Gluster users - Mar 2011 - RedHat 5.5 and automounting with fstab

All,

I wanted to pass on experience that I have trying to get GlusterFS to
automount on RedHat 5.5.  The documentation for /etc/fstab
(http://gluster.com/community/documentation/index.php/Mounting_a_GlusterFS_Volume)
doesn't quite work.

I have two servers 192.168.104.151 and 192.168.104.152.  Both are running
GlusterD and both offer one of the bricks for the volume.  Both mount also
the volume locally.

The working /etc/fstab entry on 192.168.104.151 is:

192.168.104.151:/gluster-data /var/gluster/data glusterfs
auto,_netdev,rw,allow_other,default_permissions,max_read=131072 0 0

Likewise, the working /etc/fstab entry on 192.168.104.152 is:

192.168.104.152:/gluster-data /var/gluster/data glusterfs
auto,_netdev,rw,allow_other,default_permissions,max_read=131072 0 0


If you don't have the "auto,_netdev" options on the mount, then
you can
mount correctly from the command line with an explicit mount option or a
"mount -a" but, when you reboot the server, the mount action in the
initscripts locks up the server.

Just thought it might help someone...  YMMV, etc and all other caveats
apply.  If anyone else can confirm the above is more generally applicable
then it might make sense to update the documentation...

Many thanks for GlusterFS, by the way, it "just works".  Well done.

- Roger

Hi Roger,

>
> If you don't have the "auto,_netdev" options on the mount,
then you can
> mount correctly from the command line with an explicit mount option or a
> "mount -a" but, when you reboot the server, the mount action in
the
> initscripts locks up the server.
>
> Just thought it might help someone...  YMMV, etc and all other caveats
> apply.  If anyone else can confirm the above is more generally applicable
> then it might make sense to update the documentation...
>
>
Can you check this behavior with the latest qa release (of v3.1.3)? There
was a bug, which caused client process (the process which calls 'mount')
to
hang till it gets all the server processes up and running. It should be
fixed with v3.1.3 release.


> Many thanks for GlusterFS, by the way, it "just works".  Well
done.
>
>
Thanks.

Regards,
Amar