On Friday, 06/02/2006 at 08:27 EST, "McKown, John"
<[EMAIL PROTECTED]> wrote:
> I forgot to menton why I was curious. Basically it is "protect me from
> myself". I've seen a number of people here post about Linux in an LPAR,
> not as z/VM guest, and sharing DASD. They, like me, are likely from a
> z/OS background where we've done this for decades reliably (and
> sometimes not so reliably). If the "mount" command could put a reserve
> up on the DASD volume containing the filesystem, that would simply make
> it impossible for another Linux instance to do a mount for a filesystem
> on that same DASD volume. It literally could not get access to the DASD
> device.
>
> OK, it is stupid. I'll go back to lurking again.

I'm not sure why you say it's stupid; using a RESERVE that way isn't
stupid by any means, but it *is* a very large hammer.  Reserving the
volume in order to mark the VTOC, say, with a "busy" indicator, then
releasing the volume provides a good advisory locking mechanism, but it
does require that the participating systems all honor the lock.  (A device
reservation isn't perfect, either.  It can be released if another system
issues an Unconditional Reserve.)

I don't think even z/OS holds a continuous reserve, does it?  Only if
someone does an exclusive ENQ with the right GRS (?) configuration?  As
you say, "sometimes not so reliably", which is more of a comment about the
sysprog than the system.  :-)

Alan Altmark
z/VM Development
IBM Endicott

----------------------------------------------------------------------
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