lbaas Unable to retrieve ready devices

can u try to increase the rpc_response_timeout in neutron.conf and have a try then?

how many compute nodes in your system? how many neutron agents in all? This should not happen in light load system.

to reduce the rpc message, try to increase report_interval and agent_down_time.

Thanks very much!
after I update the system, I can't reproduce this error now,but I meet another question which may be ubuntu's issue
https://bugs.launchpad.net/ubuntu/+source/neutron/+bug/1240712

I get this error and related trace messages each time I deploy Havana with LBaaS enabled.

2013-12-05 16:57:10.243 16238 ERROR neutron.services.loadbalancer.drivers.haproxy.agent_manager [-] Unable to retrieve ready devices
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager Traceback (most recent call last):
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager File "/usr/lib/python2.6/site-packages/neutron/services/loadbalancer/drivers/haproxy/agent_manager.py", line 212, in sync_state
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager ready_logical_devices = set(self.plugin_rpc.get_ready_devices())
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager File "/usr/lib/python2.6/site-packages/neutron/services/loadbalancer/drivers/haproxy/agent_api.py", line 36, in get_ready_devices
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager topic=self.topic
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager File "/usr/lib/python2.6/site-packages/neutron/openstack/common/rpc/proxy.py", line 130, in call
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager exc.info, real_topic, msg.get('method'))
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager Timeout: Timeout while waiting on RPC response - topic: "q-lbaas-process-on-host", RPC method: "get_ready_devices" info: "<unknown>"
2013-12-05 16:57:10.243 16238 TRACE neutron.services.loadbalancer.drivers.haproxy.agent_manager

This appears to be the exact same issue as the reporter experience. If I restart the neutron-lbaas-agent I no longer see this error.

Could you provide neutron-server logs during this error is seen on the agent?

http://paste.openstack.org/show/54595/

Notice that the lbaas log reports the error right around the time the neutron server connects to the RPC server.

It mostly caused by configuration problems.

Pls check the conifg items below:

service_provider in /etc/neutron/neutron_lbaas.conf
device_driver in /etc/neutron/lbaas_agent.ini
service_plugins in /etc/neutron/neutron.conf

The config would look like below if use v1 plugin:
service_provider = LOADBALANCER:Haproxy:neutron_lbaas.services.loadbalancer.drivers.haproxy.plugin_driver.HaproxyOnHostPluginDriver:default
device_driver = neutron.services.loadbalancer.drivers.haproxy.namespace_driver.HaproxyNSDriver
service_plugins = lbaas

and use v2 plugin:
service_provider = LOADBALANCERV2:Haproxy:neutron_lbaas.drivers.haproxy.plugin_driver.HaproxyOnHostPluginDriver:default
device_driver = neutron_lbaas.drivers.haproxy.namespace_driver.HaproxyNSDriver
service_plugins = lbaasv2(you may need to config the lbaasv2 path in neutron's entry_points.txt)