leaked on pvs invocation.
    Apr 16 12:42:26 main-menu[3674]: (process:16811): Parent PID 16992: /bin/sh Apr 16 12:42:26 main-menu[3674]: (process:16811): File descriptor 6 (/dev/pts/6) leaked on pvs invocation.
    Apr 16 12:42:26 main-menu[3674]: (process:16811): Parent PID 16992: /bin/sh Apr 16 12:42:27 main-menu[3674]: INFO: Falling back to the package
    description for brltty-udeb
    Apr 16 12:42:29 main-menu[3674]: INFO: Falling back to the package
    description for brltty-udeb
    Apr 16 12:42:29 main-menu[3674]: INFO: Menu item 'partman-base' selected

    I may power the machine off and try again to see why this is happening.


    --
    Dennis Clarke
    RISC-V/SPARC/PPC/ARM/CISC
    UNIX and Linux spoken
    GreyBeard and suspenders optional

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From John Paul Adrian Glaubitz@21:1/5 to Dennis Clarke on Fri Apr 16 15:10:02 2021
    On 4/16/21 2:55 PM, Dennis Clarke wrote:
    Then if I accept this :

    Your boot partition is not located on the first primary partition of
    your hard disk. This is needed by your machine in order to boot.
    Please go back and use your first primary partition as a boot
    partition.

    etc etc

    As I said, there is a bogus error message where you just have to say "<NO>"
    to continue. I will fix this error later.

    Adrian

    --
    .''`. John Paul Adrian Glaubitz
    : :' : Debian Developer - glaubitz@debian.org
    `. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
    `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Dennis Clarke@21:1/5 to John Paul Adrian Glaubitz on Fri Apr 16 15:40:01 2021
    On 4/16/21 9:01 AM, John Paul Adrian Glaubitz wrote:
    On 4/16/21 2:55 PM, Dennis Clarke wrote:
    Then if I accept this :

    Your boot partition is not located on the first primary partition of
    your hard disk. This is needed by your machine in order to boot.
    Please go back and use your first primary partition as a boot
    partition.

    etc etc

    As I said, there is a bogus error message where you just have to say "<NO>" to continue. I will fix this error later.


    AH HA!

    Well I instantly re-did the install with the default options and merely
    went past that message and ALL IS WELL!



    admsys@enceladus:~$
    admsys@enceladus:~$ uname -a
    Linux enceladus 5.10.0-6-powerpc64 #1 SMP Debian 5.10.28-1 (2021-04-09)
    ppc64 GNU/Linux
    admsys@enceladus:~$
    admsys@enceladus:~$ file /usr/bin/ls
    /usr/bin/ls: ELF 64-bit MSB pie executable, 64-bit PowerPC or cisco
    7500, version 1 (SYSV), dynamically linked, interpreter
    /lib64/ld64.so.1,
    BuildID[sha1]=4fbf6330cc04ddc123654ee5654035d8f97e0016, for GNU/Linux
    3.2.0, stripped
    admsys@enceladus:~$
    admsys@enceladus:~$ ldd /usr/bin/ls
    linux-vdso64.so.1 (0x00007fff82640000)
    libselinux.so.1 => /lib/powerpc64-linux-gnu/libselinux.so.1 (0x00007fff82570000)
    libc.so.6 => /lib/powerpc64-linux-gnu/libc.so.6 (0x00007fff82340000)
    libpcre2-8.so.0 => /lib/powerpc64-linux-gnu/libpcre2-8.so.0 (0x00007fff82280000)
    libdl.so.2 => /lib/powerpc64-linux-gnu/libdl.so.2
    (0x00007fff82250000)
    /lib64/ld64.so.1 (0x00007fff82660000)
    libpthread.so.0 => /lib/powerpc64-linux-gnu/libpthread.so.0 (0x00007fff82200000)
    admsys@enceladus:~$
    admsys@enceladus:~$ uptime
    09:34:03 up 6 min, 1 user, load average: 0.05, 0.30, 0.20 admsys@enceladus:~$

    I have not seen this running neatly in well over six months!

    Excellent work !


    --
    Dennis Clarke
    RISC-V/SPARC/PPC/ARM/CISC
    UNIX and Linux spoken
    GreyBeard and suspenders optional

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rick Thomas@21:1/5 to John Paul Adrian Glaubitz on Mon Apr 19 10:00:01 2021
    I tried it this afternoon on my PowerMac G5 - PowerMac7,3. Works great! I used the default installation and "all in one partition" partitioning. Let me know if you would like me to try something else.

    Thanks so much for all your work! This is fantastic...

    Enjoy!
    Rick

    On Fri, Apr 16, 2021, at 2:40 AM, John Paul Adrian Glaubitz wrote:
    Hello!

