Identifying the data set to contain the dump

By default, MVS™ places the dump you request on a SYS1.DUMPnn data set. However, you can code the DCB parameter on SDUMPX to specify the data set that you want MVS to use. Because SVC dumps include data from key 0 storage, the data set you use should be made secure. If you use the DCB parameter to specify a data set, your program:

Opens the data set
Issues SDUMPX with the DCB parameter and, possibly, the ECB or SRB parameter
Closes the data set.

If your program uses the DCB and ECB parameters, MVS notifies your program when the write phase completes by posting the ECB. Then your program can close the dump data set. If your program uses the DCB and SRB parameters, MVS schedules the SRB when the write phase completes. As part of its processing, the SRB routine can initiate the process to close the dump data set that your program opened.

If you use SDUMPX to request a scheduled SVC dump, IBM® recommends that you code DCB=dcb addr with either ECB=(ecb addr,WRITE) or SRB=(srb addr,WRITE) to have MVS post the ECB or schedule the SRB when the write phase of the dump completes. Whenever you code the DCB parameter, MVS automatically posts your program's ECB or schedules its SRB after the write phase completes, but coding the WRITE option makes your source code easier for programmers to understand.

If you use the SDUMP macro to request an SVC dump, code DCB=dcb addr with either ECB=ecb addr or SRB=srb addr. SDUMP does not recognize the WRITE option. However, MVS automatically posts your program's ECB or schedules its SRB after the write phase completes.