Re: Problems with ntp
>Well, there is chrony, which is feature complete with ntp as far as I
Except that it is not. NTP is hard, and attempts to make it simpler
probably are not "feature-complete".
> Even without the licensing issues, I would recommend chrony
>anyway: chrony converges to the time servers much faster than ntp. I
>have long suspected that there is something wrong with the ntp
>phase-lock loop algorithm (the clock is rate-adjusted to somewhere close
>to the time server after some hours, and then diverges again ...), but
>email exchange with Dr. Mills reaches nowhere. He is convinced that he
>is right, but I am convinced that there is something wrong after I look
>closely into the logs.
Him being one of the world's experts on time metrology makes me think
that you are wrong. If you want to get more opinions, ask about it on