keystone refresh takes a while

[Chris Ricker]

I see this sporadically during puppet runs:

Info: Class[Keystone::Db::Sync]: Scheduling refresh of Service[keystone]
Debug: Executing '/sbin/status keystone'
Debug: Executing '/sbin/initctl --version'
Debug: Executing '/sbin/status keystone'
Debug: Executing '/sbin/restart keystone'
Notice: /Stage[main]/Keystone/Service[keystone]: Triggered 'refresh' from 1 events
Debug: /Stage[main]/Keystone/Service[keystone]: The container Class[Keystone] will propagate my refresh event
Debug: Prefetching keystone resources for keystone_role
Debug: Executing '/usr/bin/keystone --endpoint http://127.0.0.1:35357/v2.0/ role-list'
Error: /Stage[main]/Keystone::Roles::Admin/Keystone_role[_member_]: Could not evaluate: Execution of '/usr/bin/keystone --endpoint http://127.0.0.1:35357/v2.0/ role-list' returned 1: Unable to establish connection to http://127.0.0.1:35357/v2.0/OS-KSADM/roles

Suspect we need a longer sleep in our keystone refresh to allow the service time to stop and start before we try to use it

[Chris Ricker (chris-ricker)]

https://review.openstack.org/59762 which will need a backport to stable/havana if accepted