On Wed, Jun 17, 2015 at 12:36 PM, juncheng bai <baijuncheng@xxxxxxxxxxxxxxx> wrote: > Hi. > > For cloned image, it is stateless about the object. when the write > offset is smaller than the parent_ovelap,it need to call > rbd_img_obj_exists_submit to check the object status every time. > > I want to save the status of the object which it is within the scope > of parent_overlap. > > The rbd client io unit is object, so i design that use idr to save > the status. The id of object is equal to write_offset >> obj_order. > In the idr, said object already exists. And the idr can provide a > quick query. > > In this way, don't need the server do anything, but need query again > when remap. > > Looking forward to your opinion. This looks like a part of what object map is going to allow. rbd object map has been implemented in librbd in hammer and will be implemented in the kernel client as well in the future (there are few things that need to happen before that, namely exclusive locking support, which Doug is working on). https://wiki.ceph.com/Planning/Blueprints/Hammer/librbd%3A_shared_flag,_object_map Thanks, Ilya -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
