On Tuesday 03 November 2009 17:40:35 Rob van der Heij wrote:
> On Tue, Nov 3, 2009 at 4:14 PM, Paul Gilmartin <[email protected]> wrote:
> >> This is cheating. Sure, if you assume the operating system has the
> >> right time, you just ask there...
> >
> > So it's cheating to ask your own OS, but not cheating to ask
> > a remote OS? What if your system lacks network connection?
>
> You just don't simply ask anyone, you ask a complex of synchronized
> servers that have a close connection to an trusted source. Experience
> shows that sometimes the operator will set the z/VM LPAR TOD offset
> not using a reliable source.At that point your System z clock is still
> stable, but very wrong.
>
> For Linux on z/VM, the best approach is to make sure z/VM has the
> right time at IPL. When your applications really can't afford the
> typical drift of a System z, make the hardware steer the LPAR. That
> will make all Linux guests have the same clock. Often better than each
> trying to sync itself with the outside.
>
> Rob
Philosophical question: Has anyone posted a requirement to SHARE or
some
such to get an interface to an NTP source into the firmware so that the TOD
clock can be synchronized to a reliable source (other than the operator's
wristwatch or the computer room's wall clock)?
I've wondered about that ever since Linux-390 happened. :-)
Leslie Turriff
