[nSLUG] Why is ntp so terrible?

Dave Flogeras dflogeras2 at gmail.com
Sat Mar 11 00:41:19 AST 2017


On Fri, Mar 10, 2017 at 6:14 PM, <billdavidson at eastlink.ca> wrote:

> NTP will resist making large changes, so on the hosts that are still far
> off you may need to stop the local ntpd, run ntpdate, then start ntpd
> before the clock can skew too far.  I haven't played around with this stuff
> in a while, but ISTR a switch that would tell ntpd to "slew" to make large
> changes.
>

This one has gotten me in the past on embedded systems that may be off for
months/years and slew a lot, then be turned on an expected to perform
without any user interaction.  The option is "-g" which allows the first
(and first only IIRC) adjustment to be arbitrarily large.  Under normal
circumstances, it will only adjust by maybe 10min or so, and if it has to
be larger it will just panic and die.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://nslug.ns.ca/pipermail/nslug/attachments/20170311/50af37f6/attachment.html>


More information about the nSLUG mailing list