On Thu, Apr 28, 2016 at 6:48 PM, Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> wrote:

I'm running a NRPE monitor that monitors openLDAP. To compute delta values, the NRPE plugin stores the "dn: cn=Uptime,cn=Time,cn=Monitor" reported by openLDAP to compare it with the current value (to build the time difference). If the time difference is negative, the plugin reports a "server restart" event.

So far, so good...

However, very rarely (once every few weeks) a false server restart is reported. Viewing the historic logs, I can only assume that openLDAP reports a bad uptime value now and then. Could this be?

If uptime is a calculation based on the system clock, is there some process such as ntpd or vmware tools, periodically updating the system clock, sometimes backwards ?

If so, maybe look for an option to limit backwards updates.