Cluster HA fence_ifmib

[vince_122]

Hi there,

I am trying to enable HA for my cluster with 3 nodes.
As we're out of money, the only solution to "fence" my nodes is to use the fence_ifmib command.

I'm trying to implement it but I can't find where is my error.

Here is my cluster.conf :

<?xml version="1.0"?>
<cluster config_version="43" name="CLUSTER-PROXMOX">
<cman keyfile="/var/lib/pve-cluster/corosync.authkey"/>
<fencedevices>
<fencedevice agent="fence_ifmib" ipaddr="[SW_IP]" community="[MY_COM]" snmp_version="2c" name="fence"/>
</fencedevices>
<clusternodes>
<clusternode name="vxplfsrv17" nodeid="1" votes="1"/>
<fence>
<method name="FENCE">
<device name="fence" port="1002"/>
</method>
</fence>
<clusternode name="vxplfsrv18" nodeid="2" votes="1"/>
<clusternode name="vxplfsrv19" nodeid="3" votes="1"/>
</clusternodes>
<rm/>
</cluster>

Where is my error ? When I'm saving my file, i get :

[dcdb] notice: wrote new cluster config '/etc/cluster/cluster.conf'
[dcdb] crit: cman_tool version failed with exit code 1#010

If i'm trying directly from command line :

./fence_ifmib -o status -a [SW_IP] -d 2c -c [MY_COM] -n 1002
Status: ON

Thanks for helping,
Vince.

 

[dietmar]

You increased the version number before save?

 

[vince_122]

Yes I made modification at each time.

But I looked all syslogs, and I see the error only on my 1st node, not on the others :

Node1 :

[dcdb] notice: wrote new cluster config '/etc/cluster/cluster.conf'
[dcdb] crit: cman_tool version failed with exit code 1#010

Node 2 & 3 :
pmxcfs[1495]: [dcdb] notice: wrote new cluster config '/etc/cluster/cluster.conf'

On all nodes, In /etc/cluster/cluster.conf I have version 42
But "cman_tool version" returns :
6.2.0 config 41

Normal ?

How can I debug this error about cman_tool ?

Thanks for all !
Vince.

 

[vince_122]

Auto reply :

I deleted a line, and all worked !

Now it looks like this :

<?xml version="1.0"?>
<cluster config_version="43" name="CLUSTER-PROXMOX">
<cman keyfile="/var/lib/pve-cluster/corosync.authkey"/>
<fencedevices>
<fencedevice agent="fence_ifmib" ipaddr="135.117.232.231" community="complf" snmp_version="2c" name="fence"/>
</fencedevices>
<clusternodes>
<clusternode name="vxplfsrv17" nodeid="1" votes="1"/>
<clusternode name="vxplfsrv18" nodeid="2" votes="1"/>
<clusternode name="vxplfsrv19" nodeid="3" votes="1"/>
</clusternodes>
<rm>
</rm>
</cluster>

But what was the error on this group ? :

<clusternode name="vxplfsrv17" nodeid="1" votes="1"/>
<fence>
<method name="FENCE">
<device name="fence" port="1002"/>
</method>
</fence>

Thanks again..