It is indeed wise to list all volumes first, ICKDSF command:
 CPVOL LIST UNIT(xxxx) NVFY
If it list something else than PERM, you can be sure that this dummy VTOC is
there, and that you don't need to rerun a CPVOL ALLOCATE.nor CPVOL FORMAT
RANGE(0,0)
Stronger: you only need a CPVOL FORMAT RANGE(0,0) if a CPVOL LIST reports
   ... Cylinder zero not in CP FORMAT
because in all other cases, CPVOL FORMAT has been run before.

And, beware: if you ever run CPVOL FORMAT RANGE(0,xx) or ALLOCATE on a pack
that has DRCT area(s), be sure to run DIRECTXA afterwards, because ALLOCATE
will write X'40' for DRCT cylinders, meaning "cylinder  free to write a new
DIRECTORY"; DIRECTXA searches for X'40' will write X'C0' to indicate where
the active CP directory has been written
Beware2: if you'd format cilinder 0 of the VM resident, you not only need to
run ALLOCATE, but you mlust rewrite the Standalone Program Loader (SAPL),
with the SALIPL utility.

The guide has spoken again.

-- 
Kris Buelens,
IBM Belgium, VM customer support

2010/1/20 Mike Walter <mike.wal...@hewitt.com>

> Not sure why you're installing z/VM 5.3 when 5.4 is available for the
> foreseeable future (to support boxes older than z10's).  But...
>
> Gotchas: Yeah... before you format the DASD, be sure to run CPFMTXA from
> z/VM (you can probably do the ALLOCATE via ICKDSF from z/OS, too - but why
> get more systems involved?).  CPFMTXA EXEC is an IBM-supplied handy z/VM
> front-end to ICKDSF which automatically sets all  the appropriate
> parameters so that the dummy VTOC is written to each DASD so that they
> look 100% full to z/OS (preventing it from blithely writing over your
> minidisks, CP area, etc.   Had CPFMTXA been used, you probably would not
> gotten into this mess, Lucy!
>
> When running CPFMTXA (or ICKDSF from z/OS), and BEFORE any formatting, be
> sure to run the ALLOCATE function to display the existing allocation
> bitmap.  This will tell you how each cylinder is allocated (normally DRCT,
> PAGE, PARM, PERM, SPOL, or TDSK).  You don't want to change anything -
> just be sure you document how each cylinder range is currently allocated.
>
> Whether you use CPFMTXA or z/OS, you'll need to ensure that the
> allocations remain (or are re-allocated) the same after the format.  I'm
> not sure if the bit indicating the active DRCT cylinders will be affected
> by formatting cylinder zero of the sysres.  Someone else can address that.
>
> You might want to test on a volume that is allocated as all PERM space
> before trying anything allocated as DRCT, PAGE, PARM, SPOL, or TDSK.
>
> Welcome to z/VM!  :-)
>
> Mike Walter
> Hewitt Associates
> The opinions expressed herein are mine alone, not my employer's.
>
>
>
> "Mike Day" <y...@csi-soft.com>
>
> Sent by: "The IBM z/VM Operating System" <IBMVM@LISTSERV.UARK.EDU>
> 01/20/2010 03:24 PM
> Please respond to
> "The IBM z/VM Operating System" <IBMVM@LISTSERV.UARK.EDU>
>
>
>
> To
> IBMVM@LISTSERV.UARK.EDU
> cc
>
> Subject
> Backing up VM Volumes from z/OS
>
>
>
>
>
>
> We installed a new z/VM 5.3 from tape a while back and are now trying to
> back up the z/VM volumes using ADRDSSU from a z/OS system which runs in a
> different LPAR.
>
> We get a message that says "the VM-formatted volume does not have an
> OS-compatible VTOC".
>
> I found some instructions for using ICKDSF (see below) to format cylinder
> 0
> which is supposed to solve the problem.  I am wondering if this is the
> only
> way to handle this, or if there are other way of doing it or gotcha's I
> should watch out for.
>
> Thanks in advance for any information or suggestions.
>
>
> Mike Day
> Confident Software
> y...@csi-soft.com
>
>
>
>
> ---------------------------------------------------------------
>
>
> > Subject: Re: ADRDSSU backup of VM volume
> >
> >
> > I guess that 510W01 was never formatted with CPVOL.  Because, CPVOL is
> > supposed to not only create CP's allocation bytemap, but also a VTOC
> telling
> > z/OS (& co) that the disk is full.
> > So,
> >
> >
> > 1. Assure you have no minidisk on 510W01 starting in cylinder 0 (you can
> > ignore the minidisk of user $ALLOC$)
> > 2. Run ICKDSF, press enter twice (to indicate CONSOLE as in- and output
> >
> >
> > 1. Link to a FULLPACK overlay on 510W01, or use DEFINE MDISK:
> >  cp define mdisk 1234 0 end 510W01
> >
> > 2. Verify that there is no CP area on the disk (it should not, otherwise
> CPVOL
> > would have created a dummy VTOC, but you wouldn't like to loose a CP
> area).
> > Run
> >  cpvol list unit(1234) verify(510W01)
> > I expect it will tell: ICK33001I 510W01 CYLINDER ZERO NOT IN CP FORMAT.
> If it
> > would -unexpectedly- list CP areas (SPOL, PAGE, DRCT, TEMP, ...  stop
> here.
> >
> > 3. Format cylinder 0:
> >  cpvol format unit(1234) verify(510W01) range(0,0) ) volid(510W01)
> > 4. END
> >
> > 3. DETACH  1234
>
>

Reply via email to