klibc - Aug 2015 - ipconfig - single option for specifying hostname

Hi all,

I've recently been experimenting with Xen VMs booted using DHCP/NFS, and
that's looking good.

One nuisance is that I'm unable to get the DNS to update on the server
since there's no hostname in the DHCP request packet.

I'm aware that I can specify a hostname using the long-format device name,
but this doesn't seem to play well with initramfs. So, I'd like to add a
patch to ipconfig that allows a parameter to specify the hostname. Are
there any immediate objections to this, or reasons why this wouldn't be
accepted?

Alternatively, it may be that the correct approach would be to leave
ipconfig alone, and add a minimal nsupdate to klibc so that the hostname
can be read out of root/etc/hostname later in the pre-boot, then passed on
to DNS. Thoughts on the relative merrits of the two approaches welcomed.

Finally, and sorry for asking a 'howto' question, are there any pointers
on
getting klibc to duild on Ubuntu? I'm being frustrated by location of
kernel headers, and don't want to go adding symlinks from /usr/src.

Phil
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://www.zytor.com/pipermail/klibc/attachments/20150713/218ac525/attachment.html>

On 07/13/2015 12:03 AM, Phil Lello wrote:
> Hi all,
> 
> I've recently been experimenting with Xen VMs booted using DHCP/NFS,
and
> that's looking good.
> 
> One nuisance is that I'm unable to get the DNS to update on the server
> since there's no hostname in the DHCP request packet.
> 
> I'm aware that I can specify a hostname using the long-format device
> name, but this doesn't seem to play well with initramfs. So, I'd
like to
> add a patch to ipconfig that allows a parameter to specify the hostname.
> Are there any immediate objections to this, or reasons why this
wouldn't
> be accepted?
> 
> Alternatively, it may be that the correct approach would be to leave
> ipconfig alone, and add a minimal nsupdate to klibc so that the hostname
> can be read out of root/etc/hostname later in the pre-boot, then passed
> on to DNS. Thoughts on the relative merrits of the two approaches welcomed.
> 
> Finally, and sorry for asking a 'howto' question, are there any
pointers
> on getting klibc to duild on Ubuntu? I'm being frustrated by location
of
> kernel headers, and don't want to go adding symlinks from /usr/src.
> 
Typically you would execute a secondary DHCP request to make that
happen.  You *have* to do that anyway, because ipconfig will not renew a
lease.

	-hpa

Many thanks for that - I'd overlooked the need to refresh leases, so will
look at changes elsewhere in the system.

Phil

On Fri, Aug 7, 2015 at 10:14 PM, H. Peter Anvin <hpa at zytor.com> wrote:
> On 07/13/2015 12:03 AM, Phil Lello wrote:
> > Hi all,
> >
> > I've recently been experimenting with Xen VMs booted using
DHCP/NFS, and
> > that's looking good.
> >
> > One nuisance is that I'm unable to get the DNS to update on the
server
> > since there's no hostname in the DHCP request packet.
> >
> > I'm aware that I can specify a hostname using the long-format
device
> > name, but this doesn't seem to play well with initramfs. So,
I'd like to
> > add a patch to ipconfig that allows a parameter to specify the
hostname.
> > Are there any immediate objections to this, or reasons why this
wouldn't
> > be accepted?
> >
> > Alternatively, it may be that the correct approach would be to leave
> > ipconfig alone, and add a minimal nsupdate to klibc so that the
hostname
> > can be read out of root/etc/hostname later in the pre-boot, then
passed
> > on to DNS. Thoughts on the relative merrits of the two approaches
> welcomed.
> >
> > Finally, and sorry for asking a 'howto' question, are there
any pointers
> > on getting klibc to duild on Ubuntu? I'm being frustrated by
location of
> > kernel headers, and don't want to go adding symlinks from
/usr/src.
> >
>
> Typically you would execute a secondary DHCP request to make that
> happen.  You *have* to do that anyway, because ipconfig will not renew a
> lease.
>
>         -hpa
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://www.zytor.com/pipermail/klibc/attachments/20150810/8ea043a9/attachment-0001.html>