Le 30/06/2015 06:42, Majdi S. Abbas a écrit :
> On Mon, Jun 29, 2015 at 11:04:55PM -0500, Jerry Pasker wrote:
>> So by this output, I am guessing that galaga.usshc.com, tick.usshc.com,
>> tock.usshc.com and clock.nyc.he.net are all not doing the leap second
>> correctly? (I guess I have some work to do tomorrow)
> Possibly. Old versions announced the leap the entire month,
> current revisions try to announce 24 hours in advance. But you can
> fix it now, since most clients will check in at least every 17 minutes.
>
>> With 4 machines saying leap=01 and four saying leap=00, what would happen
>> if the above mention hosts are not corrected to leap=01 before leap
>> second o'clock? Is it majority rules? What happens in the above case
>> where it's a tie?
> Provided you're using more recent code, if you have a reference
> clock with leap information, it should override. If you do not, and
> are relying on Internet time, it should be majority rules.
>
> Finally, if you have a leapfile configured, and are not using
> a directly attached reference clock, that should work.
>
> What is your server saying when you query it with 'rv'?
>
> If you are not returning leap=01 you may wish to download
> and configure the leapfile.
>
> --msa
>
On mine I see leap_disarmed and leap=01, which is Ok ?
_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool
