Can't find local node name in cluster.conf
 

Well I'm a newbie in clustering and I'm trying to set a 1 node cluster and I can't start cman, i get:
cman not started: Can't find local node name in cluster.conf /usr/sbin/cman_tool: aisexec daemon didn't start

my cluster.conf is:

?xml version="1.0"?>
<cluster name="kluster" config_version="1">

<cman expected_votes="1">
</cman>

<nodes>
<node name="server.gfs" votes="1">
<fence>
<method name="single">
<device name="human" ipaddr="192.168.1.53"/>
</method>
</fence>
</node>

</nodes>

<fence_devices>
<device name="human" agent="fence_manual"/>
</fence_devices>

</cluster>

I dissabled selinux and already check the permisions in /temp folder (777), can anyone hel me please???

Is your cluster node server.gfs resolvable? Best way is to add the id and name to your /etc/hosts. I suppose you should add a line like as follows:

192.168.1.53 server.gfs


Then this should work.

Thanks elcody02, I forgot to mencione that server.gfs is resolvable. In /etc/hosts I have:

# Do not remove the following line, or various programs
# that require network functionality will fail.
192.168.1.53 server.gfs server localhost.localdomain localhost
192.168.1.21 athlon1
127.0.0.1 localhost.localdomain localhost
::1 localhost6.localdomain6 localhost6


And I don't have a clue why i can't start cman

Ok. So next is to have a look at what syslog tells you when cman is started. Also the ccsd messages are important. Check if ccsd is running and what it tells you about quorum.

I'm not sure, but I remember having troubles with using nodenames with . in the name. Try to not use server.gfs. Instead use servergfs or something the like. But thats only a wild guess. I'm really not sure about that.

Have fun.

Thanks elcody02

I've tried changing the cluster name from server.gfs to servergfs and I also added "192.168.1.53 servergfs" to /etc/hosts file. But i keep geting the same response.

In /var/log/messages I don't have any line about cman, only lines of ccsd:
Nov 22 13:12:43 server ccsd[14345]: Starting ccsd 2.0.60:
Nov 22 13:12:43 server ccsd[14345]: Built: Mar 14 2007 08:17:44
Nov 22 13:12:43 server ccsd[14345]: Copyright (C) Red Hat, Inc. 2004 All rights reserved.
Nov 22 13:12:43 server ccsd[14345]: cluster.conf (cluster name = kluster, version = 1) found.
Nov 22 13:13:13 server ccsd[14345]: Unable to connect to cluster infrastructure after 30 seconds.
Nov 22 13:13:43 server ccsd[14345]: Unable to connect to cluster infrastructure after 60 seconds.
Nov 22 13:14:13 server ccsd[14345]: Unable to connect to cluster infrastructure after 90 seconds.


Any idea?

I used system-config-cluster to made a new cluster.conf file:
<?xml version="1.0" ?>
<cluster alias="kluster" config_version="3" name="kluster">

<fence_daemon post_fail_delay="0" post_join_delay="3"/>
<clusternodes>

<clusternode name="servergfs" nodeid="1" votes="1">

<fence/>

</clusternode>

</clusternodes>
<cman/>
<fencedevices>

<fencedevice agent="fence_manual" name="Yoyosaurio"/>

</fencedevices>
<rm>

<failoverdomains/>
<resources/>

</rm>

</cluster>


And it seems to work much better than the old cluster.conf but I still can't start cman, the last lines of /var/log/messages goes like this:
Nov 23 11:56:48 server openais[18649]: [SERV ] Initialising service handler 'openais CMAN membership service 2.01'
Nov 23 11:56:48 server openais[18649]: [CMAN ] CMAN 2.0.60 (built Mar 14 2007 08:17:48) started
Nov 23 11:56:48 server openais[18649]: [SYNC ] Not using a virtual synchrony filter.
Nov 23 11:56:48 server openais[18649]: [IPC ] ERROR: Could not bind AF_UNIX: Address already in use.
Nov 23 11:56:48 server openais[18649]: [MAIN ] AIS Executive exiting (-7).
Nov 23 11:56:48 server ccsd[18638]: Initial status:: Quorate
Nov 23 11:57:14 server ccsd[18638]: Unable to connect to cluster infrastructure after 30 seconds.

I think that the problem is someting to do with the address, but I'm clueless of how to solve it

Hm, everything looks quite ok so far. What you'll need is the expected_votes="1" in cman.

But that didn't work before either. I haven't seen this before.
Hm, did you try to stop everything:

Wild guess:
You could also try to setup a two node cluster by adding a second node to the cluster.conf. Where no second node is available. Perhaps that is a problem. Check for cman being something like this then:
Hope you'll get a step further.

Solved
 

Well, I disabled a few daemons and also the firewall and it work pretty fine, thanks elcody02 for all your help. I don't know what configuration should I add to the iptables config, but that's a topic for a another post.

Great to hear that everything is working now.

For iptables config you might want to have a look at the following URL:
http://www.open-sharedroot.org/faq/a...uster-software

I'm curious: which daemons did you disable, and how did you disable the firewall? Because I'm having the same problem.

I used iptables -F on my firewall and then ran the script at http://www.open-sharedroot.org/faq/a...uster-software (for RHEL 5).

RedHat Cluster Suite config
 

Hi, I'm trying to start ccsd service, but is showing, service is already running. If I try stop, it is showing, it is showing Unrecognized service

What version of Red Hat an cluster are you using and what exactly is the output of the commands you issue?

Hi Elcody02, I'm using RHEL 5.2

# service ccsd start
unrecognized service

The service cman does the whole job.

It's starting ccsd, aisexec, cman, joins the fencedomain starts qdisk (if needed).

So try to start cman and watch the logs very careful.

Roboam,

Could you list the deamons you disabled?

Thanks,

Joe.