Hello.
I succesfully compiled recent (17.02) CVS code with vanilla 2.6.10 kernel
on Gentoo distro. As it's upgrade from old version, I shut down GFS on both nodes
and I'm trying to start only one node for a test. Everything seems to be fine,
until I mount the GFS file system:
share2 root # mount -t gfs /dev/sdb1 /eva/
mount: permission denied
Dmesg says after that:
GFS: Trying to join cluster "lock_dlm", "cluster1:eva"
lock_dlm: fence domain not found; check fenced
GFS: can't mount proto = lock_dlm, table = cluster1:eva, hostdata =
What might be wrong? Here's additional info:
share2 root # cat /proc/cluster/status
Protocol version: 5.0.1
Config version: 2
Cluster name: cluster1
Cluster ID: 26777
Membership state: Cluster-Member
Nodes: 1
Expected_votes: 1
Total_votes: 1
Quorum: 1
Active subsystems: 3
Node name: share2
Node addresses: 10.1.20.29
share2 root # cat /proc/cluster/nodes
Node Votes Exp Sts Name
1 1 1 M share2
share2 root # cat /proc/cluster/services
Service Name GID LID State Code
Fence Domain: "default" 1 2 join S-6,20,1
[1]
DLM Lock Space: "clvmd" 2 3 run -
[1]
I have no clue why canot mount the fs. Do I have to start both nodes
simultaneously? Than you in advance for any hints!
Piotr
My cluster.conf:
<?xml version="1.0"?>
<cluster name="cluster1" config_version="2">
<cman two_node="1" expected_votes="1">
</cman>
<clusternodes>
<clusternode name="share1" votes="1">
<fence>
<method name="single">
<device name="human" ipaddr="10.1.20.30"/>
</method>
</fence>
</clusternode>
<clusternode name="share2" votes="1">
<fence>
<method name="single">
<device name="human" ipaddr="10.1.20.29"/>
</method>
</fence>
</clusternode>
</clusternodes>
<fencedevices>
<fencedevice name="human" agent="fence_manual"/>
</fencedevices>
</cluster>
----------------------------------------------------------------------
INTERIA.PL >>> http://link.interia.pl/f1853
