Peter Roozemaal wrote: > Chuck wrote: >> i have read several places to use many servers in my servers >> listings.. i presently have 12. is there a practical limit where >> after ntp only gets sluggish and the extra servers dont really help >> it? i figured an 'even dozen' should be a good round number :) > > I started with 8 uplinks in my ntpd.conf over 2 years ago. Only 4 of > them are alive at the moment... > For a pool server, 3 active uplinks with different "owners" is IMO a > minimum; I recommend configuring 5-7 servers in the ntpd.conf to have > some redundancy. > (I picked servers "close by", on my continent, low "ping" / ntp delay.)
Since, I haven't seen Brad Knowles around these parts lately, I've take up the flag of Champion of False Tickers... Below is the section from the NTP site about false tickers. 4 servers is the minimum to tell if one server is wrong. > 5.3.3. Upstream Time Server Quantity > > Many people wonder how many upstream time servers they should list in their > NTP configuration file. The mathematics are complex, and fully understood by > very few people. However, we can boil them down to some simple rules-of-thumb: > > * If you list just one, there can be no question which will be considered to > be "right" or "wrong". But if that one goes down, you are toast. > * With two, it is impossible to tell which one is better, because you don't > have any other references to compare them with. > o This is actually the worst possible configuration -- you'd be better off > using just one upstream time server and letting the clocks run free if > that > upstream were to die or become unreachable. > * With three servers, you have no protection against "falsetickers", and ntpd > operation will be degraded and unreliable. > * With at least four upstream servers, one can be a "falseticker" (or just > unreachable) and the system can still figure out which one that is and > which one of the three remaining is the best "truechimer" to sync with. > > According to Brian Utterback, the math officially goes like this: > > While the general rule is for 2n+1 to protect against "n" falsetickers, this > actually isn't true for the case where n=1. It actually takes 2 servers to > produce a "candidate" time, which is really an interval. The winner is the > shortest interval for which more than half (counting the two that define the > interval) have an offset (+/- the dispersion) that lies on the interval and > that > contains the point of greatest overlap. _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
