Symptoms:
Two or more database servers (local or cloud-based) report different SQL clock times.
Impact: High
Recommended action :
Examine the computers that are running at different times, and resolve any issues.
Background
When two or more computers are running replicating databases, their clocks must be synchronized properly, otherwise, important SQL functions such as rollbacks, saves, etc may fail and data may be corrupted or lost.
In many cases, server time sync issues stem from the Kerberos protocol, which has a security feature that looks at the time stamp on Kerberos tickets in an effort to prevent them from being reused. A ticket is rejected if it is more than five minutes old. Therefore, if the clocks fall out of sync by more than five minutes, Kerberos will begin to break down.
Normally, time synchronization doesn’t pose a problem where the same NTP is used across the whole range, but in mixed local/cloud environments, this may not be possible. Where some production servers are virtualized or cloud-based, clock synchronization can become an issue, also where multiple Active Directory forests exist.