That last makes for a good question. I do not have any guests other than
CMS, GCS and TPF varieties, so I have no way to trace to see if it is
still a problem. The only time I encountered it was in a non-VM shop.
Fortunately, we were able to wrest the disk away from the system before
there was any real harm. I had the onerous task of restoring the dummy
VTOC using the then state of the art IEHDASDR and SuperZap. 

I cannot reproduce the situation here. The DASD Storage Management folks
are adamant about not having VM volumes accessible to z/OS and
vice-versa, which suits me just fine. That eliminates the only system I
know of that has the potential to cause the problem.    

Regards, 
Richard Schuh 

 

> -----Original Message-----
> From: The IBM z/VM Operating System 
> [mailto:[EMAIL PROTECTED] On Behalf Of Alan Altmark
> Sent: Monday, September 29, 2008 1:33 PM
> To: [email protected]
> Subject: Re: CPFMTXA VOLUME FORMATTING QUESTION
> 
> On Monday, 09/29/2008 at 02:06 EDT, "Schuh, Richard" <[EMAIL PROTECTED]>
> wrote:
> > Unless things have changed in the last few years, the VTOC written 
> > when formatting the disk as a CPVOL does not go far enough. 
> If the MVS 
> > DASD Storage Allocation Routine is interrupted while the 
> checking for 
> > free space on the volume, a bit (formerly known as the DADSM 
> > Interruption Bit or DOS VTOC bit, currently known as the 
> VSE bit, name 
> > DS4DOSBT in the F4 DSCB DSECT) is left on. The next time 
> MVS attempts 
> > to allocate space on the volume, it will try to create proper free 
> > space records. It does this buy starting with an F5 DSCB that shows 
> > all space available on the disk. It then runs the F1 and F3 chains, 
> > allocating each described extent. Since there are no 
> allocated extents 
> > on a CPVOL formatted disk, it shows the entire volume as 
> being available for space allocation.
> > There needs to be at least 1 F1 DSCB allocating the entire 
> volume to a 
> > space-holder dataset to prevent this highly unlikely occurrence.
> 
> You need to open a PMR with the ICKDSF team on this if you're 
> interested in getting it fixed (assuming it is still a problem).
> 
> Alan Altmark
> z/VM Development
> IBM Endicott
> 

Reply via email to