On Jan 1, 2009, at 7:59, Ask Bjørn Hansen wrote:
> Count of server measurements (about 2.2 per server per hour) and > servers that were accurate within ~80ms, except one second ahead of > the monitoring server. > > Basically it was none until midnight (most hours on the day before are > missing) and then 158; going down to 13 now. Two of the RedHat's > which are stratum 1 using a CDMA device IIRC. The two RedHat servers got fixed yesterday; so it's finally basically over. Summary: Not a great idea to have the leap second on a holiday. mysql> select date(ts) as day, count(*) as tests, count(distinct server_id) as servers from log_scores where ts > '2008-12-31' and offset > 0.92 and offset < 1.08 group by date(ts); +------------+-------+---------+ | day | tests | servers | +------------+-------+---------+ | 2008-12-31 | 17 | 3 | | 2009-01-01 | 1073 | 175 | | 2009-01-02 | 185 | 12 | | 2009-01-03 | 156 | 4 | | 2009-01-04 | 109 | 3 | | 2009-01-05 | 66 | 3 | | 2009-01-06 | 4 | 1 | +------------+-------+---------+ 7 rows in set (1.62 sec) -- http://develooper.com/ - http://askask.com/ _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