    As promised, new images:

    https://cdimage.debian.org/cdimage/ports/snapshots/2021-04-16/

    ppc64 should install fine on PowerMac G5 now.

    Adrian

    --
    .''`. John Paul Adrian Glaubitz
    : :' : Debian Developer - glaubitz@debian.org
    `. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
    `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913



    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rick Thomas@21:1/5 to Rick Thomas on Tue Apr 20 03:50:02 2021
    Well... I'm not very good at letting well enough alone, so I tried installing again. This time I used the "default install" and when it came to partitioning, I chose the LVM option. This time, it complained that there was no partition for /boot (
    though when I poked around a bit, there _was_ an hfs partition for /boot/grub .) It asked if I wanted to continue without /boot and I answered "yes". It continued and installed grub using the /boot/grub partition and the /boot directory in the root LVM
    partition.

    Then it rebooted. I was fully expecting it to get hung up unable to find /boot partition and so be unable to get the initrd and kernel, but (lo and behold!) the new grub seems to be able to dig inside of an LVM and find what it needs! ((-: It
    continued with an error message (twice) about "can't open device" (I don't know which device it's talking about) and paused with "Press any key to continue". While I was puzzling what to do, the pause timed out and it continued without my hitting any
    key. And to my great surprise and glee, it booted and gave me a "login:" prompt!

    So...
    (0) Congratulations!
    (1) Can we get rid of the seemingly unnecessary error message about not having a /boot partition?
    (2) Are the messages necessary that complain about "can't open device" and "press any key to continue"?
    (3) If not, can we get rid of them?

    Big *thanks* for this!
    Rick

    On Mon, Apr 19, 2021, at 12:35 AM, Rick Thomas wrote:
    I tried it this afternoon on my PowerMac G5 - PowerMac7,3. Works
    great! I used the default installation and "all in one partition" partitioning. Let me know if you would like me to try something else.

    Thanks so much for all your work! This is fantastic...

    Enjoy!
    Rick

    On Fri, Apr 16, 2021, at 2:40 AM, John Paul Adrian Glaubitz wrote:
    Hello!

    As promised, new images:

    https://cdimage.debian.org/cdimage/ports/snapshots/2021-04-16/

    ppc64 should install fine on PowerMac G5 now.

    Adrian

    --
    .''`. John Paul Adrian Glaubitz
    : :' : Debian Developer - glaubitz@debian.org
    `. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
    `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913





    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From John Paul Adrian Glaubitz@21:1/5 to Rick Thomas on Tue Apr 20 08:40:01 2021
    On 4/20/21 3:42 AM, Rick Thomas wrote:
    (1) Can we get rid of the seemingly unnecessary error message about not having a /boot partition?

    As I said, it's a known issue and I'm going to fix it at some point. But
    since I am doing this in my freetime and I'm not getting paid for that,
    I'm not working on it every day.

    Fixing installer issues is not trivial and requires lots of testing, getting
    it working last week took me four days and 10-15 installation runs, so I
    cannot work on it every day.

    (2) Are the messages necessary that complain about "can't open device" and "press any key to continue"?

    I don't know the context of these. You need to provide more context, use a
    net console if you cannot access the logs otherwise.

    Adrian

    --
    .''`. John Paul Adrian Glaubitz
    : :' : Debian Developer - glaubitz@debian.org
    `. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
    `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From John Paul Adrian Glaubitz@21:1/5 to Rick Thomas on Tue Apr 20 11:40:01 2021
    On 4/20/21 11:16 AM, Rick Thomas wrote:
    Presumably, this message, and the check that issues it, are relics of the past when
    grub was not able to dig into LVM partitions to find the kernel and initrd. Now that
    grub understands LVM, the check and associated message are no longer necessary.

    As I said, those are known issues and I know how these messages are triggered. They are
    part of the checks in the partman-* packages. I simply didn't have the time to work on
    that yet as changing these things involves lots of testing and trial and error to make
    sure the checks are correct.

    I don't know the context of these. You need to provide more context, use a >> net console if you cannot access the logs otherwise.

    The context of these messages is immediately after the "bong" when the machine reboots
    itself -- after grub has loaded and displayed its menu but before the kernel and initrd
    are loaded. I'm guessing that they come from grub and are byproducts of the process of
    grub digging around in the LVM looking for the /boot directory.

    I'm not even there yet to work on LVM. I don't understand the hurry to make everything
    perfect at once. This issue might not even specific to PowerMacs, so it would make more
    sense first to test this on x86 systems and report a bug against src:grub2 or partman-lvm
    if the issue shows there as well.

