Re: [openstack-dev] [Glance] Recall for previous iscsi backend BP

2014-11-20 Thread Duncan Thomas
It is quite possible that the requirement for glance to own images can be achieved by having a glance tenant in cinder, and using clone and volume-transfer functionalities in cinder to get copies to the right place. I know there is some attempts to move away from the single glance tenant model for

Re: [openstack-dev] [Glance] Recall for previous iscsi backend BP

2014-11-19 Thread Alex Meade
Hey Henry/Folks, I think it could make sense for Glance to store the volume UUID, the idea is that no matter where an image is stored it should be *owned* by Glance and not deleted out from under it. But that is more of a single tenant vs multi tenant cinder store. It makes sense for Cinder to at

Re: [openstack-dev] [Glance] Recall for previous iscsi backend BP

2014-11-19 Thread henry hly
Hi Flavio, Thanks for your information about Cinder Store, Yet I have a little concern about Cinder backend: Suppose cinder and glance both use Ceph as Store, then if cinder can do instant copy to glance by ceph clone (maybe not now but some time later), what information would be stored in glance

Re: [openstack-dev] [Glance] Recall for previous iscsi backend BP

2014-11-19 Thread Duncan Thomas
I think that having a stand-alone (client of cinder) rich data streaming service (http put/get with offset support, which can be used for conventional glance plus volume upload/download directly), and rich data-source semantics exposed so that it can be used in an optimal way by/for nova, need not

Re: [openstack-dev] [Glance] Recall for previous iscsi backend BP

2014-11-19 Thread Flavio Percoco
On 19/11/14 15:21 +0800, henry hly wrote: In the Previous BP [1], support for iscsi backend is introduced into glance. However, it was abandoned because of Cinder backend replacement. The reason is that all storage backend details should be hidden by cinder, not exposed to other projects. Howeve