Leap second

serialdie

Is pfsense affected by the leap second change this year?
TIA.

Gertjan

A pfSense running in most parts in Europe has to deal with a time shift of one hour twice a year.
Works pretty well for the last 6 years.

So, 1 second …... ;)

edit : My pfSense being used at work 'WOL' the coffee machine in the morning...... I'm gona cross my fingers.

doktornotor

If you run NTPd on pfSense and care about the leap second on June 30, 2015 at 23:59:60 UTC

1/ Get the leap seconds file from one of the following locations:

ftp://time.nist.gov/pub/leap-seconds.3629404800
ftp://tycho.usno.navy.mil/pub/ntp/leap-seconds.3629577600
http://www.ietf.org/timezones/data/leap-seconds.list

2/ Upload (or paste) the entire file via the GUI under Services - NTP - Leap seconds and Save.

3/ Done. ;)


# grep leapfile /var/etc/ntpd.conf
leapfile /var/db/leap-seconds


Jun 22 22:45:02	ntpd[69902]: 0.0.0.0 c01e 0e TAI 36 leap 201507010000 expires 201512010000
Jun 22 22:45:02	ntpd[69902]: leapsecond file ('/var/db/leap-seconds'): loaded, expire=2015-12-01T00:00Z last=2015-07-01T00:00Z ofs=36
Jun 22 22:45:02	ntpd[69902]: leapsecond file ('/var/db/leap-seconds'): good hash signature

Using the NIST Leap Second File

P.S. Don't forget to post your crash experience/logs if you survive till July 1. :D

cmb

We've already been through 3 leap seconds in our 11 years, and I've never heard of any issues. Never heard of any issues with stock FreeBSD either.

There isn't a need to do anything, things will continue to just work. Only stratum 1 time servers, and applications that are extremely time-sensitive (none of which you're running on your firewall) generally need anything there.

doktornotor

Well last time it caused loads of issues with Java junk and MySQL on Linux. When you don't know how buggy your software is, you might be better off lying to NTP clients and pretending there's no leap second at all – like Google. :D

NOYB

With 61 Seconds in a Minute, Markets Brace for Trouble

NOYB

Everybody survive the extra second okay?

Harvy66

@Gertjan:

A pfSense running in most parts in Europe has to deal with a time shift of one hour twice a year.
Works pretty well for the last 6 years.

So, 1 second …... ;)

edit : My pfSense being used at work 'WOL' the coffee machine in the morning...... I'm gona cross my fingers.

Local time is based on an offset of UTC. The yearly DST time change doesn't really affect time. The leapsecond affects UTC.

doktornotor

@NOYB:

Everybody survive the extra second okay?

Apparently…


Jun 30 02:25:34	ntpd[97093]: kernel reports leap second insertion scheduled
Jul 1 02:26:42	ntpd[97093]: kernel reports leap second has occurred

Harvy66

NTPD RRD shows an average of 0.14ms offset and a maximum 0.4ms offset over the past 24 hours.