Thanks for your analysis and hints.

Our default seems to be DYNALOC=N. This is why I tried to overwrite with
DYNALLOC=SYSDA. I was hoping that DF-Sort will figure out how many data
sets to allocate based on the estimated size. It seems it does not, but
instead it allocates either a default number (4) or the number specified
via DYNALLOC=.
There is a drawback of specifying the number via DYNALLOC=: DF-Sort
always allocates this number or work data sets, no matter how much space
will be needed. 

The job in discussion runs multiple times a day and usually has only a
few megabytes to sort, but occasionally there will be a much larger
amount of data to work on. If I specify, say DYNALLOC=(SYSDA,100), then
there always will be 100 data set allocated by DF-Sort. 

Will DF-Sort adjust the size of the dynamic allocation based on the
estimated size?

Is there any other option I could try, or am I stuck to
DYNALLOC=(SYSDA,nnn) with a reasonably high "nnn"?

--
Peter Hunkeler

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On
Behalf Of David Betten
Sent: Mittwoch, 3. Oktober 2012 09:59
To: [email protected]
Subject: Re: Help with DF-Sort not being able to allocat enough work
data set space

Based on the record statement and the FILSZ specified, DFSORT estimated
it
would need 469577MB of disk work space.  Your installation default for
DYNALOC is probably the shipped default of SYSDA,4.  This is going to
cause
DFSORT to allocate 4 work data sets.  The work space allocation will be
spread across three of those with one set aside to be used for recovery
if
the work space requirement is larger than expected.  So with such a
small
number of work data sets, you need to have at least 3 volumes with
156526MB
of free space.  Since there were not any volumes available with that
much
free space, your sort failed.  On your OPTION statement you might want
to
change DYNALLOC=SYSDA to DYNALLOC=(SYSDA,30).  This would cause DFSORT
to
spread that 469577MB allocation across 29 work data sets (with 1 set
aside).  So now you need at least 29 volumes that have 16192MB
available.
If your storage group has mostly MOD 54s or MOD27s defined, that should
be
fine.  If it has something like a bunch of MOD 9s defined, you'd have to
increase DYNALLOC further to spread the work space allocation across
more/smaller work data sets.


Have a nice day,
Dave Betten
DFSMS Performance Engineer
IBM Corporation
email:  [email protected]
DFSORT/MVSontheweb at http://www.ibm.com/storage/dfsort/

IBM Mainframe Discussion List <[email protected]> wrote on
10/03/2012 03:00:19 AM:

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to