RECON data set access

While the DBRC API is running, it uses up to three RECON data sets, which are allocated using the DD names RECON1, RECON2, and RECON3, or their respective alternate DD names. MDA members must be created for the data sets before the DBRC API can allocate RECONs.

The RECONs can be allocated:

Dynamically by the API when the DSPAPI FUNC=STARTDBRC request is issued–this is the recommended method
Dynamically by your application program
Through your JCL

If the DBRC API allocates the RECONs, they are deallocated when the DSPAPI FUNC=STOPDBRC request is issued. Before the DBRC API can allocate the RECONs, MDA members must have been created for RECON1, RECON2, and RECON3. or their respective alternate DD names. The MDA members must exist either in a library allocated as //IMSDALIB, or in the //JOBLIB or //STEPLIB libraries. The DBRC API first searches in //IMSDALIB, if it exists, for the MDA members.

The IMS libraries concatenated to JOBLIB or STEPLIB are usually APF Authorized. If your program runs APF Authorized and the library containing the MDA members is not APF Authorized, allocate it using IMSDALIB as the DD name.

Requirement: All jobs accessing a set of RECON data sets must allocate the same data set by the same DD name. For example, RECON1=dsn1, RECON2=dsn2, and RECON3=dsn3. Failure to follow this convention causes serious damage to the RECONs.

The API uses only one set of RECONs between a FUNC=STARTDBRC request and its associated FUNC=STOPDBRC request. Your program, however, can use multiple sets of RECONs by deallocating the RECONs after the FUNC=STOPDBRC request and then allocating a new set of RECONs before issuing the next FUNC=STARTDBRC request. In order for this to work, your program must dynamically allocate the RECONs.