Loading ...
Sorry, an error occurred while loading the content.

Re: Debian/NSLU2 using NFS

Expand Messages
  • Phil Endecott
    ... In my case, I installed to a flash drive (sda1), removed the flash drive from the NSLU2 and plugged it in to the fileserver where it is mounted as
    Message 1 of 2 , Jul 5 5:13 AM
    • 0 Attachment
      kuehmi79 wrote:
      > I am currently using a manually bootstraped Debian (DebianSlug/LE)
      > mounting the rootfs over NFS. There are several upgrade procedures to
      > Debian/NSLU2 (etch) available (e.g. on
      > http://www.nslu2-linux.org/wiki/), but all assume the rootfs on a
      > external disk (usually /dev/sda1) and provide a special image for this
      > case. The etch installer on the other hand does not support an
      > installation to a NFS partition (not surprising). Does anyone has an
      > idea on
      >
      > (a) how to do an upgrade using the existing NFS partition (in this
      > case an appropriate firmware image is needed, not assuming rootfs on
      > /dev/sda1)

      In my case, I installed to a flash drive (sda1), removed the flash
      drive from the NSLU2 and plugged it in to the fileserver where it is
      mounted as /slugroot and NFS-exported.

      I had to make some minor changes e.g. editting /etc/fstab, but it was
      nothing complicated.

      > (b) how to change the apex configuration of a Debian/NSLU2
      > installation to boot from a NFS partition (I already have a slug with
      > an installation on a USB stick for playing).

      Well what do you really mean here? I don't think that anyone has got
      their Slug to actually boot (i.e. load its kernel) across the network,
      though there may be support for that in some future version of APEX
      (someone please correct me if I'm wrong). So the kernel is still in
      the Slug's flash. Then, in a default Debian install you will have a
      large initramfs also stored in the internal flash. This initramfs
      contains lots of modules and the Intel network microcode and I believe
      that it is sufficient to mount the eventual root partition over NFS;
      all that you need to do is to pass the right kernel command line
      parameters. There is Wiki documentation describing how to change the
      command line that APEX passes to the kernel.

      If you want to boot without the initramfs for some reason it all gets
      more complex, mostly because of the microcode. If you really want to
      do this, tell me why and I'll write some more.

      (One thing that I have not yet got to work is having multiple slugs use
      the same NFS-mounted read-only root filesystem. Has anyone done that?)


      Regards,

      Phil.
    Your message has been successfully submitted and would be delivered to recipients shortly.