1.
Ran PowerCLI scripts against all hosts to report on NTP configuration
a.
Found some misconfigured and fixed them.
b.
Time-Sync improved initially, but soon drifted again.
2.
Choosing one cluster, I tried setting the NTP servers to particular servers, IP etc. Here are the results:
a.
Setting to IP address vs. FQDN found time synced reliably.
b.
Setting NTP server to incorrect server caused time to drift almost immediately & it stayed drifted.
c.
Stopping NTP service also caused time to drift & remain drifted.
d.
Setting the NTP server to aliases or FQDN of NTP servers produced mixed results.