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

Terje Mathisen Terje.Mathisen at tmsw.no
Wed Oct 7 19:31:43 UTC 2009


rotordyn at yahoo.com wrote:
> And while a hardware time source would cleanly solve the issue,
> it isn't feasible to retrofit that to thousands of existing
> installations.
> Even if it were $500, there is the expense of getting it installed
> into
> thousands of commercial data centers. Which would be a logistical
> nightmare, given issues of leased spaces, being deep inside
> skyscrapers, etc. It isn't always as simple as sticking a GPS antenna
> to a window.
>
> As to trolling, as I mentioned previously, I'm willing to hire
> someone with the the right level of expertise as a consultant.

If requirement (1) is consistent time across your cluster and then (2) 
preferably UTC, then there is a fairly simple config which will solve 
your problem:

a) Define all your machines as an island, i.e. so that they will stay in 
sync with each other when the external input is lost.

b) Setup a _single_ external source, and use that on either a single or 
4+ to all of your machines.

If the external source goes away, you have a consistent island, possibly 
drifting away, if the external source starts lying you'll follow it, but 
all will be the same.

As long as it (your ref server) works you'll also have true UTC time.

Terje

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




More information about the questions mailing list