    Adrian

    --
    .''`. John Paul Adrian Glaubitz
    : :' : Debian Developer - glaubitz@debian.org
    `. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
    `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rick Thomas@21:1/5 to John Paul Adrian Glaubitz on Tue Apr 20 11:20:02 2021
    On Mon, Apr 19, 2021, at 11:33 PM, John Paul Adrian Glaubitz wrote:
    On 4/20/21 3:42 AM, Rick Thomas wrote:
    (1) Can we get rid of the seemingly unnecessary error message about not having a /boot partition?

    As I said, it's a known issue and I'm going to fix it at some point. But since I am doing this in my freetime and I'm not getting paid for that,
    I'm not working on it every day.

    Fixing installer issues is not trivial and requires lots of testing, getting it working last week took me four days and 10-15 installation runs, so I cannot work on it every day.

    Understood. I'm grateful for the work you can afford to do, and I don't intend to pressure you. I'm just trying to do my part -- sorry, I'm not a developer, so I can't help with the code -- but I will do what I can, which is to provide my observations
    as I test your code. In particular, the error message mentioned above comes from the partitioner section of the installer. Presumably, this message, and the check that issues it, are relics of the past when grub was not able to dig into LVM partitions
    to find the kernel and initrd. Now that grub understands LVM, the check and associated message are no longer necessary.

    (2) Are the messages necessary that complain about "can't open device" and "press any key to continue"?

    I don't know the context of these. You need to provide more context, use a net console if you cannot access the logs otherwise.

    The context of these messages is immediately after the "bong" when the machine reboots itself -- after grub has loaded and displayed its menu but before the kernel and initrd are loaded. I'm guessing that they come from grub and are byproducts of the
    process of grub digging around in the LVM looking for the /boot directory. If that explanation doesn't clear up the ambiguity, I'll get out my camera and take some pictures of the screen for you.

    Is there some way to tell grub to log its progress to some durable media like a USB stick?

    Adrian

    Rick

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rick Thomas@21:1/5 to John Paul Adrian Glaubitz on Tue Apr 20 22:20:02 2021
    On Tue, Apr 20, 2021, at 2:31 AM, John Paul Adrian Glaubitz wrote:
    I don't understand the hurry to
    make everything
    perfect at once.

    I'm sorry if it seems that way. That is not my intent. I'm just reporting what I see.
    It's definitely not mission critical. The machine still boots, just with these puzzling messages.
    You can ignore it if you want.

    This issue might not even specific to PowerMacs, so it
    would make more
    sense first to test this on x86 systems and report a bug against
    src:grub2 or partman-lvm
    if the issue shows there as well.

    I'll try that. Would the "daily build" sid iso for amd64 have the same version of grub2 as your ppc64 iso?

    Thanks for all your efforts!
    Rick
    Adrian

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From John Paul Adrian Glaubitz@21:1/5 to Rick Thomas on Wed Apr 21 13:30:02 2021
    On 4/20/21 10:13 PM, Rick Thomas wrote:
    On Tue, Apr 20, 2021, at 2:31 AM, John Paul Adrian Glaubitz wrote:
    I don't understand the hurry to
    make everything
    perfect at once.

    I'm sorry if it seems that way. That is not my intent. I'm just reporting what I see.
    It's definitely not mission critical. The machine still boots, just with these puzzling messages.
    You can ignore it if you want.

    I do appreciate the feedback. But a more complete feedback with some pre-testing would be nice,
    anything that helps me spend less time and effort to get the issues sorted out.

    So, in general if you see a problem, try different configuration settings, check the logs yourself
    for suspicious messages and maybe try to reproduce the error on different architectures. Any test
    that may help pinpoint the problem is useful because it means I will have to spend less time doing
    that myself and can fix the issues more efficiently.

    You can also grep the debian-installer, GRUB and related sources on Salsa and sources.debian.org
    to see where the error message comes from. That information is very helpful, too, and it also
    helps you becoming more familiar with the code.

    This issue might not even specific to PowerMacs, so it
    would make more
    sense first to test this on x86 systems and report a bug against
    src:grub2 or partman-lvm
    if the issue shows there as well.

    I'll try that. Would the "daily build" sid iso for amd64 have the same version of grub2 as your ppc64 iso?

    This should do the trick, yes:

    https://cdimage.debian.org/cdimage/daily-builds/daily/current/amd64/iso-cd/

    Adrian

    --
    .''`. John Paul Adrian Glaubitz
    : :' : Debian Developer - glaubitz@debian.org
    `. `' Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
    `- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rick Thomas@21:1/5 to John Paul Adrian Glaubitz on Fri Apr 23 04:20:01 2021
    On Wed, Apr 21, 2021, at 4:26 AM, John Paul Adrian Glaubitz wrote:
    This issue might not even specific to PowerMacs, so it
    would make more
    sense first to test this on x86 systems and report a bug against
    src:grub2 or partman-lvm
    if the issue shows there as well.

