On Wed, 15 Jun 2016 10:15:56 -0700, retired mainframer wrote: > >Create two new PDSes. One with the original DCB attributes of the broken >dataset. The second with the new DCB attributes of that dataset. > Make backups; work with copies. Will HBACKDS/HRESTORE RENAME deal with inconsistent member attributes?
>Run IEBCOPY to copy only the new member(s) produced by LISTC to the second >dataset. (Use the SELECT control card.) The DD cards for this job should not >specify DCB information for either PDS. > >Run a second IEBCOPY to copy only the original members created before you ran >the first LISTC operation. (Use the EXCLUDE command.) On the DD statement >pointing to the broken dataset, specify LRECL=80, RECFM=FB, and BLKSIZE=32760. > (For disk datasets, a BLKSIZE greater than the longest block is not a >problem.) IEBCOPY will reblock the records to match the BLKSIZE specified >when the output dataset was created. > JCL may insist that BLKSIZE be a multiple of LRECL. Will IEBCOPY insist on using the DSCB attributes of SYSUT1, copying to SYSUT2? Ugh! DFSMS should *never* allow changing the attributes of an existing nonempty data set. DFSMS should *almost*never* allow overwriting the directory when DSORG=PO. (If the programmer explicitly overrides to DSORG=PS, programmer's gun; programmer's foot.) -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN