Thank You Ed!

On 10/22/2017 1:47 PM, King Beowulf wrote:
> On 10/22/2017 12:49 PM, Dave Lien - W7DAL wrote:
>> /... //the install process usually sets it up for you./
>>
>> Thanks Jim and Larry!
>>
>> I've been watching the flow of contacts using FT8 over the past several
>> days without doing anything about time correction. With continuously
>> positive results over many hours of system ON time I now have no reason
>> to believe drift is a factor to be addressed.
>>
>> Also, I've never seen a reference to having to add special time adjust
>> software when Linux is used with this time critical digital
>> communications mode. Such a software addition is pretty much mandatory
>> with Windows and is well publicized by FT8 and other digital mode users.
>>
>> So, I've decided to just leave things alone. . TNX.
>
> Just as an FYI (and late to the party as usual), I have this is my start
> up /etc/rc.d/rc.local, along with the appropriate settings in :
>
> # Set the time
> /usr/sbin/ntpd -qg
> /sbin/hwclock -w
>
> This syncs up both hardware and system clocks at the end of the boot cycle.
>
> The motherboard hardware clock can and does drift a bit over long time
> scales. These clock chips have always been awful.  It does not matter
> the OS you are running.  The system (software) clock is even worse,
> although does not seem to be as bad as in Windows.
>
> http://www.tldp.org/HOWTO/Clock-2.html
>
> Running ntp continuously is a bit overkill for my needs.  Here is my
> time drift on recent reboots for kernel and nvidia driver updates:
>
> # last reboot
> reboot   system boot  4.4.88           Fri Oct 13 00:58 - 13:10 (9+12:11)
> reboot   system boot  4.4.88           Thu Oct 12 15:07 - 00:57  (09:50)
> reboot   system boot  4.4.75           Mon Sep  4 13:32 - 15:05 (38+01:32)
>
> # cat messages.2 |grep ntpd
> Oct 12 15:07:20 gandalf ntpd[1431]: ntpd [email protected] Thu Apr 20
> 17:49:20 UTC 2017 (1): Starting
> Oct 12 15:07:20 gandalf ntpd[1431]: Command line: /usr/sbin/ntpd -qg
> Oct 12 15:07:20 gandalf ntpd[1431]: proto: precision = 0.114 usec (-23)
> ...
> Oct 12 15:07:29 gandalf ntpd[1431]: ntpd: time set +2.169528 s
> Oct 13 00:59:03 gandalf ntpd[1432]: ntpd [email protected] Thu Apr 20
> 17:49:20 UTC 2017 (1): Starting
> Oct 13 00:59:03 gandalf ntpd[1432]: Command line: /usr/sbin/ntpd -qg
> Oct 13 00:59:03 gandalf ntpd[1432]: proto: precision = 0.111 usec (-23)
> ...
> Oct 13 00:59:12 gandalf ntpd[1432]: ntpd: time slew +0.032806 s
>
> Thus, after an uptime of 38 days, the time drift was +2.169528 s and
> then after 9 days, +0.032806 s.  For comparison, after 20 days up, my
> antique ~10 yrs old HP Pentium 4 box (Intel motherboard) drifted
> +81.659436 s
>
> It all depends on your application and needs as to how accurate you
> computer time needs to be. For time critical software, depending on the
> actual motherboard drift, you can calculate any required ntp update
> interval.
>
> -Ed
>
> _______________________________________________
> PLUG mailing list
> [email protected]
> http://lists.pdxlinux.org/mailman/listinfo/plug

_______________________________________________
PLUG mailing list
[email protected]
http://lists.pdxlinux.org/mailman/listinfo/plug

Reply via email to