DD statements are required for input and output data sets. There must be one DD statement for each unique data set used in the job step. You must specify a unique output DD statement (for the unload data set) for every input data set that you unload. You cannot unload multiple data sets to the same unload data set.

Data sets that are used as input data sets in one copy operation can be used as output data sets in another copy operation.

Input data sets cannot be concatenated together on the same DD statement.

Fixed or variable records can be reblocked, but you cannot convert fixed format records into variable format records, or variable format records into fixed format records. Reblocking or deblocking is performed automatically when the block size of the input data set is not equal to the block size of the output data set.

For variable format data sets, the output LRECL must be greater than or equal to the input LRECL. For fixed format data sets the input and output LRECL must match. For undefined format data sets, the output BLKSIZE must be greater than or equal to the input BLKSIZE, and any LRECL is ignored.

A PDSE might require somewhat more space than a partitioned data set requires to hold the same data. When copying from a partitioned data set to a PDSE, IEBCOPY will override a secondary space allocation of zero for the output PDSE, rather than stop the job. The copy will obtain any additional space it requires in one unit increments of the primary allocation (tracks, cylinders, or blocks). The secondary space quantity in the data set label will not be changed.

When IEBCOPY must supply DCB parameters for the output data set, it will use the corresponding values from the input data set. In particular, system determined block size will not be used for the output data set block size, nor will the system determined block size or reblockable flags be set in the output data set label.

You should allow IEBCOPY to choose the DCB parameters for an unload data set. If you must specify a block size, specify a value that is at least 20 bytes greater than the input data set block size. (This avoids creating spanned records in the unload data set.)

OPTCD=W may be used to request write verification of the data written and causes DASD cached controllers (for instance, 3880-21, 3990-3) to perform a "write through" operation.

Starting with ESCON support, OPTCD=W also causes a different set of channel programs to be used for writing to ECKD™ capable DASD and for reading from all DASD. These programs could run much slower than the default ones; therefore, there is a significant performance cost for using OPTCD=W.

The write verification requested by OPTCD=W is usually unnecessary, because extensive error recovery occurs in both hardware and error recovery procedures (ERP) used by the MVS I/O supervisor. OPTCD=W does not cause the data read from the device to be compared with the data in virtual storage and cannot detect data garbled in transmission from virtual storage to the device control unit. OPTCD=W causes the control unit to read data from the device (after it has been written) to validate the error checking codes stored with the data.

Related reading: For more information, see IEBCOPY Unload Data Set DCB Parameters .