On Wed, Sep 1, 2010 at 11:33 AM, H Chang <[email protected]> wrote: > Hi Ceph team, > > I'm a newbie in Ceph, so please bear with me if my question sounds stupid. > > I'd like to explore the possibility to support multiple placement > policies in Ceph. > > For example, suppose each user using Ceph has its own placement policy > which may be different from one user to another. > So, all the files created by the user and their replicas will be placed based > on > the user's policy. Essentially each user would have its own CRUSH map. > > Do you think Ceph could somehow be modified to support this kinda scenario? > If so, where do we start? > > Regards, > HS This is something we've been asked about before and plan to look at in the future. Unfortunately it may turn out to be a bit of a chore. The CRUSH map and data distribution policies are handled by RADOS (the underlying object store) rather than the Ceph filesystem components. Specifying separate data placement policies for each user will require that each user has their own RADOS pool and that those pools all be accessible via the filesystem. Currently the Ceph filesystem has just two pools -- one for data and one for metadata management -- and it puts all the user's data in the "data" pool. Anybody implementing this kind of change would also be enabling some stronger security, which also makes it look desirable.
But it will require changing several pretty low-level structs and modifying the metadata server and both versions of the client to enable, so it's not a good introductory project and you shouldn't expect it soon. :( -Greg -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
