[Gluster-users] [gluster client] what's different between mount.glusterfs and glusterfs

[Niels de Vos]

On Thu, Dec 11, 2014 at 11:59:18AM +0000, Jifeng Li wrote:
> Hi,
> 
> When mounting Gluster volumes to access data , I find that there are two ways listed below:
> 
> 
> 1)      glusterfs -p /var/run/glusterfs.pid --volfile-server=<vol_server> --volfile-id=<volfile_id>  <mount_point>
> 
> 2)      mount -t glusterfs -o direct-io-mode=disable,_netdev  backupvolfile-server=<backup_vol_server>   <vol_server>: <volume_name>   <mount_point>
> 
> 
> so,
> 
> 1.       what's the different between the above ways?
> 
> 2.       If using way 1, is there option similar backupvolfile-server ?

The glusterfs binary is the low-level tool that functions as a GlusterFS
client. The NFS-server, self-heal and other processes are actually
running the glusterfs binary too, just with a different set of options.

The /sbin/mount.glusterfs script is a mount helper. When executing a
standard "mount -t <fs-type>" command, the /sbin/mount.<fs-type> helper
get executed (if one exists). The script itself will parse the options
that were passed on the mount commandline, and will re-format/arrange
them and executes the glusterfs binary with those options.

You can pass multiple --volfile-server=.. options to the glusterfs
binary. This should try mounting from the next server in case the
current one fails.

HTH,
Niels
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20141211/de123ef5/attachment.sig>