Best Wishes!
> 在 2014年5月23日，19:27，Geert Lindemulder <glindemul...@snow.nl> 写道： > > Hello Greg and Haomai, > > Thanks for the answers. > I was trying to implement the osd leveldb backend at an existing ceph > test cluster. > > At the moment i am removing the osd's one by one and recreate them with > the objectstore = keyvaluestore-dev option in place in ceph.conf. > This works fine and the backend is leveldb now for the new osd's. > The leveldb backend looks more efficient. Happy to see it, although I'm still try to improve performance for some workloads. > > The error gave me the idea that migrating from non-leveldb backend osd > to new type leveldb was possible. > Will online migration of existings osd's be added in the future? Still not, I think it's a good feature. We can implement it at ObjectStore class and simply convert one type to another > > Thanks, > Geert > >> On 05/23/2014 11:31 AM, GMail wrote: >> implement the osd leveldb backend at an existing ceph test >>>> cluster. > _______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
