Monday, July 21, 2014

Neutron + Pacemaker for HA Gives error

I was trying to configure HA for neutron server in icehouse implementation. I was able to set up ha for all other services except neutron. I was trying to use pacemaker for setting up HA  by following http://docs.openstack.org/high-availability-guide/content/_add_neutron_l3_agent_resource_to_pacemaker.html

but still i get following error. dhcp agent and metadata agent is showing no error but l3 agent is not working.

output of crm_mon -1
Last updated:FriJul1814:03:252014Last change:FriJul1813:54:042014 via cibadmin on network1 Stack: classic openais (with plugin)Current DC: network2 - partition with quorum Version:1.1.10-14.el6_5.3-368c7262Nodes configured,2 expected votes 4Resources configured

Online:[ network1 network2 ]

p_api-ip (ocf::heartbeat:IPaddr2):Started network2

p_neutron-dhcp-agent (ocf::openstack:neutron-dhcp-agent):Started network1

p_neutron-metadata-agent (ocf::openstack:neutron-metadata-agent):Started network1

Failed actions: p_neutron-l3-agent_start_0 on network2 'unknown error'(1): call=13, status=TimedOut,last-rc-change='Fri Jul 18 04:32:06 2014', queued=20091ms,exec=0ms p_neutron-l3-agent_start_0 on network1 'unknown error'(1): call=23, status=TimedOut,last-rc-change='Fri Jul 18 14:03:01 2014', queued=20010ms,exec=0ms[root@network1 openstack]#

Solution

The neutron-agent-l3 script to blame as it tries to communicate with neutron server
directly on port 9696, while communication is handled by AMQP service
(Qpid in my case). We need to modify the script to use Qpid port and not neutron server one.

No comments:

Post a Comment