Before an operator can start BDT you must provide a start procedure. The start procedure identifies the BDT program that is to run first and the data sets that BDT is to use. Invoking the start procedure creates a BDT address space.

You must store the start procedure as a member of SYS1.PROCLIB or as a member of a data set that is concatenated to SYS1.PROCLIB. You may give the member any name you wish. There is no naming restriction like there is for the TQI start procedure.

Information about invoking the start procedure that you write is in z/OS BDT Commands. A discussion of cold, warm, and hot starting is included in that book. Note that starting BDT for the first time after it is installed requires a cold start.

IBM provides a sample BDT start procedure in SYS1.SBDTSAMP (member name BDT$V2SP). It is shown in Figure 1. A description of each statement that appears in the example follows the figure.

//BDTA1    PROC
//BDT      EXEC PGM=BDTINTK,REGION=5000K,TIME=1440
//STEPLIB  DD DISP=SHR,DSN=SYS1.SBDTLIB
//BDSPOOL  DD DISP=OLD,DSN=BDT1.BDTSPOOL
//CRSPOOL  DD DISP=OLD,DSN=BDT1.BDTSPOOL
//DATAFILE DD DISP=SHR,DSN=BDT1.TQIDATA
//BITMAPS  DD DISP=SHR,DSN=BDT1.TQIBITS
//BDTM001  DD DISP=SHR,DSN=BDT1.MSG0001
//BDTOUT   DD SYSOUT=A
//SYSABEND DD SYSOUT=A
//BDTABEND DD SYSOUT=A
//BDTIN    DD DISP=SHR,DSN=BDT.INISH.DECKS(BDT$FTF)

//BDTA1 PROC

This statement is required. However, the procedure name can be a name other than BDTA1.

//BDT EXEC

This statement is required. You must code PGM=BDTINTK. REGION and TIME are optional.

//STEPLIB DD

specifies that BDTINTK is in SYS1.SBDTLIB.

//BDSPOOL DD

//CRSPOOL DD

These two statements define the BDT work queue data set and are required to be in the start procedure. You may not use the DYNALLOC initialization statement instead of these DD statements to allocate the data set. The ddnames must be BDSPOOL and CRSPOOL and the dispositions must be OLD. The data set name you use must be the same one you specified when you allocated the data set in Step 2. Allocate a Data Set for the BDT Work Queue.

//BDTM001 DD

Each message data set used by your BDT subsystem requires a DD statement in the start procedure or a DYNALLOC statement in the initialization stream. Valid ddnames are BDTMx, where x is any 1 to 4 alphanumeric characters. A suggestion is to start with the name BDTM001 and proceed sequentially: BDTM001, BDTM002, BDTM003, and so forth. The data set names you use must be the same ones you specified when you allocated the data sets in Step 7. Allocate Message Data Sets. The disposition of the data sets must be SHR.

//BITMAPS DD

This statement is optional. It is required if any processors in a complex will run TQI. It identifies the TQI bit-map data set. The ddname must be BITMAPS. The data set name must be the same one you specified when you allocated the data set in Step 6. Allocate the TQI Bit-Map Data Set. The disposition of the data set must be SHR.

//DATAFILE DD

This statement is optional. It is required if any processors in a complex will run TQI. It identifies the TQI checkpoint data set. The ddname must be DATAFILE. The data set name must be the same one you specified when you allocated the data set in Step 5. Allocate the TQI Checkpoint Data Set. The disposition of the data set must be SHR.

//BDTOUT DD

This statement identifies the data set to which BDT will write initialization statements and initialization messages. It is required to be in the start procedure. You may not use the DYNALLOC initialization statement instead of BDTOUT to allocate the data set.

//SYSABEND DD

This statement is optional. If you want a formatted MVS storage dump in the event BDT abnormally terminates, include this statement or a SYSUDUMP DD statement. To get an unformatted dump, include a SYSMDUMP DD statement.

To determine where the dump is to be sent, use the DUMP parameter of the OPTIONS initialization statement.

//BDTABEND DD

This statement is optional. It defines the data set to which BDT is to write the formatted dump.

//BDTIN DD

This statement identifies the data set that contains the BDT initialization stream. This statement is required to be in the start procedure. You may not use the DYNALLOC initialization statement instead of BDTIN to allocate the data set.

By means of the data set name you specify on the BDTIN DD statement, the operator can be given flexibility in choosing which initialization stream to use. When starting BDT, the operator can reply to message BDT3037 in any of the following ways:

• If the BDTIN DD statement specifies a sequential data set, the operator can reply N (for normal) to use the initialization stream in that data set.
• If the BDTIN DD statement specifies a partitioned data set without a member name, the operator can reply M=member to use the initialization stream in that member, or M=xx to use the initialization stream in member BDTINxx, where xx is any one or two alphanumeric characters.
• If the BDTIN DD statement specifies a partitioned data set with a member name, the operator can reply N to use the initialization stream in that member, M=member to use the initialization stream in member, or M=xx to use the initialization stream in member BDTINxx, where xx is any one or two alphanumeric characters.

If you have a system GMJD library (allocated in Step 3. Allocate a System GMJD Library (File-to-File Customers Only)) you must include a DD statement for it in the start procedure or a DYNALLOC statement for it in the initialization stream. The ddname must be GMJDLIB.