Glusterfs /etc/init.d/glusterfs-server stop
Hello.
I think that I found a bug/mistake/feature in init-script of
glusterfs-server.
I have some kind of a cluster of 4 nodes (HP DLs). Debian Eth 4.0 amd64
is installed on each node.
A node is a glusterfs-server and glusterfs-client at the same time. Each
glusterfs-server export 'www-data' volume, each client combines these
four 'www-data' server volumes into one RAID1 (afr) client volume.
Some kind of RAID1 on four servers...
Both glusterfs-client and glusterfs-server are running on each node:
node3:/srv/export/cluster# ps aux|grep gluster
root 6570 0.0 0.0 29544 1788 ? Ssl 17:15 0:00
[glusterfs]
root 7059 0.1 0.0 95692 1652 ? Ssl 18:55 0:03
[glusterfs]
root 7159 0.0 0.0 3936 720 pts/0 S+ 19:25 0:00 grep
gluster
node3:/srv/export/cluster#
'www-data' volume is mounted on /mnt directory.
When I'm trying to stop glustrefs-server on a node I do
'/etc/init.d/glusterfs-server stop'. After that I got the following error:
node3:/srv/export/cluster# ls -l /mnt
ls: /mnt: Transport endpoint is not connected
node3:/srv/export/cluster#
And:
node3:/srv/export/cluster# ps aux|grep gluster
root 8195 0.0 0.0 3940 724 pts/0 S+ 19:34 0:00 grep
gluster
node3:/srv/export/cluster#
So init script kills both of glusterfs client and server daemons...
I used the following software (build from unstable debian repository
sources):
node3:/srv/export/cluster# dpkg -l|grep glusterfs
ii glusterfs-client 1.3.12-1
clustered file-system
ii glusterfs-server 1.3.12-1
clustered file-system
ii libglusterfs0 1.3.12-1
GlusterFS libraries and translator modules
node3:/srv/export/cluster#
I hope someone got the same problem already...
Can anybody help me to submit this bug? I think that I know buggy line
in the init script, but I'm unsure:)
Thanks.
Reply to: