[Gluster-users] nfs-ganesha with disperse volume

[Niels de Vos]

On Fri, Jan 08, 2016 at 04:18:26PM +0200, Serkan Çoban wrote:
> Below is the snippet from ganesha.conf documentation, what will be the
> Path if i don't need to mount gluster volume on server?
> 
> EXPORT{
> Export_Id = 1 ; # Export ID unique to each export
> Path = "volume_path"; # Path of the volume to be exported. Eg: "/test_volume"

This is the path that can be mounted by NFS-clients. NFSv3 will be able
to show this path when you execute "showmount $NFS_SERVER".

If you set the Path="/test_volume" like the example, an NFS-client will
be able to do:

  # mount -t nfs $NFS_SERVER:/test_volume /mnt

Cheers,
Niels

> 
> On Fri, Jan 8, 2016 at 2:40 PM, Niels de Vos <ndevos at redhat.com> wrote:
> > On Fri, Jan 08, 2016 at 02:04:53PM +0200, Serkan Çoban wrote:
> >> Thanks, I understand it. Now I am trying to export a volume through
> >> nfs-ganesha but have a question again.
> >> In the EXPORT block in ganesha.conf, for the export path I need the
> >> enter the full path to gluster volume.
> >> So this means that I need to mount the gluster volume on the server
> >> right? Currently I don't have gluster volume mounted on servers.
> >> If this is true each server also will be a client right?
> >
> > No, you do not need to mount the Gluster volume on the system that has
> > NFS-Ganesha running. The documentation should explain pretty good what
> > you need to configure:
> >
> >   http://gluster.readthedocs.org/en/latest/Administrator%20Guide/NFS-Ganesha%20GlusterFS%20Intergration/
> >
> > Niels
> >
> >>
> >> Serkan
> >>
> >> On Fri, Jan 8, 2016 at 12:09 PM, Niels de Vos <ndevos at redhat.com> wrote:
> >> > On Fri, Jan 08, 2016 at 11:55:52AM +0200, Serkan Çoban wrote:
> >> >> Hi,
> >> >>
> >> >> I want to ask if we can use nfs-ganesha with disperse volumes? If so,
> >> >> how the underlying process works? Gluster client split the file to
> >> >> chunks and write each chunk to different server in cluster; in case of
> >> >> nfs how this will work?
> >> >
> >> > Yes, that should work just fine.
> >> > The NFS server acts as a Gluster client. So, on one side the NFS-server
> >> > speaks the standard NFS protocol to the clients, on the other side it
> >> > talks to the bricks in the volume. The logic of the distribution and
> >> > splitting of data is done inside the NFS-server (by libgfapi in
> >> > FSAL_GLUSTER for NFS-Ganesha).
> >> >
> >> > HTH,
> >> > Niels
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20160108/6745961f/attachment.sig>