check_ntp_time

[djberriman]

Hi,

Appears the latest stable version of check_ntp_time has an issue with the upcoming leap second, all my servers running this version of check_ntp_time has as of a few hours ago started reporting as follows:-

sending request to peer 0
response from peer 0: offset 0.0008511543274
sending request to peer 0
response from peer 0: offset 0.0007092952728
sending request to peer 0
response from peer 0: offset 0.0007601976395
sending request to peer 0
response from peer 0: offset 0.0007613897324
discarding peer id 0: flags=1
overall average offset: 0
NTP CRITICAL: Offset unknown

It seems related to this in the system

Dec 31 02:18:10 ntpd[1356]: kernel reports leap second insertion scheduled

Any ideas?

[dwhitfield]

I am by no means an expert, but I was told "NTP will slow down our clocks a bit until we have compensated for that second" -- reference at https://datamost.com/clacke/comment/YrS ... xI4oS8_4IQ so you can judge the validity of the statement in context.

Did you have any checks that had problems over the weekend? We observed New Years Day yesterday, so we are just getting to the forums this morning.

[djberriman]

I believe the slow down was only if you were using google smearing, as standard a leap second is inserted by the clock registering 23:59:60 which did at one time causes some software to crash hence google coming up with smearing.

The problem here is that ntp ran fine and reported no issues however the check failed with the above error but only servers with the latest source build were affected. It was reported for approximately 24 hours (12 hours either side).

It therefore appears to be a bug which will happen every time there is a leap second once the kernel detects a leap second will be inserted.

Either way check_ntp_time should not incorrectly report an issue with ntp.

[rkennedy]

Can you please post the exact command definition, and service definition you're using? This will help us to try and recreate it in house.

More than likely a case will need to be created here - https://github.com/nagios-plugins/nagios-plugins/issues - but, I can try to see if it's anything obvious pointing out prior to that.