[ntp:questions] NTP request retry?
unruh at invalid.ca
Tue Jan 28 01:55:33 UTC 2014
On 2014-01-28, Harlan Stenn <stenn at ntp.org> wrote:
> Brian Utterback writes:
>> On the other hand, I have definitely observed that phenomenon as a
>> source of asymmetric round trip time. The outgoing request packet gets
>> delayed for ARP request and reply at each hop, but the return packet has
>> no such delay. Quite a while ago I suggested a special burst mode where
>> two packets were sent, one shortly after the other and the first one
>> would be ignored. Dr. Mills said that the first would generally be
>> ignored because of the longer round trip time (delay), but I thought
>> that treating the first packet as a throw away would be better because
>> otherwise you end up with half the number of "good" samples in the
>> billboard. Anyway, nothing every came of the discussion.
> I'm game to see this discussion warmed up.
chrony has this option. It will send out a first packet to wake up the
link and then a second real packet to use the link.
If the timing measurements being made by `chronyd' are the only
network data passing between two computers, you may find that some
measurements are badly skewed due to either the client or the
server having to do an ARP lookup on the other party prior to
transmitting a packet. This is more of a problem with long
sampling intervals, which may be similar in duration to the
lifetime of entries in the ARP caches of the machines.
In order to avoid this problem, the `presend' option may be used.
It takes a single integer argument, which is the smallest polling
interval for which a pair of packets will be exchanged between the
client and the server prior to the actual measurement being
initiated by the client. For example, with the following option
included in a `server' directive :
when the polling interval is 512 seconds or more, a UDP echo
datagram will be sent to the server a short time (currently 4
seconds) before the NTP client mode datagram.
More information about the questions