Specifying the control data set

The recovery operation uses the control data set to determine what data XRC must still write to the secondary volumes. The control data set contains control records that indicate:

The last set of data that is written to the secondary volumes
The amount of unwritten data that exists in the journal
The location of this unwritten data

XRC uses a control data set name in one of the following forms:

SYS1.XCOPY.session_id.CONTROL  or  hlq.XCOPY.session_id.CONTROL

The XRC session ID can be up to eight characters long and must be the same name that you will specify with the session_id parameter of the XSTART command.

Example: The following is an example of a control data set name:

 SYS1.XCOPY.DALLAS.CONTROL

Guidelines: Use the following guidelines to allocate the control data set:

Place the control data set on a different volume from the journal data sets, ideally on its own volume.
Allocate the data set with physical sequential organization, and without extended format.

Prior to its first use, initialize the data set with IEBDG. This is sample JCL for allocation and initialization:

  //STEP1    EXEC PGM=IEBDG                               
  //SYSPRINT DD SYSOUT=*                                  
  //DD1      DD DISP=(NEW,CATLG),                         
  //            DSN=SYS1.XCOPY.TPC.CONTROL,               
  //            STORCLAS=(JRNLSTD),                       
  //            DCB=(RECFM=FB,LRECL=15360,BLKSIZE=15360,  
  //            DSORG=PS),SPACE=(TRK,(1,0))               
  //SYSIN    DD  *                                        
    DSD    OUTPUT=(DD1)                                   
    CREATE QUANTITY=2                              
    END                                            
  /*

Do not allocate the control data set in the Extended Address Space (EAS) of an Extended Address Volume (EAV).

Notes:Any valid combination of RECFM, LRECL and BLKSIZE that results in two single-record blocks on the track may be used. However, the system data mover uses EXCP to write the data set, and may write blocks of different size than you specify. Therefore, you will not be able to browse or otherwise process the dataset with standard access methods once it has been used in an XRC session. The DCB attributes shown in the sample JCL have been chosen to create this situation, and discourage editing or other processing of the data set.

Once a control data set has been used in an XRC session, do not attempt to initialize it again. Should the data set become corrupted, you will not be able to XRECOVER or XADVANCE the session, nor will you be able to restart a coupled session. Restart is possible only by first uncoupling the session through the XCOUPLE PURGE command.

XRC also supports allocation of a PDSE control data set with LRECL=4096 and BLKSIZE=4096. However, IBM® recommends sequential allocation as it provides much higher performance, and is also required if Coupled XRC is to be used.