Your message dated Sun, 1 Aug 2010 12:33:01 -0400
with message-id <[email protected]>
and subject line Re: Bug#591244: nbd-client: specifying nbd device with nbdroot
not very useful
has caused the Debian Bug report #591244,
regarding nbd-client: specifying nbd device with nbdroot not very useful
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [email protected]
immediately.)
--
591244: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=591244
Debian Bug Tracking System
Contact [email protected] with problems
--- Begin Message ---
Package: nbd-client
Version: 1:2.9.16-2
Severity: minor
nbdroot is working rather nicely now. thanks! :)
however, specifying the nbd block device via the nbdroot= parameter seems less
than useful, as it also needs to be specified via the root= boot parameter,
otherwise the initramfs-tools scripts/local doesn't know what root device to
mount... simply specifying:
nbdroot=192.168.0.1,2000,nbd0
nbd-client configures the nbd device properly, but fails to mount root.
since the scripts/local-top/nbd script handles getting the root device from the
root= parameter, having it also specified in nbdroot= can actually lead to
inconsistant behavior between nbdroot/root that leads to a boot failure:
root=/dev/nbd0 nbdroot=192.168.0.1,2000,nbd1
causes nbd1 to get configured via nbd-client, but then fails to mount root as
it tries to use nbd0.
if there was some way for scripts/local-top/nbd to change the root= parameter
in a way that the scripts/local respects, it might be useful to keep the code
for setting the nbd device with the nbdroot parameter. if that's not feasible,
and there isn't some use i can't think of at the moment, probably best to drop
the code handling it.
live well,
vagrant
--- End Message ---
--- Begin Message ---
On Sun, Aug 01, 2010 at 07:06:27AM -0700, Vagrant Cascadian wrote:
> nbdroot is working rather nicely now. thanks! :)
>
> however, specifying the nbd block device via the nbdroot= parameter seems less
> than useful, as it also needs to be specified via the root= boot parameter,
> otherwise the initramfs-tools scripts/local doesn't know what root device to
> mount... simply specifying:
>
> nbdroot=192.168.0.1,2000,nbd0
>
> nbd-client configures the nbd device properly, but fails to mount root.
>
> since the scripts/local-top/nbd script handles getting the root device
> from the root= parameter, having it also specified in nbdroot= can
> actually lead to inconsistant behavior between nbdroot/root that leads
> to a boot failure:
>
> root=/dev/nbd0 nbdroot=192.168.0.1,2000,nbd1
>
> causes nbd1 to get configured via nbd-client, but then fails to mount
> root as it tries to use nbd0.
Don't do that, then :-)
I might be persuaded to add some code which checks whether the root=
device looks like an NBD device, and if so, whether it would conflict
with the device as specified in the nbdroot= parameter. However, that
could be problematic, and I'd rather not deal with that if it's not
absolutely necessary.
> if there was some way for scripts/local-top/nbd to change the root=
> parameter in a way that the scripts/local respects, it might be useful
> to keep the code for setting the nbd device with the nbdroot
> parameter. if that's not feasible,
It isn't. The regular scripts that mount the root filesystem parse the
kernel command line, and you can't change the kernel command line once
the kernel's booted (for obvious reasons).
However, specifying the nbd device in nbdroot is optional; it is only
required if the device to use would be ambiguous otherwise.
> and there isn't some use i can't think of at the moment, probably
> best to drop the code handling it.
There is another use case: if you wish to use root-on-LVM-on-NBD, or
root-on-dmcrypt-on-NBD, or something else equally insane, then you
cannot specify the NBD device through the root= parameter. That is the
only reason why the nbd-client init.d script supports specifying it
anywhere else than the root= parameter; but in the common case, you
should really only be specifying it in root=.
Hey, I won't be doing something layerd on NBD myself, but it's not an
invalid configuration per se, and if you're interested in doing
something crazy, who am I to stop you?
It might be a good idea to document this better, however; patches to the
documentation are certainly welcome.
Thanks,
--
The biometric identification system at the gates of the CIA headquarters
works because there's a guard with a large gun making sure no one is
trying to fool the system.
http://www.schneier.com/blog/archives/2009/01/biometrics.html
--- End Message ---
