SES5: "time_init: ntp" does not work as expected

[smithfarm]

Description of Issue/Question

When using DeepSea to deploy a Ceph cluster on SLE-12-SP3 (SES5), Stage 3 fails with:

admin: Failures summary:
    admin: ceph.time (/srv/salt/ceph/time):
    admin:   node3.ses5.com:
    admin:     start ntp: The named service ntpd is not available
    admin:   node1.ses5.com:
    admin:     start ntp: The named service ntpd is not available
    admin:   node2.ses5.com:
    admin:     start ntp: The named service ntpd is not available
Command '['vagrant', 'up']' failed: ret=1 stderr:

This happens when:

• the admin node ("admin.ses5.com" in this case) has NTP installed and ntpd.service running
• the NTP server on the admin node is configured to sync against external NTP servers
• there is no ntp package installed on any of the cluster nodes
• DeepSea has time_init: ntp in the pillar.

Analysis

This failure is apparently caused by a bug in Salt 2016.11.4 - the YAML in question is:

start ntp:
  service.running:
    - name: ntpd
    - enable: True

My expectation is that this should cause the ntpd.service to be enabled and started. But that doesn't happen.

[smithfarm]

1823 was merged