On 25 October 2011 18:05, Burnash, James <jburnash@knig...> wrote:

> Top posting for visibility:****
>
> ** **
>
> Here is the pertinent line in the log:****
>
> ** **
>
> [2011-10-25 11:39:56.762720] E [glusterfsd-mgmt.c:628:mgmt_getspec_cbk]
> 0-glusterfs: failed to get the 'volume file' from server
> [2011-10-25 11:39:56.762775] E [glusterfsd-mgmt.c:695:mgmt_getspec_cbk]
> 0-mgmt: failed to fetch volume file (key:/mnt)
>
> ****
>
> ** **
>
> Did you move your volume file or in some other way make it inaccessible for
> the client? The volume file tells the client essentially how to access files
> and directories on the GlusterFS – if the client can get that, the timeout
> is expected (it figures that the server has gone away and that you will try
> something different.****
>
> ** **
>
> I suggest trying this mount line instead (assuming that your servers are
> named consistently:****
>
> ** **
>
>                 mount -t glusterfs gluster02:/volume01 /mnt/gluster****
>
> ** **
>
> and see if you get the same message****
>
> ** **
>
> James Burnash****
>
> Unix Engineer****
>
> Knight Capital Group****
>
>

I thank you for the advice.

I've gave another look to the bricks, ando on gluster03 someone tried to
mount the volume inside the /mnt, so when the client machine tried to mount
/mn/gluster the gluster03 crashed all the servers.

Problem solved.

Thanks
MV.