[Gluster-devel] glusterfs with vservers

Martin Fick mogulguy at yahoo.com
Fri Nov 21 18:21:41 UTC 2008


--- On Fri, 11/21/08, Anand Avati <avati at zresearch.com> wrote:

> > B)
> >
> > 2008-11-20 17:40:32 W
> [client-protocol.c:280:client_protocol_xfer] trans:
> > attempting to pipeline request type(2) op(4) with
> handshake
> > 2008-11-20 17:40:32 E [fuse-bridge.c:2702:init]
> glusterfs-fuse: fuse_mount
> > failed (Transport endpoint is not connected)
> >
> > 2008-11-20 17:40:32 E [glusterfs.c:547:main]
> glusterfs: Initializing FUSE
> > failed
> >
> 
> The above error means there is a stale mount on /mnt.
> umount /mnt and try again. 

Hmm, well to ensure that is not the case, I just 
rebooted my machine.  There are no mounts on /mnt
or on /mnt in the guest.  I then retired only the

 vnamespace -e 221 glusterfs -s 10.10.20.11 mnt


and got the same client and server errors.  From
the looks of the server logs it seems like the 
client attempts to connect twice, once looking
for a special config for its IP (which is
confused as the server's IP) and once without
looking for such a config.  Is that correct?

Could this double attempt be the problem somehow?
Could the first attempt be leaving around a stale
mount so that the second attempt cannot succeed?



> Not quite sure about your other errors, seems not to
> be glusterfs related. Do you know if having support 
> for glusterfs client to bind() to a specific local 
> IP address will resolve this issue?

I assume it is actually a vserver problem, but since
I noticed the wrong IP problem, I thought that it might
be the root of the problem.  No, I do not know that 
binding to a specific IP will solve the mounting 
problem, but as I mentioned it could potentially be 
an authorization issue for me & others, and a worthwhile
switch to have in general, do you agree?

Thanks,

-Martin



      





More information about the Gluster-devel mailing list