[Linux-HA] How to get HA2 to recognize the other node?

[Dejan Muhamedagic]

On Thu, Sep 07, 2006 at 06:59:57PM +0530, Oza Dhairesh wrote:
> Hi,
> 
>   I'm trying the most simplistic of setups. Two nodes running HA. One
> is a 64 bit node and the other a 32 bit node. After starting heartbeat
> on both the nodes, each node comes up as a cluster of a single node. I
> have deleted all the /var/lib/heartbeat files (crm/*, hb_* etc) and
> restarted heartbeat on both m/cs but still I can't get them to realize
> that they are part of the same cluster. The ha-debug on both files has
> just these line pertaining the the other node:
> 
> heartbeat[3754]: 2006/09/07_16:54:58 WARN: node other-node: is dead   &
> 
> crmd[3768]: 2006/09/07_16:55:00 WARN: get_uuid:utils.c Could not
> calculate UUID for other-node
> crmd[3768]: 2006/09/07_16:55:00 WARN: populate_cib_nodes:control.c
> Node other-node: no uuid found
> 
> The heartbeat package is heartbeat-2.0.5-7.10.
> 
> crm_mon output is as follows:
> 
> ============
> Last updated: Thu Sep  7 18:25:49 2006
> Current DC: sribhu (de70b532-5e8e-4d98-8e35-c9eb9aa23fbc)
> 1 Nodes configured.
> 0 Resources configured.
> ============
> 
> Node: node-32 (de70b532-5e8e-4d98-8e35-c9eb9aa23fbc): online
> 
> On the other node it's the other way round.
> 
> I am able to ping the other node form each of the node and firewall is
> disabled.
> 
> This setup was working a few days back but now I just can't get it up
> after having heartbeat shutdown for a couple of weeks. Any pointers on
> how to debug this?
> 
> Log attached.

Did any of the hostnames changed? Does uname produce the same
result as what you have specified in node options in ha.cf?

Cheers,

Dejan