[Linux-cluster] Updating cluster.conf

[isplist at logicore.net]

Thanks for the input.

> You could scp the cluster.conf to the member machines.  Make sure you are
> updating the version number in your cluster.conf.

Thing is, I can't run any cman commands which include updating the conf file. 
Since I cannot start the cluster, I cannot run any cluster commands. Isn't 
that part of my problem, needing to figure out why/if cman is running?

It looks like cman won't start and because of that, fenced won't start either. 
There are plenty of errors in the logs which I've posted many times. I'm no 
longer sure what I'm looking for.

> cluster.conf, you could then use system-config-cluster to update your

system-config-cluster has never worked on any of my machines using ANY version 
of RH. I've posted about that but have not found any answers.

> cluster configuration which will propogate the new file for you.  Make
> sure you inspect /var/log/messages .. it can be very helpful.  Make sure
> that no new iptables firewall rules could be getting in your way.

Well, here's another paste of what's going on now...

First, how I'm starting my cluster, then some logging then my cluster.conf;

depmod -a
modprobe dm-mod
modprobe gfs
modprobe lock_dlm
ccsd
cman_tool join -t 120 -w
fence_tool join -t 120 -w
clvmd
vgchange -aly
mount -t gfs /dev/vgcomp/rimfire /var/www/

Then a snip of my logging;

Oct 20 09:06:03 cweb92 kernel: CMAN: Waiting to join or form a Linux-cluster
Oct 20 09:06:04 cweb92 kernel: CMAN: sending membership request
Oct 20 09:06:04 cweb92 kernel: CMAN: Cluster membership rejected
Oct 20 09:06:04 cweb92 kernel: CMAN: Waiting to join or form a Linux-cluster
Oct 20 09:06:04 cweb92 ccsd[4788]: Connected to cluster infrastruture via: 
CMAN/SM Plugin v1.1.7.1
Oct 20 09:06:04 cweb92 ccsd[4788]: Initial status:: Inquorate
Oct 20 09:06:08 cweb92 kernel: CMAN: sending membership request
Oct 20 09:06:08 cweb92 kernel: CMAN: Cluster membership rejected
Oct 20 09:06:08 cweb92 ccsd[4788]: Cluster manager shutdown.  Attemping to 
reconnect...
Oct 20 09:06:23 cweb92 clvmd: Can't open cluster manager socket: Network is 
down
Oct 20 09:06:23 cweb92 ccsd[4788]: Remote copy of cluster.conf is from quorate 
node.
Oct 20 09:06:23 cweb92 ccsd[4788]:  Local version # : 42
Oct 20 09:06:23 cweb92 ccsd[4788]:  Remote version #: 42
Oct 20 09:06:23 cweb92 ccsd[4788]: Remote copy of cluster.conf is from quorate 
node.
Oct 20 09:06:23 cweb92 ccsd[4788]:  Local version # : 42
Oct 20 09:06:23 cweb92 ccsd[4788]:  Remote version #: 42
Oct 20 09:06:27 cweb92 ccsd[4788]: Unable to connect to cluster infrastructure 
after 30 seconds.
Oct 20 09:06:57 cweb92 ccsd[4788]: Unable to connect to cluster infrastructure 
after 60 seconds.

Cluster.conf;

<?xml version="1.0"?>
<cluster config_version="46" name="vgcomp">
<fence_daemon clean_start="1" post_fail_delay="0" post_join_delay="3"/>
    <clusternodes>
        <clusternode name="cweb92" votes="1">
                <fence>
                    <method name="1">
                        <device name="fence_mcdata" port="16"/>
                    </method>
                </fence>
        </clusternode>
        <clusternode name="cweb93" votes="1">
                <fence>
                    <method name="1">
                        <device name="fence_mcdata" port="17"/>
                    </method>
                </fence>
        </clusternode>
        <clusternode name="cweb94" votes="1">
                <fence>
                    <method name="1">
                        <device name="fence_mcdata" port="18"/>
                    </method>
                </fence>
        </clusternode>
        <clusternode name="qm250" votes="1">
                <fence>
                    <method name="1">
                        <device name="fence_mcdata" port="19"/>
                    </method>
                </fence>
        </clusternode>
        <clusternode name="qm249" votes="1">
                <fence>
                    <method name="1">
                        <device name="fence_mcdata" port="20"/>
                    </method>
                </fence>
        </clusternode>
        <clusternode name="qm248" votes="1">
                <fence>
                    <method name="1">
                        <device name="fence_mcdata" port="21"/>
                    </method>
                </fence>
        </clusternode>
        <clusternode name="qm247" votes="1">
                <fence>
                    <method name="1">
                        <device name="fence_mcdata" port="22"/>
                    </method>
                </fence>
        </clusternode>
        <clusternode name="dev" votes="1">
                <fence>
                    <method name="1">
                        <device name="fence_mcdata" port="23"/>
                    </method>
                </fence>
        </clusternode>
    </clusternodes>
<fencedevices>
    <fencedevice agent="fence_mcdata" ipaddr="x.x.x.x" login="user" 
name="ED5000" passwd="xxxx"/>
</fencedevices>
</cluster>

Mike