We use Nagios 3.0.6 system on CentOS 5 to monitor around 200 host servers with SNMP, with approx. 6 services per server and are using RRD and PNP to graph the results.
Nagios has twice in the past week has detected a system time change of:
0d 2h 49m
0d 3 46m
The message in /usr/local/nagios/var/nagios.log reads:
Warning: A system time change of 0d 3h 46m 53s (forwards in time) has been detected. Compensating...
All service checks thereafter are orphaned for about the same amount of time indicated in the log which causes gaps in the graphs.
We are running NTP to sync time and other than minor adjustments (1/100 of a second change) the system time appears to be stable.
Any insight to what could be causing this behavior and how to fix?
Thanks in advance.