Re: Creating /dev/md* device files

From: Mitchell Laks (mlaks_at_verizon.net)
Date: 01/27/05

  • Next message: Ralph Aichinger: "What to do when a bug report is not accepted?"
    To: debian-user@lists.debian.org
    Date: Wed, 26 Jan 2005 20:28:01 -0500
    
    

    On Wednesday 26 January 2005 01:59 pm, Steve Witt wrote:

    > Thanks for the help, mknod worked but MAKEDEV didn't (I was root). And
    > devices made with mknod don't survive a reboot!?! So, in digging a bit
    > more, it seems, unbeknownst by me, that this machine is using the udev
    > scheme for managing devices (it has a 2.6.8 kernel, which I should have
    > mentioned). So at least I know why MAKEDEV didn't work and now need to
    > research udev and find out how to get md devices to consistantly show up.
    I am running multiple software raid systems.
    I also had the trouble with udev. In the end I just removed udev, ran
    mknod /dev/md0 b 9 0 and went from there.

    There is no problem if you compile the modules md, raid5 (or raid1) and md
    into the kernel. However if you load these as modules, then there is a
    problem. Udev apparently doesnt create the devices early enough.

    The issue is that it is hard to be able to have udev create the devices early
    enough so that the raid array will exist early during the boot process
    so that when fsck is run during /etc/rcS.d
    (ie in single user mode - when we run S30checkfs.sh, and all the devices
    in /etc/fstab have fsck run on them) we need that /dev/md0 should exist.

    Anyway, this is apparently a known issue with udev. If you have a solution
    using initrd and modules other than recompiling the kernel, let me know. To
    find the bug report, look for my prior postings on debian-user and md0.

    I quote from Greg Folkert

    >
    > i am curious. i am running 2.6.8 kernel on sarge. i tried to avoid
    recompiling
    > my kernel and would prefer to load the raid5 driver as a module, and thus
    use
    > initrd.img. however, in my initial setup, after i ran mknod /dev/md0 and
    then
    > rebooted, /dev/md0 dissapeared, (perhaps this udev? is the culpret? who
    > knows?).

    Well you found the same bug I found doing it that way. Remove udev. file
    a bug against kernel-source-2.6.8-1. Ask the Debian Kernel Team. As far
    as the bug, here it is and read Marco's response and re-assign to the
    kernel.

    http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=273182

    > well what do i do to make it appear each boot? or perhaps, once i
    > create the initrd with raid5 module in it, then it will appear without my
    > doing mknod /dev/md0 each time, or do i need a script in /etc/rcS.d
    > or /etc/rc2.d to do
    > mknod /dev/md0 b 9 0
    > each reboot?
    > any ideas?
    > mknod /dev/md0

    Oh, yes my child... all you haav to do is remove udev for the time being
    or recompile a kernel with raid* compiled in statically.

    -- 
    greg, greg <at> gregfolkert.net
    Good Luck
    Mitchell Laks
    -- 
    To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org 
    with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
    

  • Next message: Ralph Aichinger: "What to do when a bug report is not accepted?"

    Relevant Pages

    • Re: /etc/udev/rules.d and permissions in /dev/usb
      ... The udevsynthesize/uevent approach actually found a bug in the kernel ... >> It was there from the start of udev. ... Well, it's not a problem for you, but I've got SCSI disks... ...
      (uk.comp.os.linux)
    • Re: /etc/udev/rules.d and permissions in /dev/usb
      ... > The udevsynthesize/uevent approach actually found a bug in the kernel ... >>> It was there from the start of udev. ... >> reboot a little too early. ... > I fitted a new disk last week and moved my old /usr over onto it with dd. ...
      (uk.comp.os.linux)
    • Re: Re: "Waiting for root file system..." hang solved
      ... bug, a kernel package bug, a udev bug, or operator error. ... you upgraded the kernel and the new kernel ... Just installed Etch but on first boot after installation I got the "Waiting for root file system..." ... kernel modules are loaded before udev daemon is started though...... ...
      (Debian-User)
    • Re: [patch 00/13] devtmpfs patches
      ... Initramfs logic works by just waiting for the device node udev creates ... We have many tings to make the kernel ... reason alone to kill it, was the dumb new naming scheme, which broke ... Sysfs doesn't even support per user ACLs which means its not much use for ...
      (Linux-Kernel)
    • Re: devfs vs udev, thoughts from a devfs user
      ... into sysfs, what about a little thought experiment. ... contest is basically the same thing for the kernel. ... devfs is just one more filesystem of many. ... But again, you're talking about udev and devfs, when you ...
      (Linux-Kernel)