> +Proposed changes
> +================
> +
> +Definitions
> +-----------
> +
> +* All disks have exactly one *disk type*, which determines where and how the
> +  disk can be stored. If the disk is transfered, the disk type remains
> +  unchanged.

This is a bit misleading, as we usually use the term "disk type" to referr to
the way the disk is built up, i.e., things like "drbd", "plain", and "file".
However, already now we can specify which volume group a disk for an instance
is to be taken from---and that is supposed to remain unchanged over instance
moves. A typical use case for this already existing feature is to have two 
volume
groups formed of different physical disks, e.g., solid-state disk and hard 
drives.

So, is this notion supposed to be "disk type" in the sense we use it for
disk templates (then the assumptions that all providers of that disk type
are equal does not hold, already now) or is it supposed to be an abstract
name space supposed to take also into account how storage for that type is
provided (then I'd prefere a name not easily confusable with the notion
of disk type we already have)?
 

-- 
Klaus Aehlig
Google Germany GmbH, Dienerstr. 12, 80331 Muenchen
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Geschaeftsfuehrer: Graham Law, Christine Elizabeth Flores

Reply via email to