CA Datacom Why is EXTRACT slower than BACKUP?

Products

• Datacom
• DATACOM - AD
• Datacom/DB

Issue/Introduction

There is a difference in performance when running a CA Datacom DBUTLTY EXTRACT and physical BACKUP against the same area.
The EXTRACT is much slower than the BACKUP. It uses more CPU and EXCPs,

Environment

Release : 15.0, 15.1
Component : CA DATACOM/DB

Resolution

The  reason the EXTRACT takes longer to run is that  it has to uncompress the records first. By default the BACKUP specifies CMPRS=YES which leaves the records compressed in the backup. This is explained in the documentation section Back Up Data Area

Compression

If CA Datacom®/DB compression or user compression is being used, the data can be written expanded or compressed. If the backup is being done with a following load which changes the type of compression, the backup should be expanded. Backing up compressed data generally runs faster.

There are some things you can do to speed up the EXTRACT mentioned in the documentation section When to Use EXTRACT:

The data area EXTRACT function uses a form of GSETL/GETIT or GSETP/GETPS command for reading the data blocks. This includes reading multiple blocks per I/O. You can allocate sequential buffers to use for reading the area by specifying the SEQBUFS= keyword.

Also see article 18772 - Is there an easy way to speed up CA Datacom DBUTLTY BACKUP or EXTRACT processing?