[ntp:questions] Clarification Required about NTP

Steve Kostecke kostecke at ntp.org
Fri Feb 6 15:49:16 UTC 2009

On 2009-02-05, balajisundar at midascomm.com <balajisundar at midascomm.com> wrote:
> Dear All,
> Currently i am using Red Hat 7.2 Linux PC
> I have 3 Red Hat 7.2 Linux PCs for Testing
> I have using 2 Linux PC as Redundant PC
> I have configured floating IP as NTP Server on 3 Linux PCs
> Floating IP is
> PC1 IP is
> PC2 IP is
> PC3 IP is
> I have Written a script to monitoring the PC1 to PC2 vice-versa every
> 30 Seconds
> Assume PC1 is currently Active and its Network is goes down or
> Shutdown the PC1 then the Floating IP will be host in PC2 and
> respectively

If you are attempting to provide some sort of redunancy or fail over you
are going about this the wrong way.

Are these 3 systems operating in a time island (i.e. disconnected

> *Clarification Required is* On this situation whether i need to
> restart the ntp server or it will automatically take care NTP Server
> IP Address changes and if its takes automatically then how much time
> its take to sync with configured ntp Server

ntpd, when properly configured, will automatically handle fail-over.
Once a remote time server, or ref-clock, becomes unreachable ntpd
selects the next best sys_peer from the surviving group of candidates.

The default maximum time between polls is 1024 seconds, or 16.84
minutes. So that is the maximum about of time which could elapse between
the poll of an IP address on one server and the poll of that IP address
after it has moved to another server. However, as has been mentioned
elsewhere in this thread, your scheme introduces a number of problems.

Steve Kostecke <kostecke at ntp.org>
NTP Public Services Project - http://support.ntp.org/

More information about the questions mailing list