[LEAPSECS] Coming of age in the solar system
M. Warner Losh
imp at bsdimp.com
Sun Sep 5 14:18:26 EDT 2010
In message: <1009051753.AA08511 at ivan.Harhan.ORG>
msokolov at ivan.Harhan.ORG (Michael Sokolov) writes:
: Tony Finch <dot at dotat.at> wrote:
: > This reminds me of 2008-12-31 23:59:60 Z when a lot of Oracle systems
: > crashed or rebooted because the Solaris clock went backwards.
: That's Solaris' fault then. It wouldn't happen under 4.3BSD-Quasijarus,
: Folks, this is ALL it takes to have a trouble-free internal time scale
: in the presence of leap seconds!
Systems that have requirements for synchronization in the millisecond
range can't use this solution. Systems that require passing a leap
second test and displaying 23:59:60 can't use this solution. Many
Oracle databases back financial trading software, which operate in
markets that mandate closer time synchronization than the 1s
synchronization this technique gives. So solaris is back to being
stuck in the "undefined" portion of the POSIX specification, which
interacts poorly with Oracle's underlying assumptions about the OS.
More information about the LEAPSECS