I'm trying to use fence_cisco_mds to fence an unresponsive node cutting access to storage. I've read the part about fencing on proxmox and I think I've pinpointed the problem to the cluster.rng file in /usr/share/cluster/
cluster.conf looks like this:
<?xml version="1.0"?>
<cluster name="Cluster-VMS" config_version="5">
<cman keyfile="/var/lib/pve-cluster/corosync.authkey">
</cman>
<clusternodes>
<clusternode name="athos" votes="1" nodeid="1">
<fence>
<method name="off">
<device name="fence_mds" action="off" port="fc1/15"/>
<device name="fence_mds" action="off" port="fc1/16"/>
</method>
</fence>
</clusternode>
<clusternode name="aramis" votes="1" nodeid="2">
<fence>
<method name="off">
<device name="fence_mds" action="off" port="fc1/12"/>
</method>
</fence>
</clusternode>
<clusternode name="portos" votes="1" nodeid="3">
<fence>
<method name="off">
<device name="fence_mds" action="off" port="fc1/11"/>
</method>
</fence>
</clusternode>
</clusternodes>
<fencedevices>
<fencedevice agent="fence_cisco_mds" name="fence_mds" ip="10.1.1.2" community="community" />
</fencedevices>
</cluster>
the output from ccs_config_validate is :
Relax-NG validity error : Extra element fencedevices in interleave
tempfile:31: element fencedevices: Relax-NG validity error : Element cluster failed to validate content
Configuration fails to validate
I'm guessing the problem is with attribute ip and community. I've edited the cluster.rng file to add those attributes to fencedevices but the changes don't get saved. I've also tried to issue a ccs_update_schema but this didnt work either. Can someone tell me how to fix this? Also, is there a proper documentation for this ?
PS: I've run the fenced agent directly and it works so it must be from cluster.rng.
cluster.conf looks like this:
<?xml version="1.0"?>
<cluster name="Cluster-VMS" config_version="5">
<cman keyfile="/var/lib/pve-cluster/corosync.authkey">
</cman>
<clusternodes>
<clusternode name="athos" votes="1" nodeid="1">
<fence>
<method name="off">
<device name="fence_mds" action="off" port="fc1/15"/>
<device name="fence_mds" action="off" port="fc1/16"/>
</method>
</fence>
</clusternode>
<clusternode name="aramis" votes="1" nodeid="2">
<fence>
<method name="off">
<device name="fence_mds" action="off" port="fc1/12"/>
</method>
</fence>
</clusternode>
<clusternode name="portos" votes="1" nodeid="3">
<fence>
<method name="off">
<device name="fence_mds" action="off" port="fc1/11"/>
</method>
</fence>
</clusternode>
</clusternodes>
<fencedevices>
<fencedevice agent="fence_cisco_mds" name="fence_mds" ip="10.1.1.2" community="community" />
</fencedevices>
</cluster>
the output from ccs_config_validate is :
Relax-NG validity error : Extra element fencedevices in interleave
tempfile:31: element fencedevices: Relax-NG validity error : Element cluster failed to validate content
Configuration fails to validate
I'm guessing the problem is with attribute ip and community. I've edited the cluster.rng file to add those attributes to fencedevices but the changes don't get saved. I've also tried to issue a ccs_update_schema but this didnt work either. Can someone tell me how to fix this? Also, is there a proper documentation for this ?
PS: I've run the fenced agent directly and it works so it must be from cluster.rng.