Re: Uploaded pv-grub-menu to NEW.

On 26 January 2014 03:37, Charles Plessy <plessy@debian.org> wrote:

Le Sun, Aug 11, 2013 at 11:42:03AM +0200, Anders Ingemann a écrit :

> On 11 August 2013 08:13, Charles Plessy <plessy@debian.org> wrote:
> >

> > I just uploaded pv-grub-menu to the NEW queue. This package borrows from the
> > grub-legacy package the scripts to create and manage the file
> > /boot/grub/menu.lst.
>

> Nicely done! So I wont need my custom grub script any longer, right?
> (https://github.com/andsens/build-debian-cloud/blob/master/grub.d/40_custom)

Hi Anders and everybody,

I just updated pv-grub-menu as follows:

pv-grub-menu (1.3) unstable; urgency=medium

* Never assume the presence of a device.map file; create a temporary one.
[840ba0c] Closes: #732408.
* Removed code for various operations that are not needed in the context
of pv-grub-menu: RAID support [c9b7f86], transition to GRUB 2 [4a702dc],
alternative locations for menu.lst [d8aae34], relative paths [1a419be],
splash images [e4d1608].

-- Charles Plessy <plessy@debian.org> Sun, 26 Jan 2014 08:59:26 +0900

I have tested that an Amazon instance can reboot with the menu.lst file created
by pv-grub-menu. However, I notice that it uses UUIDs by default, and I
suspect that it may be problematic as they may change after the image was
initially created.

I guess that the most robust solution would be to ensure that the root partition
has a label, and use it ?

Have a nice Sunday,

--
Charles Plessy
Tsurumi, Kanagawa, Japan

--
To UNSUBSCRIBE, email to debian-cloud-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Archive: [🔎] 20140126023753.GB21910@falafel.plessy.net" target="_blank">http://lists.debian.org/[🔎] 20140126023753.GB21910@falafel.plessy.net

Hello Charles

> I just updated pv-grub-menu as follows

Nice! Looking forward to test it.

> However, I notice that it uses UUIDs by default, and I suspect that

> it may be problematic as they may change after the image was initially created.

The UUIDs stays the same, even after you snapshot the volume and make a clone.

That's actually a problem when you want to debug an instance and attach the volume

to an instance that was created with the same AMI - you get a duplicate UUID error

and need to fiddle with the mount options.

> I guess that the most robust solution would be to ensure that the root partition

> has a label, and use it ?

That's only possible with GPT though, correct?

Anders