Gluster users - Nov 2009 - "No space left on device" on Namespace using 2.0.8

Hi Glusterers,

I`ve got some problems using unify/namespace with glusterfs 2.0.8.
The OS (debian) reports a no space left on device on the Namespace mount 
but from the 5GB the device has only 180MB are used.
As you can see in the attached Config Files we are using 2 afr volumes, 
one for data and one for namespaces and over it lies the unify device.

What am i doing wrong?

Thanks for your help!!

--

Glusterfsd.Log:
[2009-11-10 09:57:30] E [posix.c:1724:posix_create] posix-ns: open on 
/lala failed: No space left on device
--

df -h:
/dev/mapper/vg0-data 825698728 165062512 618693176  22% /export/data
/dev/mapper/vg0-data--ns
                        5160576    179392   4719040   4% /export/data-ns
--

glusterfsd.vol:
volume posix
         type storage/posix
         option directory /export/data
end-volume

volume locks
         type features/locks
         subvolumes posix
end-volume

volume brick
         type performance/io-threads
         option thread-count 8
         subvolumes locks
end-volume

volume posix-ns
         type storage/posix
         option directory /export/data-ns
end-volume

volume locks-ns
         type features/locks
         subvolumes posix-ns
end-volume

volume brick-ns
         type performance/io-threads
         option thread-count 8
         subvolumes locks-ns
end-volume

volume server
         type protocol/server
         option transport-type tcp
         option auth.addr.brick.allow *
         option auth.addr.brick-ns.allow *
         subvolumes brick brick-ns
end-volume
--

glusterfs.vol:
volume brick1
         type protocol/client
         option transport-type tcp/client
         option remote-host 192.168.164.1
         option remote-subvolume brick
end-volume

volume brick2
         type protocol/client
         option transport-type tcp/client
         option remote-host 192.168.164.2
         option remote-subvolume brick
end-volume

volume brick1-ns
         type protocol/client
         option transport-type tcp/client
         option remote-host 192.168.164.1
         option remote-subvolume brick-ns
end-volume

volume brick2-ns
         type protocol/client
         option transport-type tcp/client
         option remote-host 192.168.164.2
         option remote-subvolume brick-ns
end-volume

volume afr1
         type cluster/afr
         subvolumes brick1 brick2
end-volume

volume afr-ns
         type cluster/afr
         subvolumes brick1-ns brick2-ns
end-volume

volume unify
         type cluster/unify
         option scheduler nufa
         option namespace afr-ns
         subvolumes afr1
end-volume

-- 
Andreas Lehr

mail: mail at andreas-lehr.com

On Tue, Nov 10, 2009 at 10:06:58AM +0100, Andreas Lehr
wrote:
> I`ve got some problems using unify/namespace with glusterfs 2.0.8.
> The OS (debian) reports a no space left on device on the Namespace mount 
> but from the 5GB the device has only 180MB are used.
> As you can see in the attached Config Files we are using 2 afr volumes, 
> one for data and one for namespaces and over it lies the unify device.
> 
> What am i doing wrong?
> 
> Thanks for your help!!
> 
> Glusterfsd.Log:
> [2009-11-10 09:57:30] E [posix.c:1724:posix_create] posix-ns: open on 
> /lala failed: No space left on device
> --
> 
> df -h:
> /dev/mapper/vg0-data 825698728 165062512 618693176  22% /export/data
> /dev/mapper/vg0-data--ns
>                        5160576    179392   4719040   4% /export/data-ns
IMHO your namespace is out of inodes. Namespace contains a lot of empty
files, so you end with a lot of free disk space and all inodes used.
Try df -i.
If you desperately want to keep the namespace on a separate filesystem,
use a filesystem, which allocates the inodes on an as-needed basis
or, for a filesystem with fixed number of inodes, create the fs with
largest possible number of inodes (and a little free space for
directory entries and symlinks ;).
Krzysztof