[ntp:questions] when change ntp servers, log print "localhost: timed out, nothing received"

Terje Mathisen terje.mathisen at tmsw.no
Wed Feb 11 13:33:03 UTC 2015


Harlan Stenn wrote:
> catherine.wei1989 at gmail.com writes:
>> The toish is a tool made by our system, the purpose of "toish
>> setobject cfg.ntp.servers 192.168.1.101" is to change the ntp server.
>> I've searched and find that the time out error may be related to
>> ntpdc's deprecation. Since when I change the server, ntpdc should
>> remove the old server from the remote serverlist and add the new
>> server to it, during which, a time out error occurs. So now the
>> problem is how can I still use ntpdc, it will be really huge task if I
>> use ntpq to replace it in our system so I don't want to give it up at
>> present.
>
> You don't want to use ntpdc for this.  You want to use ntpq instead.
> Or perhaps you could use a "pool" directive that points to a local DNS
> name that contains an updated list of servers you want your clients to
> talk to.

The latter is by far the best suggestion!

We have 7-9 servers behind a single ntp.<corporate_domain> DNS entry, 
with an updated ntpd and a single pool configuration line that's all 
we'll ever need:

Any changes/additions/deletions from that private pool will be detected 
very quickly by the clients.

Terje

-- 
- <Terje.Mathisen at tmsw.no>
"almost all programming can be viewed as an exercise in caching"



More information about the questions mailing list