[ntp:hackers] smearing the leap second

brian utterback brian.utterback at oracle.com
Wed Jun 24 15:51:48 UTC 2015


On 6/24/2015 11:07 AM, Harlan Stenn wrote:
> Danny Mayer writes:
>> > On 6/23/2015 10:53 PM, Harlan Stenn wrote:
>>> > > Danny Mayer writes:
>>>> > >> I've been following this thread with interest and I have the following
>>>> > >> comments.
>>>> > >>
>>>> > >> 1) Stealing bits is wrong. That not a moral judgement but an
>>>> > >> architectural one. Whether you get them from the RefID or the fuzz bits
>>>> > >> of the timestamps you are not going to get what you want. If you take it
>>>> > >> from the RefID then the reference information will be incorrect and
>>>> > >> you've lost loop detection. Moreover it won't be usable unless it's a
>>>> > >> stratum 1 server and then you lose the information about the RefClock.
>>>> > >> If you take it from the fuzz bits you lose the purpose of those bits.
>>>> > >> There's actually a specific purpose to them. I don't remember all the
>>>> > >> details but they are important.
>>> > > 
>>> > > It is possible to convey the smear  offset value in an extension field.
>>> > > 
>>> > > The only instances of NTP that will know what to do with this are "new"
>>> > > instances of NTP.
>>> > > 
>>> > > What happens with the tens of millions (could be 100M, not sure) that
>>> > > don't know about this?
>>> > > 
>> > 
>> > There is actually nothing you can do about this. ...
> We disagree.  If the servers send smeared time with no leap indication
> old clients will do the right thing.
>

I agree. Any solution we can eventually agree on will almost certainly
have to send the original timestamps adjusted with the smear. The
extension field would have to contain the amount the timestamps have
been smeared so an aware ntpd could recover the actual time. But this is
all just an intermediate solution. The correct thing is to engineer the
smear into the clients, not the servers. I think the server side smear
should be a configuration option which if compiled in requires an active
measure to enable (not default) and prints a big scary message to the
system log when it is in use. The server smear breaks the NTP spec in
multiple ways no matter how we do it and it should never be done lightly
or haphazardly. Client side smear is pretty much just another feature.
-- 
Oracle <http://www.oracle.com>
Brian Utterback | Principal Software Engineer
Phone: +1 6038973049 <tel:+1%206038973049>
Oracle Systems/RPE Solaris Network
1 Oracle Dr. | Nashua, NH 03062
------------------------------------------------------------------------
This space for sale or rent.
------------------------------------------------------------------------
Green Oracle <http://www.oracle.com/commitment> Oracle is committed to
developing practices and products that help protect the environment


More information about the hackers mailing list