Corrupt Filesystem - PROXMOX GLITCH

[ejc317]

This is the 2nd time this has happened. The vm shows up as stopped when the file system access is still open. HA starts multiple copies of the same vm and our vms become corrupt as the disks are getting written to at the same time ... this is really dangerous and we cannot use this

 

[tom]

Proxmox HA never starts a VM multiple times. If you see this issue, you have a major configuration problem ...

 

[ejc317]

it doesn't "start it" per se but the vm is still running and the iscsi connection is still on so it becomes corrupt

 

[tom]

hardware fencing should prevents such a behavior. post your cluster.conf.

 

[ejc317]

Here you go. My issue is I think its the web gui - it shows node is off even tho its on. I've had to routineely restart pvestard and apache

<?xml version="1.0"?>
<cluster config_version="105" name="PL-CLOUD">
<cman keyfile="/var/lib/pve-cluster/corosync.authkey"/>
<fencedevices>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.220" login="test" name="ipmi1" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.221" login="test" name="ipmi2" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.222" login="test" name="ipmi3" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.223" login="test" name="ipmi4" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.232" login="test" name="ipmi5" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.233" login="test" name="ipmi6" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.234" login="test" name="ipmi7" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.235" login="test" name="ipmi8" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.236" login="test" name="ipmi9" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.237" login="test" name="ipmi10" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.238" login="test" name="ipmi11" passwd="test" power_wait="5"/>
<fencedevice agent="fence_ipmilan" ipaddr="10.10.1.239" login="test" name="ipmi12" passwd="test" power_wait="5"/>
</fencedevices>
<clusternodes>
<clusternode name="PL-C-001" nodeid="1" votes="1">
<fence>
<method name="1">
<device name="ipmi1"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-002" nodeid="2" votes="1">
<fence>
<method name="1">
<device name="ipmi2"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-003" nodeid="3" votes="1">
<fence>
<method name="1">
<device name="ipmi3"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-004" nodeid="4" votes="1">
<fence>
<method name="1">
<device name="ipmi4"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-005" nodeid="5" votes="1">
<fence>
<method name="1">
<device name="ipmi5"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-006" nodeid="6" votes="1">
<fence>
<method name="1">
<device name="ipmi6"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-007" nodeid="7" votes="1">
<fence>
<method name="1">
<device name="ipmi7"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-008" nodeid="8" votes="1">
<fence>
<method name="1">
<device name="ipmi8"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-009" nodeid="9" votes="1">
<fence>
<method name="1">
<device name="ipmi9"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-010" nodeid="10" votes="1">
<fence>
<method name="1">
<device name="ipmi10"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-011" nodeid="11" votes="1">
<fence>
<method name="1">
<device name="ipmi11"/>
</method>
</fence>
</clusternode>
<clusternode name="PL-C-012" nodeid="12" votes="1">
<fence>
<method name="1">
<device name="ipmi12"/>
</method>
</fence>
</clusternode>
</clusternodes>
<rm/>
</cluster>