Gluster users - Mar 2010 - Strange performance issues on Debian/Lenny compared to Ubuntu

Hi -

I'm experiencing strange results when comparing the same setup on either 
Debian/Lenny and Ubuntu/Hardy.

The server is composed of two replicated nodes (both with Debian/lenny 
with GlusterFS 3.0.3, 1GbE).

Config files were generated with the following options:

glusterfs-volgen --name store1 --raid 1 gl1:/export/nunki gl2:/export/nunki

Two clients, one is an Ubuntu/Hardy (LTS, up-to-date) when the other is 
a Debian/Lenny. Altough a dd's issued on the Ubuntu client is quite 
satisfying, the write performances are disastrous on the Debian client.

ubuntu:# dd if=/gl_nunki/file_test of=/dev/null
524288+0 records in
524288+0 records out
268435456 bytes (268 MB) copied, 3,15924 s, 85,0 MB/s

lenny:# dd if=/dev/zero of=/mnt/file_test count=262144 bs=1024
262144+0 enregistrements lus
262144+0 enregistrements ??crits
268435456 bytes (268 MB) copied, 82,408 s, 3,3 MB/s

Notes that might be of interest:
- I have tried to mount the storage directly on both servers and I face 
the same poor performances
- I have tried with v3.0.0 at the client side, same result
- I have tried on three different Lenny clients (with different 
hardwares), always same results
- With all performance translators disabled, the result is even worse 
(less than 1 MB/s)

Actually, I suspect a kernel option in Debian/Lenny to be the cause of 
the problems, but TBH I don't know where to focus my attention.

Any help would be greatly appreciated!

Thanks in anticipation,
-- 
Olivier

Ooops, I have cc the wrong command on the Ubuntu server. Please correct:

 > ubuntu:# dd if=/gl_nunki/file_test of=/dev/null
 > 524288+0 records in
 > 524288+0 records out
 > 268435456 bytes (268 MB) copied, 3,15924 s, 85,0 MB/s

with:

# dd if=/dev/zero of=/mnt/file_test count=262144 bs=1024
262144+0 records in
262144+0 records out
268435456 bytes (268 MB) copied, 5,56942 s, 48,2 MB/s

Sorry for that!

-- 
Olivier