[one-users] OpenNebula and ceph: why "rbd copy" rather than "rbd clone"?
jmelis at opennebula.org
Fri Feb 21 01:35:27 PST 2014
The initial OpenNebula Ceph driver implementation used "rbd copy"
instead of "rbd clone"
basically for simplicity reasons. The driver has evolved (thanks to a
patch submitted by Bill Campbell) and it now uses "rbd clone" as you
The code will be available in OpenNebula 4.6, but if you want to try
it out right away, it's merged into the master branch, so you can git
clone and compile.
On Fri, Feb 21, 2014 at 7:16 AM, Stuart Longland <stuartl at vrt.com.au> wrote:
> Hi all,
> Silly question. I've just thrown together an OpenNebula cluster for
> evaluation purposes and have it using a back-end Ceph cluster for storage.
> Image creation works, but one thing I observe is that the VMs sit in the
> PROLOG stage for a long time. During this process, I see an rbd process
> busily copying the image to a temporary rbd for the instance.
> The call specifically is a 'rbd copy' command. I see in the manual
> pages for rbd:
>> clone [parent-snapname] [image-name]
>> Will create a clone (copy-on-write child) of the parent snapshot. Object order will be identical to
>> that of the parent image unless specified. Size will be the same as the parent snapshot.
> Surely this would be more efficient than creating copies of the images.
> Is there a reason why 'copy' was chosen rather than 'clone' in OpenNebula?
> Stuart Longland
> Systems Engineer
> _ ___
> \ /|_) | T: +61 7 3535 9619
> \/ | \ | 38b Douglas Street F: +61 7 3535 9699
> SYSTEMS Milton QLD 4064 http://www.vrt.com.au
> Users mailing list
> Users at lists.opennebula.org
OpenNebula - Flexible Enterprise Cloud Made Simple
www.OpenNebula.org | jmelis at opennebula.org
More information about the Users