8:28 a.m.
At one point Adam had really suggested we look at how guests are defined
and changed. In particular, straight copies of XML and or straight
edits of XML could lead (after time) to the XML digressing into
something libvirt might not understand at some point.
To prevent this sort of problem, we thought that each time a guest is
edited, the XML is created entirely anew but of course using the same
values. This has the distinct advantage of making future additions of
options easier to integrate. It is also the exact basis I think would
be appropriate for any sort of cloning and dovetails nicely into your
proposal. About a month ago I had a working version of this in draft
form.
Would anyone be interested in reviewing it and or adding to it prior to
implementing the clone function? If so, ping me on IRC and I'll get you
a copy of the patch.
Brent
On Thu, 2014-10-02 at 15:05 -0300, CrÃstian Viana wrote:
Hi everyone,
I'm presenting here my proposal for the feature "Guest cloning" which
is expected to be implemented for Kimchi 1.4.
Description
Cloning a guest means creating a new guest with a copy of the settings
and data of the original guest. All data described by its XML will be
copied completely, with the following exceptions:
* name: the new guest will have an automatically generated name.
We can append "-clone<n>" to the original guest's name,
where
<n> is related to the number of clones created from that
guest. For example, cloning a guest named "myfedora" will
create a new guest named "myfedora-clone1"; if another clone
for that same guest is requested, it will be named
"myfedora-clone2".
* uuid: the new guest will have an automatically generated UUID.
We can create a random UUID for every cloned guest.
* devices/interface/mac: the new guest will have an
automatically generated MAC address for every network
interface. We can create random MAC addresses for every cloned
guest.
* devices/disk: the new guest will have copies of the original
guest's disks. Depending on the storage pool type of each
disk, a different procedure may be used to copy that disk:
* DIR, NFS, Logical: the disk file will be copied to a
new file with a modified name (e.g. "disk.img" ->
"disk-clone1.img") on the same storage pool.
* SCSI, iSCSI: the volume data will be copied as a new
disk file on the storage pool "default".
REST API
Only one new REST command will be added.
Syntax
POST /vms/<vm-name>/clone
Parameters:
None.
Return:
An asynchronous Task with "target_uri" containing
"/vms/<new-vm-name>".
As expected with any Task, the cloning process can be tracked by
checking the corresponding task's status.
Discussion
I think the most challenging part of this feature is how to deal with
different types of disks while not prompting the user with any input.
There are a lot of possibilities and a lot of things that can go wrong
during the disks copy but we still need to do whatever is easier for
the user. For example, do we really have to create the new disks in
the same storage pool as the original disk's? If that's not possible
(e.g. not available space), should we create them in another pool with
available space? Should we ask any input from the user (e.g. "Would
you
like to create the new disk on the same storage pool or on a different
one?")? What about the *SCSI pool types, is it OK to copy the volume
data to a different storage pool (i.e. "default") like I'm proposing
here? I couldn't think of a way to add a new volume in an existing
pool
of those types. How about making the *SCSI volumes shareable between
the original and the new VMs? I don't like that approach because then
both VMs will use the same disk, whatever is changed in one VM is also
changed in the other one, and that's not a clone for me, that's a
"hardlink".
Any feedback is welcome!
Best regards,
CrÃstian.
_______________________________________________
Kimchi-devel mailing list
Kimchi-devel(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/kimchi-devel