We could tweak haproxy as Mike Bayer suggests here - http://lists.openstack.org/pipermail/openstack-dev/2016-April/093562.html - to make sure that we do not start routing connections to the original active node when goes back online.
It's not really a Nova or Neutron problem - all services are potentially affected here after a failover of an active MySQL node.
Bogdan, Puppet team, what do you think?
We could tweak haproxy as Mike Bayer suggests here - http:// lists.openstack .org/pipermail/ openstack- dev/2016- April/093562. html - to make sure that we do not start routing connections to the original active node when goes back online.
It's not really a Nova or Neutron problem - all services are potentially affected here after a failover of an active MySQL node.
Bogdan, Puppet team, what do you think?