IBM Connect:Direct Data Transmission Facility (DTF) Dumps
This dump is generated when an ABEND occurs. It contains the contents of the IBM® Connect:Direct® address space, which is copied into one or more of the data sets that you specify by ddname in the DTF JCL (CDSVCDMP, SYSABEND, or SYSMDUMP).
When an ABEND occurs, IBM Connect:Direct produces an SVC dump with all of the information for the address space, regardless of whether the ABEND occurred in the main task or a User Exit subtask. Unless otherwise specified in the JCL, the dump is written to the standard SVC dump data set, SYS1.DUMPxx. You can specify an alternate data set by using the CDSVCDMP DD statement in the JCL.
Turn off the SVC dump by modifying your JCL. For more information, see Turning Off the SVC Dump.
Multiple DTF SVC Dumps Capture Using CDSVCDMP
You can use CDSVCDMP to capture multiple DTF SVC dumps. A dump is created for each ABEND and written to a separate data set. You can specify whether the dumps are written to the SYS1.DUMPxx data sets or to unique user-specified data sets.
To write dumps to user-specified data sets, you must create a data set name for each dump that can occur. The first data set name must end with .SYSMDP00. Each additional data set name must end with .SYSMDPnn, where nn is a consecutive number up to a value of 99. For example, if you want to create enough data sets to write five dumps, create five data set names beginning with xxxx.yyyy.SYSMDP00 and ending with xxxx.yyyy.SYSMDP04.
You must define the data sets with the same attributes as your SYS1.DUMPxx data sets. The data sets must be preallocated and on the same disk volume.
To record multiple SVC dumps, set the JCL statement for CDSVCDMP as follows.
The first dump is written to the .SYSMDP00 data set. When an additional ABEND occurs, a dump is written to the next data set, .SYSMDP01. Each additional ABEND creates a dump to the next .SYSMDPnn data set as long as ABENDs occur and enough .SYSMDPnn data sets are available. If an ABEND occurs and all data sets are full, the dump is not created and a message is issued stating that all .SYSMDPnn data sets are full.
You do not have to empty or reset these dump data sets. When the DTF is initialized, and you are using the .SYSMDPnn data sets, IBM Connect:Direct writes over the existing data in the data sets. If you want to save the existing data, save the data sets using a different data set name before you restart IBM Connect:Direct.
Turning Off the SVC Dump
If you want to turn off the SVC dump, place the following statement in your DTF JCL.
Changing Dump Options
For the SDATA parameter specify SDATA=(ALLSDATA). If SDATA=(ALLSDATA) is not an acceptable default for your system, ensure that the SYS1.PARMLIB member corresponding to the ddname in the JCL for the DTF specifies, at a minimum, the following.
For the PDATA parameter for IEADMPxx and IEAABDxx, specify PDATA=(ALLPDATA).
If you cannot specify ALLPDATA, at a minimum, include PSW, REGS, SA, JPA, SPLS, and SUBTASKS. Refer to the IBM manual MVS Initialization and Tuning Reference for the release of z/OS being used.
If you are unable to change the SYS1.PARMLIB member, issue an operator command to change the dump options.
The following steps guide you in issuing operator commands to change dump options:
- Issue the command DISPLAY DUMP,OPTIONS to list the dump options currently in effect.
- Issue the command CHNGDUMP SET to change the options.
- Issue the CHNGDUMP DEL or CHNGDUMP RESET command to reset the options after recreating the dump.
If the correct dump options are specified in one of the SYS1.PARMLIB members, change the ddname in the JCL for the DTF to reference the ddname corresponding to that member.
For further information on changing dump options, refer to the MVS System Commands IBM manual for the z/OS being used.