I am currently trying to get a server out of the pool before it gets
decomissioned completely.
It has been removed from the pool like two months ago, but it still
receives a lot of requests.
I have sent KoD on these requests for some time, but they keep coming in.
I think most of them are from ISP routers that have pool.ntp.org as
their default NTP server setting
and query the pool in SNTP manner to set the router clock. Queries are
usually either once per
hour or once per day, but there is a huge amount of such routers. I
think it should be put as a
requirement on manufacturers of such routers that the re-resolve the DNS
name at least once
a day, instead of resolving it only once at bootup and then using the
same IP address forever.
But also there are more NTP-like users. They have a polling interval
around 1024 seconds.
They apparently are stabilized instances of ntpd or similar, but still a
lot of them keep coming
back when receiving KoD. Maybe the processing of KoD has been removed
from ntpd just like
the generation of it?
As long as we cannot get such simple measures correctly implemented,
forget about more
complicated things. Sure you could add a lookup feature, but nobody is
going to use it.
Rob
On 7/21/19 11:51 PM, Hal Murray wrote:
Is there any way for client code to determine if an IP Address is still in the
pool? Maybe something like a DNS lookup on d.c.b.a.pool.ntp.org
If there was a way to do that, it would be easy to update servers to check
occasionally and stop using servers that have left the pool.
_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool
