On 09/22/2016 07:30 AM, Tom Marchant wrote:
> On Thu, 22 Sep 2016 13:48:42 +0200, R.S. <r.skoru...@bremultibank.com.pl> 
> wrote:
>> PDSEs were available long before z/OS brandname was born.
> Executable code could be stored in PDSE starting with MVS 4.3 in 1992.
> From the General Information manual:
> <quote>
> Additional MVS/ESA SP 4.3 support with DFSMS/MVS allows users to store and
> load executable code in a partitioned data set extended (PDSE).  Users can 
> create and manage loaded modules and load libraries in PDSEs with the same
> data management techniques as other PDSEs.
> </qoute>

Perhaps a better question would be when did IBM and the MVS user
community actually have enough confidence in PDSE reliability to  start
distributing products with PDSE program libraries, especially ones that
needed to go in the LNKLST concatenation. 

My recollection is that PDSEs did have a  significant number of issues
for a number of years including data-loss and availability APARs.  You
had to have a strong need for some of the unique PDSE features to take
the risk in the early years.   About the only way to corrupt a PDS was
to open one improperly, and with the most common error (attributes
change) most of the damage was easily repairable.  We had a number of
cases (more than 1, less than 10, at least 2 night calls) in the early
PDSE years where access rules were not violated and an entire PDSE
became corrupted and unreadable and also at least one case where a PDSE
became hung and inaccessible until next IPL!  That would have been a
much more serious issue for us had those incidents involved a product
program library instead of just an application data library..
    Joel C. Ewing

Joel C. Ewing,    Bentonville, AR       jcew...@acm.org 

For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to