[ntp:questions] NTP servers not accessible on some networks
nomail at example.com
Wed May 21 07:41:00 UTC 2014
mike cook <michael.cook at sfr.fr> wrote:
>>>> If I keep pinging different pools,
>>>> I can find another NTP server which replies.
>>> Seems strange, that would imply that the server 126.96.36.199
>>> (or their ISP) is blocking your client?
> I regularly get inaccessible servers returned by pool directives but have never worried as I have local refclocks and servers so they would only get selected if those were down. I can only guess that the pool does not make sufficient, or timely verification of service supplied before providing the address.
If anything, the pool does more verification than required and drops a
server more quickly than would be required.
This verification is done from a single server, and it happens that many
servers are dropped just because there are routing issues between that
server and the NTP servers. Especially on IPv6.
When you have regular problems with inaccessible servers it is an indication
that the NTP protocol is filtered by one of the upstream providers used
by your ISP. That happens. The pool admins cannot do anything about that.
You can try "traceroute -U -p 123 ip.ad.dr.es" to see where it is being
blocked, and if it is nearby to your ISP complain to your ISP about it.
More information about the questions