[Linux-cluster] cluster-3.0.0 problems starting up

[Fabio M. Di Nitto]

On Fri, 2009-08-14 at 15:14 -0700, Trevor Peirce wrote:
> Hello,
> 
> I've attempted to compile and install the cluster-3.0.0 suite and have 
> these problems when trying to start cman:
> 
> Aug 14 12:10:01 localhost corosync[27739]: parse error in config: parse 
> error in config: .
> Aug 14 12:10:01 localhost corosync[27739]: [MAIN ] Corosync Cluster 
> Engine ('trunk'): started and ready to provide service.
> Aug 14 12:10:01 localhost corosync[27739]: [MAIN ] Successfully read 
> config from /etc/cluster/cluster.conf
> Aug 14 12:10:01 localhost corosync[27739]: [MAIN ] Successfully parsed 
> cman config
> Aug 14 12:10:01 localhost corosync[27739]: [MAIN ] Successfully 
> configured openais services to load
> Aug 14 12:10:01 localhost corosync[27739]: [MAIN ] parse error in 
> config: parse error in config: .
> Aug 14 12:10:01 localhost corosync[27739]: [MAIN ] Corosync Cluster 
> Engine exiting with status -9 at main.c:879.
> 
> Here is cluster.conf, as generated by ccs_tool:
> 
> <?xml version="1.0"?>
> <cluster name="TestCluster" config_version="3">
> <clusternodes><clusternode name="guinneapig" votes="1" 
> nodeid="1"><fence><method name="single"><device 
> name="manual"/></method></fence></clusternode></clusternodes>
> <fencedevices><fencedevice name="manual" agent="manual"/></fencedevices>
> <rm>
> <failoverdomains/>
> <resources/>
> </rm>
> </cluster>

Check your hostname setting.

The name you use for clusternode _MUST_ resolve (either via DNS
or /etc/hosts) to the ip address of an ethernet interface. That hostname
set to localhost is suspicious.

Fabio