On Feb 5, 2016 5:42 AM, "Vijay Bellur" <vbellur@xxxxxxxxxx> wrote: > > On 02/04/2016 10:35 AM, Kaushal M wrote: >> >> On Thu, Feb 4, 2016 at 8:51 PM, Pranith Kumar Karampuri >> <pkarampu@xxxxxxxxxx> wrote: >>> >>> >>> >>> On 02/04/2016 08:48 PM, Kaushal M wrote: >>>> >>>> >>>> I'm still up to mentor the sub-directory mount support idea. >>> >>> >>> this one? http://review.gluster.org/10186 >> >> >> This seems to have been revived. IIRC, Avati's change did it >> everything server side. IMO, this is not the most suitable approach. >> One of the reasons we want this support is to do multi-tenancy which >> requires proper access control. To have access control, we need this >> to be done on the server side. >> > > Can you elaborate a bit more on the changes needed on the server side? These are the changes I expect, - Do all the changes (or similar changes) in 10186 for the fuse xlator, except in the server xlator. - Do the required changes in the server auth code to support sub-directory acls. In addition, we'd need to check for parts in the client side graph which expect the root-gfid to be null, and fix those to handle a proper gfid for root. Since this feature is mainly targetted for a cloud, multi-tenant use case, I expect access-control to be a very important requirement. ~kaushal > > -Vijay > _______________________________________________ Gluster-devel mailing list Gluster-devel@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-devel