* ceph by vote - but will accept multiple options/
HA next steps:

We have working ground-> overcloud now. But it's not HA! Lets get that done. 
 
In particular we need to identify all the things to HA and talk through single node + 2 node + 3+ node setups of them.

OpenStack HA guide:
http://docs.openstack.org/high-availability-guide/content/ch-intro.html

Overcloud Services Needing HA:
	* rabbit
		* native clustering mode (http://www.rabbitmq.com/clustering.html)
		*  activation within the code (rabbit_ha_queues flag) (http://www.rabbitmq.com/ha.html#behaviour) mirrored queues
	* qpid
		* http://openstack.redhat.com/Highly_Available_Qpid_for_OpenStack
		* http://qpid.apache.org/books/0.7/AMQP-Messaging-Broker-CPP-Book/html/ch01s08.html
		* corosync
	* db
		* Xtradb cluster
			* percona committing to having packages for fedeora/ubuntu/suse/centos
	* api instances
		* virtual IP - keepalived + haproxy
	* glance store
		* swift - mature, part of openstack - pending confirmation of 
			* 1 server works
			* 2 servers also 
		* or first patch that comes in
	* heat-engine
		* not stateless - failures will interrupt stack /create/update/delete
		* run 2 of 
	* neutron
		* dhcp - set number scheduled to some N.
			* what about stale data? - file a bug: https://bugs.launchpad.net/neutron/+bug/1248048
		* L2agent - per hypervisor - localised failures
		* L3agent - spof - no distributed router
			* routers - detect failure and reschedule networks to other routers
				* ideally neutron does this
		* metadata agent
			* stateless
			* check it runs on the network node
			* if runs on network node - need to schedule to new places as needed
	* nova-compute
		* nothing to do.
	* cinder-volume
		* e.g. ceph/glusterfs/3par ?
		* cinder bug - requests go to the originating cinder-volume even when the real data is available
	* horizon
		* N-node memcache ?
		* Run N copies.
		* virtual IP + haproxy
	* nova scheduler
		* run N copies
	* nova conductor
		* run N copies
		* Or just eliminate the conductor (use_local = True)

General solution topics:
    -- load balancer strategy
    -- state/data preservation/migration
    -- load balancer as a service


Actions