On Thu May 11 23:36:53 2017, alessandro.gherardi@yahoo.com wrote:
> Hi Thomas,Thank you for the update.
> I have a question: Since the failover protocol uses 32-bits
> timestamps, it is not y2k38 safe. Correct?

=> half correct: the failover uses unsigned timestamps,
the y2k38 issue is for signed timestamps, e.g. ANSI time_t.

> Does your fix include changing protocol to make it y2k38 safe,
> or will that still be an issue?

=> no, it should break in 21 years...