[Linux-cluster] cluster communication packets - changing interface

[Darek Skorupa]

Hi,

I have tried to change interface for heartbeat packets. In FAQ
(http://sourceware.org/cluster/faq.html) I've found information about
private natwork (-p) which should be added to nodename in /etc/hosts.
I've tried that but it doesn't work for me.

ver. 1
    /etc/hosts
        192.168.10.10   l1-p.local     l1-p
        192.168.10.11   l2-p.local     l2-p

        200.200.200.200.201   l1.local     l1
        200.200.200.200.202   l2.local     l2

          After start cluster still sends packet with source 200.200.200.x

ver.2
          192.168.10.10     l1-p
          192.168.10.11     l2-p

           200.200.200.200.201   l1
           200.200.200.200.202   l2


          Starting cman ... failed
         cman not started: Can't determine address family of nodename
/usr/sbin/cman_tool
          : aisexec daemon didn't start



What I'm doing wrong ?
daro



PS. Cluster configuration:

/etc/cluster/cluster.conf

<?xml version="1.0"?>
<cluster alias="OBN_HA" config_version="28" name="OBN_HA">
    <fence_daemon clean_start="0" post_fail_delay="0" post_join_delay="3"/>
    <clusternodes>
        <clusternode name="l2.local" nodeid="1" votes="1">
            <fence>
                <method name="1">
                    <device name="l2_fence" nodename="l2.local"/>
                </method>
            </fence>
        </clusternode>
        <clusternode name="l1.local" nodeid="2" votes="1">
            <fence>
                <method name="1">
                    <device name="l1_fence" nodename="l1.local"/>
                </method>
            </fence>
        </clusternode>
    </clusternodes>
    <cman expected_votes="1" two_node="1"/>
    <fencedevices>
        <fencedevice agent="fence_manual" name="l1_fence"/>
        <fencedevice agent="fence_manual" name="l2_fence"/>
    </fencedevices>
    <rm>
        <failoverdomains>
            <failoverdomain name="OBN" ordered="1" restricted="1">
                <failoverdomainnode name="l1.local" priority="1"/>
                <failoverdomainnode name="l2.local" priority="2"/>

            </failoverdomain>
        </failoverdomains>
                <resources>
                        <script file="/etc/init.d/vsftpd" name="vsftpd"/>
                        <ip address="200.200.200.203" monitor_link="1"/>
                </resources>
                <service autostart="1" domain="OBN" name="OBN_vsftpd"
recovery="relocate">
                        <ip ref="200.200.200.203">
                                <script ref="vsftpd"/>
                        </ip>
                </service>

    </rm>
</cluster>