As explained previous, CA Disk processing of data sets is based on the data set attributes. Sometimes these attributes conflict with the actual data or the attributes are incomplete. The following explains how CA Disk will treat a variety of possible attributes during Archive, Restore, Move, and Copy processing of physical sequential and direct access data sets.
The attributes CA Disk takes into consideration during processing include the DSORG, RECFM, LSTAR, BLKSIZE and KEYLEN. During processing, CA Disk can also compress the data set, which will have an effect on whether or not TTR addressing is maintained.
The data set organization attribute (DSORG) and its types will be referred to as:
Physical Sequential
Direct Access
Unknown
The record format attributes (RECFM) and its types will be referred to as:
Fixed or Fixed Blocked
Undefined or Variable
Fixed Block Standard
Unknown
The LSTAR is a pointer to the last record in the data set and is in TTR format. TTR format is a three-byte value. The first two bytes are the track number relative to the start of the data set, and the last byte is the record number relative to the start of the track. TTR addressing refers to a pointer for a data block.
Data set compression can occur when a data set is restored or moved to a DASD device with a different track length, if there is empty space on a track, or if there is an empty track in the middle of a data set. Each record is written on the next available space after the previous record. This can change the TTR address of the record and compress the data set. Under some conditions this is not desirable.
Data sets with a DSORG of DA and RECFM of UV require special handling. When a data set with DSORG of DA and RECFM of UV is archived by CA Disk, an ASDA flag is set, indicating the data set is archived as a direct access data set. This flag is also set if an empty track is detected anywhere in the data set prior to the end-of-file (EOF). At restore time, if this flag is set, the TTR value for each record is maintained. That is, a record is written back in the same relative position from which it was read.
This also occurs when sysparm HONORTTR is specified with a value of Y, regardless of the RECFM of the PS, DA, or UNK data set.
Note: ESF Spool data sets must use HONORTTRY if the Move, Copy, Backup and Restore commands are to function properly.
Data sets are considered empty if the LSTAR value is zero. However, a data set with DA in DSORG and RECFM is UV can have all of these attributes and not be empty. In this case, sysparm IGNLSTAR can be specified with a value of Y to cause CA Disk to check bit 0 of the DS1ND field in the format-1. For more information, see the sysparm description for IGNLSTARy in the Systems Guide or data sets with a RECFM of FBS, the BLKSIZE is important. On any other PS or DA data sets, when the EOF is written, it is written on the next track after the last data track. An FBS data set must have the EOF on the same track if another record of BLKSIZE length can be written to that track.
CVOL data sets (a type of IBM catalog whose data set name is SYSCTLG) have a BLKSIZE of 256 bytes and a KEYLEN of eight bytes. Their DSORG is usually either UNK or CTLG. However, they are really PS data sets. IBM requires CVOLs to have a KEYLEN of zero, and their tracks formatted to 25 byte blocks with an eight-byte key. The LSTAR record value must also be X'FF'. After CA Disk has processed a CVOL data set, the DSORG is defined as PS instead of UNK. This does not affect CVOL processing.
CA Disk handles any data set with a DSORG of UNK as a PS data set. Data sets with a RECFM of UNK are treated as though they have a RECFM of U. This corresponds to IBM's handling of the same data sets. Data sets with a DSORG of UNK will have a DSORG of PS after CA Disk processing.
Note: The Idle Space Release function will release space from UNK dsorg's by default. You can instruct CA Disk to bypass this type of DSORG by executing:
EXCLUDE CRI=(DSORG,EQ,UNKNOWN)
|
Copyright © 2015 CA Technologies.
All rights reserved.
|
|