[ntp:questions] question regarding NTP configuration for clusters, and "cluster time" stability

rotordyn@yahoo.com rotordyn1 at gmail.com
Tue Oct 6 18:10:14 UTC 2009


On Oct 6, 1:04 pm, Unruh <unruh-s... at physics.ubc.ca> wrote:

> That is why you do not use ONE external source, you use at least 4, so
> that such bad behaviour from one does not destroy your system.

I still have to allow for the general case, where for N servers,
N-1 go offline while the last returns bad inputs to the system.

An unlikely scenario, but not solving it is worse than the
current implementation, where we don't use external UTC
sources.

> You have
> the same problem. What happens if your internal node A jumps by 5 min.
> What do the rest of your internal machines do?

If it is among the internal servers, I believe the rest of the system
will consider it a falseticker and not use it. While the cluster
software
will detect the local clock is too far off cluster time and shut down.

What makes that less likely is that the nodes are internal to the
cluster, and are pretty tightly locked down. While any external
references are beyond our control.

thanks,
tim




More information about the questions mailing list