On Sat, Jun 30, 2012 at 10:01 PM, Dave Hart <[email protected]> wrote: > I surveyed 3326 pool servers by configuring them by IP address in ntpd > with noselect around 20:00 UTC. ntpd took about an hour (possibly due > to rate limiting somewhere of the UDP spews) to stabilize. > Surprisingly, despite the low-rent 6to4 connectivity, only 21 of those > were showing leap=11 unsynced, all due to having never been reachable. > Of the rest, 285 (8.6%) report leap=00 (no insertion pending).
I left ntpd running and re-surveyed the leap indication at 23:30 UTC. The numbers changed quite a bit. Now 18 are leap=11 due to never having received a response. Of the rest, 154 (4.7%) indicate no leap pending (leap=00). Unsurprisingly, neither scan showed any pending leap second deletion (leap=10). Older ntpd believe leap=01 if any of their sources do. Newer ones require a tie or majority of leap=01. Presumably an even smaller percentage of pool consumers will not have been warned, as typically several pool servers are in use by each direct client. Cheers, Dave Hart _______________________________________________ pool mailing list [email protected] http://lists.ntp.org/listinfo/pool
