On Thu, Mar 20, 2014 at 09:49:05AM -0400, Steve Dickson wrote:
On 03/20/2014 03:33 AM, Noam Meltzer wrote:
>
>
>
> On Thu, Mar 13, 2014 at 7:41 PM, Simo Sorce <simo(a)redhat.com
<mailto:simo@redhat.com>> wrote:
>
> On Thu, 2014-03-13 at 13:25 -0400, J. Bruce Fields wrote:
> > On Thu, Mar 13, 2014 at 12:58:59PM -0400, Simo Sorce wrote:
> > > On Thu, 2014-03-13 at 11:24 -0400, Steve Dickson wrote:
> > >
> > > > V4 servers now have an option of returning a uid string (aka
"3606")
> > > > where the NFS client just converts that into the uid.
> > >
> > > Can the client tell the server *not to do that* ?
> >
> > The client can use kerberos, in which case the server won't do that.
>
> This is sufficient, thanks.
>
> > Other than that, no, the behavior can only be controlled by server-side
> > configuration.
>
> A little sub-optimal if the server can do either but the client can have
> a choice, but ok.
>
>
> Sorry for the late response. A misplaced email filtering rule had moved it away from
my inbox.
> How can we push this forward?
>
Not sure since I don't maintain any of the code these patches change...
steved.
Does the code work with the current NFS idmap implementation or are
there any changes on either side needed?
Simo, are there still any issues that would prevent us from merging the
code into SSSD?