On Monday, 07/26/2004 at 09:36 AST, David Boyes <[EMAIL PROTECTED]>
wrote:
> On Mon, Jul 26, 2004 at 12:18:00PM -0700, Wolfe, Gordon W wrote:
>
> > Can you provide a reference to setting up VM with sysplex timers?
>
> Unfortunately, as far as I know, you can't, but I was thinking that
> (like the SIGNAL SHUTDOWN function), it's an architected function
> that's already part of the ESA and zArchitecture POP that could be
> co-opted for the purpose of allowing clock adjustment. VM tolerates
> the timer (it can get its time from a clock controlled by a timer
> updated by a sysplex member), but can't really exploit it directly.

>From my post of November 25, 2003:
| For small clock corrections, LPAR and virtual machine clocks will be
| corrected automatically by the h/w, so Linux (or any other opsys) will
| serve the correct time.  For a sufficiently large clock correction, e.g.
| reestablishment of external time reference, the clock will not change
and
| the LPAR or guest will not serve the correct time.
|
| Don't ask me what "sufficiently large" is.  I don't know.

While you can use z/VM in a CEC with a sysplex timer, z/VM will not detect
gross shifts in time.  The timer is still useful for setting the correct
TOD when the LPAR is activated.

Of course, I think it would be much nicer to figure out a way to solve the
problem so that (a) the external time references can be used and (2) no
special support within the guest is required.  "It just works the way it
oughtta.  It's a miracle."

Alan Altmark
Sr. Software Engineer
IBM z/VM Development

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to