[Gluster-users] Volume creation time?

[Jason Hilton]

Thank you Niels!  That did the trick.  I have everything worked out now and I'm even able to access the NFS shares via windows clients.  Your help was much appreciated.


-----Original Message-----
From: Niels de Vos [mailto:ndevos at redhat.com] 
Sent: Monday, March 23, 2015 4:58 AM
To: Jason Hilton
Cc: 'gluster-users at gluster.org'
Subject: Re: [Gluster-users] Volume creation time?

On Sun, Mar 22, 2015 at 06:21:52PM +0000, Jason Hilton wrote:
> So I have been trying to get this going and the volume status command 
> shows that the "NFS Server on localhost" is offline, port and pid are 
> "N/A".  I have double checked everything and NFS is disabled on start 
> up an are not currently running.  The showmount command returns
> "clnt_create: RPC: Program not registered"

Please make sure that none of the RPC programs for NFS are registered at rpcbind before starting Gluster/NFS (or rebooting).

You can check by executing 'rpcinfo' on the Gluster/NFS server. Any of the mountd, nlockmgr, status and nfs can block the registration of the NFS-server (and its helper protocols). Unregistering one of the protocols can be done like this:

    # rpcinfo
    ...
    100005    3    tcp       0.0.0.0.150.65         mountd     superuser
    ...

    # rpcinfo -d 100005 3

Do this for all the protocols (+versions) mentioned above and restart the Gluster/NFS server process:

    # gluster volume start $VOLUME force

The 'start force' will only start any missing processes, like the Gluster/NFS (a 'glusterfs' process). Running processes (like the ones for the bricks) should not be impacted.

Niels

> 
> Any ideas?
> Thanks again!
>  
> -----Original Message-----
> From: Niels de Vos [mailto:ndevos at redhat.com]
> Sent: Sunday, March 22, 2015 11:45 AM
> To: Jason Hilton
> Cc: 'gluster-users at gluster.org'
> Subject: Re: [Gluster-users] Volume creation time?
> 
> On Sun, Mar 22, 2015 at 02:59:53PM +0000, Jason Hilton wrote:
> > Thank you for the quick reply!  I didn't expect to see any response 
> > on a Sunday.  I did as you suggested and found  some messages 
> > stating that the address and port were failing to bind because it 
> > was already in use.  It turned out that the NFS service was running 
> > and interfered with glusterd.  I was intending to share my gluster 
> > volumes via NFS and I thought I had read that as of V3, gluster 
> > exported NFS shares by default, so I had started the service.  Does 
> > gluster provide its own NFS services?
> 
> Yes, Gluster comes indeed with its own NFS-server. You should not start any NFS-services, Gluster takes care of starting them. The only service that you need to have running (or activated for systemd environments), is rpcbind.
> 
> Once your volume has been created and started, you should be able to see that there is a NFS-server running with this command:
> 
>     # gluster volume status
> 
> And, with 'showmount -e' the volme should be listed as an export.
> 
> Cheers,
> Niels
> 
> > 
> > ***************************************************************
> > Jason Hilton
> > Director of Technology Development
> > 601 Madison Street, Suite 400
> > Alexandria, VA 22314
> > jason.hilton at aaae.org
> > Desk: 703.824.0500x167
> > FAX: 703.578.4952
> > 
> > AAAE Tech support:
> > IET at aaae.org
> > 703.797.2555, opt. 2
> > ***************************************************************
> > 
> > 
> > -----Original Message-----
> > From: Niels de Vos [mailto:ndevos at redhat.com]
> > Sent: Sunday, March 22, 2015 10:13 AM
> > To: Jason Hilton
> > Cc: 'gluster-users at gluster.org'
> > Subject: Re: [Gluster-users] Volume creation time?
> > 
> > On Sun, Mar 22, 2015 at 01:34:24PM +0000, Jason Hilton wrote:
> > > Hi-
> > > I'm new to GlusterFS and I have been trying to set up a gluster 
> > > volume.  The volume is 150 TB.  I started the create volume 
> > > command on Friday morning and it has not yet completed.  Since I 
> > > have no prior experience with GlusterFS, is this an expected 
> > > duration?  The server is no power house, a pair of older Xeon Quad 
> > > core processors at 2 GHz and only 4 GB of RAM.  TOP shows very 
> > > little processor usage, but IOTOP shows some disk I/O.  I don't 
> > > mind waiting it out, I just want to be sure that the process is 
> > > still proceeding.  Is there a way to monitor Gluster volume creation progress?
> > 
> > Volume creation should be very fast, there is not a lot to do for Gluster to create a volume. A couple of seconds should be sufficient.
> > 
> > Check the /var/log/glusterfs/etc-*.log to see if there are any errors listed there.
> > 
> > HTH,
> > Niels