[Nagiosplug-devel] check_ntp problem with locan NTP server

[Subhendu Ghosh]

On Sat, 22 Feb 2003, Simon L. Nielsen wrote:

> 
> Hello
> 
> I use check_ntp to monitor the NTP deamon running on the same server as
> Nagios to make sure the NTP daemon is alive. The problem is that since
> the NTP daemon is running on localhost the time is some times exactly
> the same and check_ntp thinks there is a problem and sets a critical
> error :
> 
> [02-22-2003 14:09:47] SERVICE ALERT: trillian.nitro.dk;NTP;CRITICAL;SOFT;1;CRITICAL: server 192.168.1.6, stratum 2, offset -0.000000, delay 0.02573
> 
> I don't really know what the correct fix is since I suppose the check
> for offset = 0 is there for a reason. Do anybody have an idea for a
> proper way to handle this ?
> 
> 

I don't think anybody had planned for check_ntp to be used on a localhost 
:)

Since check_ntp actually does 2 different test - ntpdate and nptq, perhaps 
a switch could be added to turn one of them off.

ntpdate could be used against remote NTP or SNTP hosts to check for 
relative synchronization.

ntpq could be used to check that NTP was running and was synchronized to a 
stratum (not 16) peer.  (no relative synchronization check)

Since this would be an enhancement, it would be added afetr the 1.3 
release.

Intermediate step would be run check_proc to see if ntpd was running and 
possbily comment out the ntpdate section where offset==0 is checked

-- 

-sg