[Linux-cluster] fencing with gfs
Enrico Popp
Enrico.Popp at unister-gmbh.de
Tue Dec 9 10:38:26 UTC 2008
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello,
i'm new in this list. I have to show no understanding for the fence
daemon. Why should i use this ? My issue:
I have a cluster with two nodes. And i want to use the dbrd 8.0 or
higher for a primary/primary state. There was a problem with fencing.
When the first node is down, the second node can not be rebootet by
root, because the shell says:
fence_tool: waiting for cluster quorum
Here my cluster.conf
<?xml version="1.0"?>
<cluster name="mail-intern" config_version="1">
<cman two_nodes="1" expected_votes="1"/>
<clusternodes>
<clusternode name="mail-i01" nodeid="1">
<fence>
<method name="human">
<device name="maili01" ipaddr="10.1.0.1"/>
</method>
</fence>
</clusternode>
<clusternode name="mail-i02" nodeid="2">
<fence>
<method name="human">
<device name="maili02" ipaddr="10.1.0.2" />
</method>
</fence>
</clusternode>
</clusternodes>
<fencedevices> <fencedevice name="human" agent="fence_ack_manual"/> </fencedevices> </cluster>
Is the behavior right or could i configure this. If that is so, i cannot
reboot the second node when the first isn't in the cluster, is so ?
Thank for all your time
Enrico
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFJPkqi39Du34F4DWIRAgQ+AJ0bdOkpn/QPG6kSp9ewUjEEejCsrgCglMFU
agIDkgVvhRp60mvHbCsadK0=
=6JTs
-----END PGP SIGNATURE-----
More information about the Linux-cluster
mailing list