In case when incorrect IP is set for switch:


neutron-ml2-ansible.yaml:
resource_registry:
  OS::TripleO::Services::NeutronCorePlugin: OS::TripleO::Services::NeutronCorePluginML2Ansible
parameter_defaults:
  NeutronMechanismDrivers: openvswitch,ansible
  NeutronTypeDrivers: local,vxlan,vlan,flat
  NeutronNetworkType: vlan
  ML2HostConfigs:
    switch1:
      ansible_network_os: junos
      ansible_host: 10.9.95.26 #Not existing IP
      ansible_user: ansible
      ansible_ssh_pass: N3tAutomation!
      #manage_vlans: false
                  
There are two issues:

1 - Loop
Error message is logged ~every 2-3 seconds into Neutron server.log:

For example:
2018-11-01 07:20:40.866 34 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
[root@overcloud-controller-0 heat-admin]# cat /var/log/containers/neutron/server.log | grep 'Name or service not known' | grep switch1 | grep 2018
2018-11-01 06:55:29.903 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:32.359 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:34.744 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:37.545 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:39.930 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:42.275 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:44.665 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:47.126 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:49.641 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:52.005 33 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:53.426 29 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:56.054 29 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}
2018-11-01 06:55:58.478 29 ERROR neutron.plugins.ml2.managers  fatal: [switch1]: FAILED! => {"msg": "[Errno -2] Name or service not known"}

As far as I know "retry mechanism" in such cases is usually started from some low value and then getting increased per failure until Final/Fatal decision is made.

2- more meaningful message
Actual message is: "Name or service not known"

I think we should improve experience for operators. One
thing is to look at the loop why we retry so many times and other thing
is to provide more meaningful message.