Image Copy function

These topics provide information about the Image Copy function.

Purpose

The batch Image Copy function generates image copy data sets of databases that have been taken offline or placed in a non-update mode.

It reads an input data set group or area and creates one to ten image copy data sets. Since the database is not available for update purposes during the image copy process, the Image Copy function creates an image copy of the data set as it appears at a particular point in time.

The following figure shows the processing flow of the Image Copy function.

Processing considerations

The Image Copy function does not support the following:

GSAM, HSAM, or MSDB database organizations
the IBM Utility Control Facility (UCF)

DBRC considerations

Having DBRC active is not required for the Image Copy function.

It is, however, highly recommended.

KSDS and ESDS overflow data set considerations

When generating image copies of a VSAM KSDS, a HISAM overflow data set, or an overflow data set for a secondary index with duplicate keys, the Image Copy function uses control interval processing to improve performance.

You can also do the following things to improve performance even more:

Specify the free space element compression option (with the ICALG keyword). A KSDS has no free space, but under this option, the function makes the output record variable blocked (VB), which improves performance. If the LRECL is less than 64 bytes, the function must write a record of at least 64 bytes, the size of the image copy header. The IMS Database Image Copy utility performs a similar expansion of the output record. By using compression, the Image Copy function can write fewer bytes than the IMS Database Image Copy utility.
Copy several data sets concurrently.
Stack multiple data sets on the same tape and run the copies asynchronously. In this case, the function can overlap the overhead of DBRC and OPEN/CLOSE processing and reduce elapsed time.
Create virtual image copies of your primary and secondary indexes. Generally, a lost primary or secondary index can be rebuilt from the database faster than the index can be recovered from an image copy if log tapes are supplied to the recovery process. For more information, see Virtual-image-copies-DBRC.
If you use the Image Copy function to copy these types of data sets, do not specify the AMP=('BUFND=n') keyword in the DD statement for the database. Let the Image Copy utility calculate the proper number of data buffers to ensure optimum performance.