[Gluster-devel] very big problem

[nicolas prochazka]

Hello,

my config is very standard, see at this end of this email.

With 2.0 branche ( configuration file with same translator) , glusterfs is working.

With 3.0.x ( 3.0.0 / 3.0.1 ) glusterfs is not working correctly : 

When I boot system ( 2 servers in replicate mode) , all is ok, as soon I desactivate network card of glusterfs server ( eth1 ) , my two servers reboot between 2 and 5 minutes !!

Log logs nothing as server reboot, i'm trying with kernel from 2.6.32 -> 2.6.33 : same things

If i running just glusterfs server, without client, nothing reboot.

I suspect a fuse module bad usage with gluster

Regards, 

NP

volume 10.98.98.1

type protocol/client

option transport-type tcp/client

option remote-host 10.98.98.1

option transport.socket.nodelay on

option remote-subvolume brick

end-volume

volume 10.98.98.10

type protocol/client

option transport-type tcp/client

option remote-host 10.98.98.10

option transport.socket.nodelay on

option remote-subvolume brick

end-volume

# -------------    Create Brick Replicate  definition

volume last0

type cluster/replicate

subvolumes 10.98.98.1 10.98.98.10

end-volume

# -------------    Create Distribute definition

volume last

type cluster/distribute

subvolumes  last0

end-volume

volume iothreads

type performance/io-threads

option thread-count 8

subvolumes last

end-volume

volume io-cache

type performance/io-cache

option cache-size 2GB             # default is 32MB

option cache-timeout 5  # default is 1

subvolumes iothreads

end-volume

volume writebehind

type performance/write-behind

option cache-size 4MB

subvolumes io-cache

end-volume

cat /etc/glusterfs/glusterfs-server.vol

volume brickless

type storage/posix

option directory /mnt/disks/export

end-volume

volume brickthread

type features/locks

subvolumes brickless

end-volume

volume brickcache

type performance/io-cache

option cache-size 2GB             # default is 32MB

option cache-timeout 2  # default is 1

subvolumes brickthread

end-volume

volume brick

type performance/io-threads

option thread-count 8

subvolumes brickcache

end-volume

volume server

type protocol/server

subvolumes brick

option client-volume-filename /etc/glusterfs/Gglusterfs-client.vol

option transport-type tcp

option transport.socket.nodelay on

option verify-volfile-checksum no

option auth.addr.brick.allow 10.98.98.*

end-volume