Comment 2 for bug 1689400

I see that the contrail-api is running independently triggered by the /etc/systemd/system/contrail-api.service(systemd)

root@a5s10(controller):/var/log/contrail# ps -ef | grep 29464
root 6946 29614 0 13:17 ? 00:00:00 grep --color=auto 29464
contrail 29464 1 0 May08 ? 00:01:18 /usr/bin/python /usr/bin/contrail-api --conf_file /etc/contrail/contrail-api.conf --conf_file /etc/contrail/contrail-keystone-auth.conf --conf_file /etc/contrail/contrail-database.conf
root@a5s10(controller):/var/log/contrail#

In Redhat systems, all the contrail services are under supervisor,
root@a5s10(controller):/var/log/contrail# supervisorctl -s unix:///var/run/supervisord_config.sock status
contrail-api:0 EXITED May 08 09:12 PM
contrail-config-nodemgr RUNNING pid 4521, uptime 16:29:44
contrail-device-manager RUNNING pid 4523, uptime 16:29:44
contrail-schema RUNNING pid 4524, uptime 16:29:44
contrail-svc-monitor RUNNING pid 6263, uptime 16:29:15
root@a5s10(controller):/var/log/contrail#

I see that the contrail-api.service(systemd) file is brought in to the system by contrail-ansible-internal
https://github.com/Juniper/contrail-ansible-internal/blob/master/playbooks/roles/contrail/config/files/systemd/contrail-api.service
https://github.com/Juniper/contrail-ansible-internal/blob/master/playbooks/roles/contrail/config/tasks/setup.yml#L19

I would let Narendra to debug further.