Final on - Re: A tale of two nameservers - resolution problems

[Robert Moskowitz]

I have learned a bit in the past time period.

First, adding a 2nd DNS server, in this case my ISP's, got around this 
timing problem, becuase the 'right fix' is not available yet.

The answer is to stop using old ntpd and use modern chronyd.  But more 
than that, ver 2.1.1 which is available in Fedora 22, but not (yet) in 
Centos 7 (perhaps 7.2).  You start chrony with the -s option; this tells 
it you have a bad, or no rtc, and to regularly save the time to a file 
which is read when chrony starts up.  I have tested this out on F22 and 
it works as advertised.

There is also systemd-timesync, but Fedora/redhat went the chrony route, 
and I got more help figuring it out.

On to the next fun challenge.

On 09/01/2015 12:16 PM, Sam Wilson wrote:
> In article <mailman.2626.1441122408.26362.bind-users at lists.isc.org>,
>   Robert Moskowitz <rgm at htt-consult.com> wrote:
>
>> I will be looking more into this.  Obvious when you get ones nose
>> dragged into time wrong on boot.  This is actually a broader problem on
>> arm SoC booting.  Your logs all have the wrong time for the boot
>> messages until there is a network to get time.  I have some ideas for a
>> process that will set time a boot to the time of the last poweroff.  at
>> least that is 'close enough' for starters.
> I believe that's the solution Apple adopted for the AppleTV, which has
> no rtc and couldn't use a certificate to connect to a wireless network
> because the certificate wasn't valid in 1970.
>
> Sam
>