HA and an APC (AP7921)

[vostor]

ok I have 3 Proxmox boxes configured, 1 storage (ISCSI Target) and an AP7921.. The problem i'm getting is if i power a socket off on the APC the servers do not failover until the power is restored and the old box is running! I am sure its to do with my cluster.config

<?xml version="1.0"?>
<cluster config_version="13" name="farm-gp01">
  <cman keyfile="/var/lib/pve-cluster/corosync.authkey"/>
  <fencedevices>
    <fencedevice agent="Fence_apc" ipaddr="x.x.x.x" login="admin" name="APC-R00" passwd="123456"/>
  </fencedevices>
  <clusternodes>
    <clusternode name="pm-srv02" nodeid="1" votes="1">
      <fence>
        <method name="power">
          <device name="APC-R00" port="1" />
        </method>
      </fence>
    </clusternode>
    <clusternode name="pm-srv01" nodeid="2" votes="1">
      <fence>
        <method name="power">
          <device name="APC-R00" port="2" />
        </method>
      </fence>
    </clusternode>
    <clusternode name="pm-srv03" nodeid="3" votes="1">
      <fence>
        <method name="power">
          <device name="APC-R00" port="3" />
        </method>
      </fence>
    </clusternode>
  </clusternodes>
  <rm>
    <pvevm autostart="1" vmid="100"/>
    <pvevm autostart="1" vmid="101"/>
  </rm>
</cluster>

Any help is always welcome!

 

[dietmar]

That behavior indicates that fencing is not working - please test.

 

[tom]

works here. and compare your settings with the example, http://pve.proxmox.com/wiki/Fencing#APC_Switch_Rack_PDU

e.g. you did not set "secure=on", and did you enable ssh on the APC?

 

[vostor]

tried with and without secure="on" SSH is enabled

all 3 servers can see the AP7921 and seem to be able to interact with it using fence_apc they can all SSH to the AP but its only when the servers are coming back up and live does it migrate.

I am confused about this section not sure about it at all

  [COLOR=#009900][COLOR=#000000][B]<rm>[/B][/COLOR][/COLOR]    [COLOR=#009900][COLOR=#000000][B]<service[/B][/COLOR] [COLOR=#000066]autostart[/COLOR]=[COLOR=#FF0000]"1"[/COLOR] [COLOR=#000066]exclusive[/COLOR]=[COLOR=#FF0000]"0"[/COLOR] [COLOR=#000066]name[/COLOR]=[COLOR=#FF0000]"TestIP"[/COLOR] [COLOR=#000066]recovery[/COLOR]=[COLOR=#FF0000]"relocate"[/COLOR][COLOR=#000000][B]>[/B][/COLOR][/COLOR]      [COLOR=#009900][COLOR=#000000][B]<ip[/B][/COLOR] [COLOR=#000066]address[/COLOR]=[COLOR=#FF0000]"192.168.7.180"[/COLOR][COLOR=#000000][B]/>[/B][/COLOR][/COLOR]    [COLOR=#009900][COLOR=#000000][B]</service>[/B][/COLOR][/COLOR]  [COLOR=#009900][COLOR=#000000][B]</rm>

[/B][/COLOR][/COLOR]

 

[dietmar]

all 3 servers can see the AP7921 and seem to be able to interact with it using fence_apc they can all SSH to the AP but its only when the servers are coming back up and live does it migrate.

You need to test fencing: 'fence_node <nodename>'

# man fence_node

I am confused about this section not sure about it at all

What is the question (This configures a failover IP address)?