[Pool] Off by a second leap second failures

Martin Burnicki martin.burnicki at burnicki.net
Sun Jul 12 08:25:08 UTC 2015


John Winters wrote:
> 
> 
> On 12/07/15 08:34, Martin Burnicki wrote:
> [snip]
>>
>> what's the output of "ntpq -c rv"?
> 
> john at pisquared:~$ ntpq -c rv
> associd=0 status=0615 leap_none, sync_ntp, 1 event, clock_sync,
> version="ntpd 4.2.6p5 at 1.2349-o Fri Apr 10 19:31:04 UTC 2015 (1)",
> processor="armv7l", system="Linux/3.18.0-trunk-rpi2", leap=00, stratum=3,
> precision=-20, rootdelay=30.824, rootdisp=66.543, refid=130.159.196.117,
> reftime=d94c9aa7.6f2887ec  Sun, Jul 12 2015  8:49:59.434,
> clock=d94c9ae5.ad881b9f  Sun, Jul 12 2015  8:51:01.677, peer=64427, tc=8,
> mintc=3, offset=0.626, frequency=-7.725, sys_jitter=0.294,
> clk_jitter=0.341, clk_wander=0.040, tai=36, leapsec=201507010000,
> expire=201512280000
> john at pisquared:~$ ls -l /etc/leap-seconds.3629404800
> -rw-r----- 1 root root 10386 Apr 14 17:15 /etc/leap-seconds.3629404800
> john at pisquared:~$
> 
>> If you have a *current* leap second file, and you have configured ntpd
>> correctly to use it, the output should contain something like:
>>
>> leapsec=201507010000, expire=201512280000
>>
>> If these terms are missing then the leap second file has not been
>> evaluated by ntpd.
> 
> So as they're present, the file *has* been evaluated by ntpd?

Yes, as far a I know. Otherwise ntpd would not have an idea of the leap
second date, and when the available information expires.

> If so, then the plot thickens.

Indeed.

Are you running ntpd with or without -x?

If without -x then ntpd should have passed the leap second warning to
the kernel, and there should be a syslog message from the kernel at UTC
midnight at the end of 2015-06-30 telling that the kernel has inserted a
leap second.

Do you still have those syslog files and can check if this is the case?

Martin



More information about the pool mailing list