This method was intended to check state of the HA router on the
node and update it in the neutron server.
Patch [1] added check of the initial status to the
neutron_keepalived_state_change_monitor process.
It also could cause some race conditions and event which is setting
correct state of the router will be not processed thus router may endup
with two nodes with "primary" state in the Neutron's DB.
Neutron_keepalived_state_change_monitor was notifying agent about
router's initial state only if this state was 'primary'.
Now it will notify agent always to let agent set router's state as
'backup' if needed (that was previously done by this removed
update_initial_state() method).
[1] https://review.opendev.org/c/openstack/neutron/+/642295
Conflicts:
neutron/agent/l3/ha_router.py
neutron/agent/l3/keepalived_state_change.py
neutron/tests/unit/agent/l3/test_dvr_local_router.py
Change-Id: I2cc58c30cf844ee0ecf0611ecdec430086464790
Closes-Bug: #1916022
(cherry picked from commit 0d8ae15767)
3b2b7f4fe7