Before I ask for help let me just say... wow! What an amazing product!
This has the potential to shake the SAN market profoundly. I was
disappointed in Luster because it made itself sounds like it didn't
require a SAN but you folks are strait forward and to the point. Kudos!
Now on to the problem:
The process for the glusterfs client is spiking to 100% cpu usage and
not responding (have to kill -9) whenever I add a stripe or unify
translator to the client volume spec.
There isn't anything in the client log, but the server logs show:
[Mar 03 19:11:01] [ERROR/common-utils.c:52/full_rw()]
libglusterfs:full_rw: 0 bytes r/w instead of 113
This only occurs on file writes. I can touch and read files just fine.
None of these problems appear when I just mount a remote volume
without the translator.
I'm using the latest glusterfs-1.3.0-pre2 code on centos with
fuse-2.6.3 (had to apply a patch before fuse module loaded)
My client volspec is below:
volume client0
type protocol/client
option transport-type tcp/client
option remote-host 192.168.1.201
option remote-port 6996
option remote-subvolume testgl
end-volume
volume client1
type protocol/client
option transport-type tcp/client
option remote-host 192.168.1.202
option remote-port 6996
option remote-subvolume testgl
end-volume
volume stripe
type cluster/stripe
subvolumes client1 client0
option stripe-size 131072 # 128k
end-volume
#volume bricks
# type cluster/unify
# subvolumes client1 client0
# option scheduler rr
#end-volume
