> I'm seeing an incorrect leap-second indicator from these pool servers:
> Anyone care to confirm?
> starbug.netinertia.co.uk
> syrinx-123.symphonics.org
> de003178.idaq.com
These look ok now.
> dns1.rmplc.co.uk
> gw1-if1-c4l.priorityonline.net
These are still reporting an incorrect leap second.
I had to remove one of the peers from my server as it was causing my
server to report a leap second. Even though I have the
leap-seconds.3427172400 file from NIST configured the current NTP code
will declare a leap second if *any* of your valid peers are indicating a
leap second! An extract from the source code of 4.2.4p7:
* Consider each peer in turn and OR the
* leap bits on the assumption that, if some of them honk
* nonzero bits, they must know what they are doing.
The current dev version of ntp appears to have changed to a voting system
so leap seconds are only declared if your local reference clock or > 50%
of your valid peers are declaring a leap second.
When I looked earlier today dns1.rmplc.co.uk and
gw1-if1-c4l.priorityonline.net were both listing one of the ja.net servers
as their sources. The ja.net servers weren't declaring a leap second, so
one or more of their other peers must have been.
Laurence
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
