On Jun 3, 2012, at 5:34, AlbyVA wrote:

> Why is the "Current Score" on server Up/Down time so sensitive for IPv6 
> servers?

IPv6 servers are monitored roughly every 14-15 minutes, IPv4 servers every 
24-25 minutes.

https://github.com/abh/ntppool/blob/master/lib/NP/Model/Server.pm#L316

I am planning to make it more frequent for IPv4 servers, too.   Some of the 
work I've done over the last months have been to make storing the historical 
data scale better and more practical, so maybe we can monitor every few minutes 
or so. 

One of the side-effects will be that bad servers will recover much faster, 
which I'm not sure is desirable; another is that if something goes wrong it'll 
go wrong much faster so I need to build out a better automatic "kill switch", 
too.

Finally as others mentioned, IPv6 routing is not always worse but pretty 
consistently less reliable than IPv4.  I'm looking into if there are any 
problems on the network with the monitoring running.


Ask
_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool

Reply via email to