[Pacemaker] Help with configuring pacemaker automatically with chef

[Todd Nine]

Hi guys,
  I'm having a hard time finding the info I need to configure pacemaker
from an input file.  I've been using Zookeeper a lot in our application
tier, so I'm familiar with clusters, however I'm struggling to adapt
that knowledge to the pacemaker configuration.


Here is an overview of our cluster.

OS: Ubuntu 10.10 Server 64 bit
Environment: Amazon EC2
Pacemaker Stack: heartbeat using unicast and pacemaker from the Ubuntu
package
1 Online node, 2 Standby nodes.  All nodes are true peers, so there is
no weight should be given to any nodes.


I have heartbeat up and running.  This is the output from my crm_mon -1
command.


============
Last updated: Thu Jan 13 22:00:55 2011
Stack: Heartbeat
Current DC: ip-10-160-58-200 (79170a75-1b64-43c8-80a1-ff4f490149e5) -
partition with quorum
Version: 1.0.8-042548a451fce8400660f6031f4da6f0223dd5dd
3 Nodes configured, unknown expected votes
0 Resources configured.
============

Online: [ ip-10-160-138-192 ip-10-160-233-27 ip-10-160-58-200 ]



Now I need to perform the following.

Only 1 node is running these services at a time, the other two do not
have these services running.  I need to check our public IP, which can
be configured from chef.


On Start
start haproxy
start stunnel
execute custom script to switch our amazon ec2 ip to the node running

on stop
stop haproxy
stop stunnel



Another caveat.  I've read the cluster_from_scratch several times.  It
uses the interactive command line to configure pacemaker.  I need to
create an xml file directly using my chef recipes.  I have no ability to
know each node's UUID, only the hostname.  Can anyone provide me with an
example of how I can configure pacemaker from an xml file to execute
several scripts?


Thanks,
Todd




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.clusterlabs.org/pipermail/pacemaker/attachments/20110114/f7c9128d/attachment.html>