Hey,
On zo, 2013-09-22 at 11:32 +0800, 王根意 wrote:
> Hi guys:
>
>
> From the ceph documentation we knew that
>
>
> From a user’s perspective, a clone is just like any other rbd
> image.
>
>
> But does anyone know the difference between "rbd clone" and "rbd cp".
>
>
> When we changed parrent block image, does it have effect to children
> image ?
>
>
> And why can "rbd clone" do "fast image creation" ?
"rbd clone" works by creating a copy-on-write clone off of an existing
snapshot. This snapshot needs to be locked first and therefore cannot
be deleted as long as the clone exists in copy-on-write form. You can
change it's parent, just like with any other snapshot. The "rbd
flatten"-command fills in the gaps in a copy-on-write clone from it's
snapshot, after which it becomes a regular image and the snapshot can be
unlocked and deleted. This flattening-process will have to run through
the entire image, taking about as much time as an "rbd cp" would have
taken.
PS: Can anyone tell me if flattening a copy-on-write clone while this
clone is actually in use by, for example, a VM is OK and stable?
Regards,
Oliver
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
