[Gluster-users] Can a gluster server be an NFS client ?

Jason Brooks jbrooks at redhat.com
Tue May 19 17:07:03 UTC 2015



----- Original Message -----
> From: "Prasun Gera" <prasun.gera at gmail.com>
> To: "Jason Brooks" <jbrooks at redhat.com>
> Cc: gluster-users at gluster.org
> Sent: Monday, May 18, 2015 3:26:02 PM
> Subject: Re: [Gluster-users] Can a gluster server be an NFS client ?
> 
> Thanks. Can you tell me what this achieves, and what the side-effects, if
> any, are ? Btw, the NFS mounts on the gluster server are unrelated to what
> the gluster server itself is exporting. They are regular nfs mounts from a
> different NFS server.

Your nfs client and the gluster nfs server are both vying for portmap,
turning off locking for the client resolves this. The side effects will
depend on your environment -- in my converged ovirt+gluster setup, 
it isn't causing an issue, but I'm looking forward to dropping this 
workaround when ovirt gets support for hosting its engine vm from
native gluster.

For you, this may or may not be a workable workaround.

Jason

> 
> On Mon, May 18, 2015 at 5:35 PM, Jason Brooks <jbrooks at redhat.com> wrote:
> 
> >
> >
> > ----- Original Message -----
> > > From: "Prasun Gera" <prasun.gera at gmail.com>
> > > To: gluster-users at gluster.org
> > > Sent: Monday, May 18, 2015 1:47:32 PM
> > > Subject: [Gluster-users] Can a gluster server be an NFS client ?
> > >
> > > I am seeing some erratic behavior w.r.t. the NFS service on the gluster
> > > servers (RHS 3.0). The nfs service fails to start occasionally and
> > randomly
> > > with
> > >
> > > Could not register with portmap 100021 4 38468
> > > Program  NLM4 registration failed
> > >
> >
> > I've encountered this before -- I had to disable file locking,
> > adding Lock=False to /etc/nfsmount.conf
> >
> > > This appears to be related to
> > > http://www.gluster.org/pipermail/gluster-users/2014-October/019215.html
> > ,
> > > although I'm not sure what the resolution is.
> > >
> > > The gluster servers use autofs to mount user home directories and other
> > > sundry directories. I could verify that stopping autofs and then starting
> > > the gluster volume seems to solve the problem. Starting autofs after
> > > gluster seems to work fine too. What's the right way to handle this ?
> > >
> > > _______________________________________________
> > > Gluster-users mailing list
> > > Gluster-users at gluster.org
> > > http://www.gluster.org/mailman/listinfo/gluster-users
> >
> 


More information about the Gluster-users mailing list