    I'll try that. Would the "daily build" sid iso for amd64 have the same version of grub2 as your ppc64 iso?

    This should do the trick, yes:

    https://cdimage.debian.org/cdimage/daily-builds/daily/current/amd64/iso-cd/

    I just tried that on an amd64 VM. I chose the default install from the installer's menu. When it got to partitioning the disk, I chose use whole disk and set up LVM. I chose separate /home logical volume. All went well, and when it came to reboot
    from the newly installed disk, I did not see any of the messages that I saw on the ppc64 PowerMac G5 install. However:

    The amd64 install created a separate /boot partition on /dev/sda1. In addition to kernel and initrd, it contained a /boot/grub directory.

    rbthomas@burnout:~$ df | grep -v tmpfs ; lsblk

    Filesystem 1K-blocks Used Available Use% Mounted on
    udev 1995140 0 1995140 0% /dev
    /dev/mapper/burnout--vg-root 6930164 962864 5593952 15% /
    /dev/sda1 480618 48562 407122 11% /boot
    /dev/mapper/burnout--vg-home 12028924 40 11396052 1% /home

    NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
    sda 8:0 0 20G 0 disk
    |-sda1 8:1 0 487M 0 part /boot
    |-sda2 8:2 0 1K 0 part
    `-sda5 8:5 0 19.5G 0 part
    |-burnout--vg-root 254:0 0 6.8G 0 lvm /
    |-burnout--vg-swap_1 254:1 0 976M 0 lvm [SWAP]
    `-burnout--vg-home 254:2 0 11.8G 0 lvm /home
    sr0 11:0 1 1024M 0 rom

    rbthomas@burnout:~$

    The ppc64 install, on the other hand, created a separate hfs type /boot/grub partition on /dev/sda2, and let the parent /boot be a directory in the ext4 root partition/logical-volume

    rbthomas@kmac:~$ df | grep -v tmpfs ; lsblk

    Filesystem 1K-blocks Used Available Use% Mounted on
    udev 1874880 0 1874880 0% /dev
    /dev/mapper/kmac--vg-root 19047080 1372148 16682064 8% /
    /dev/sda2 249988 11228 238760 5% /boot/grub
    /dev/mapper/kmac--vg-home 158766164 60 150628376 1% /home

    NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
    sda 8:0 0 232.9G 0 disk
    |-sda1 8:1 0 31.5K 0 part
    |-sda2 8:2 0 244.1M 0 part /boot/grub
    |-sda3 8:3 0 232.6G 0 part
    | |-kmac--vg-root 254:1 0 18.6G 0 lvm /
    | |-kmac--vg-swap_1 254:2 0 10G 0 lvm [SWAP]
    | `-kmac--vg-home 254:3 0 154.9G 0 lvm /home
    `-sda4 8:4 0 56.5K 0 part
    sdb 8:16 0 931.5G 0 disk
    `-backup-lvol0 254:0 0 500G 0 lvm
    sr0 11:0 1 1024M 0 rom

    rbthomas@kmac:~$

    I suspect that this difference is what's producing the "error: can't open device" messages on the ppc64, and why they do not appear on the amd64. If I did a manual partitioning install on the amd64 that mimicked the layout of the ppc64, I suspect I
    would probably see the same "can't open device" error messages there as well. (I do plan to try that, and I'll report what I find)

    Rick

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rick Thomas@21:1/5 to Rick Thomas on Sat Apr 24 04:50:01 2021
    On Thu, Apr 22, 2021, at 7:16 PM, Rick Thomas wrote:
    I suspect that this difference is what's producing the "error: can't
    open device" messages on the ppc64, and why they do not appear on the
    amd64. If I did a manual partitioning install on the amd64 that
    mimicked the layout of the ppc64, I suspect I would probably see the
    same "can't open device" error messages there as well. (I do plan to
    try that, and I'll report what I find)

    Well... "Curiouser and curiouser!"

    I did an expert install on my amd64 VM and accepted defaults up to the partitioning phase. In the partitioning phase, I chose to do a manual partitioning. In my manual partitioning I created an ext2 partition on /dev/sda1 with mount point /boot/grub.
    The rest of the installation went into an LVM partition with three volumes for root, /home, and swap. (So /boot is a directory in the root volume. If grub wants a kernel and initrd, it has to dig around in the LVM to find them)

    Unlike the ppc64 installation, the installer did not complain about this setup. (The ppc64 installation complained that there was no partition for /boot to mount on) and happily went ahead with installing a basic (non-gui) system with ssh. Somewhat
    surprisingly, when it came time for the reboot, all went smoothly. There were no "error: can't open device" messages at all, and it proceeded smoothly to a login prompt.

    Hope this helps!

    Rick

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)