IBM Connect:Direct DTF Out-of-Storage ABENDS
DTF out-of-storage ABENDS may occur during heavy IBM® Connect:Direct® activity or during phases when the DTF has run for a long period of time. This chapter explains the most common causes of DTF out-of-storage ABENDs, actions to take, and the types of data to collect.
Condition: Out-of-Storage ABEND Occurs in the DTF
The following table describes two possible causes and associated actions to take.
| Error | Cause | Action | Collect |
|---|---|---|---|
| user ABEND U0500 user ABEND U0501 |
If this condition occurs only during heavy IBM Connect:Direct activity, you may need to modify the initialization parameters or the DTF REGION parameter. | Review both the short text and long text messages. Limit the number of Processes that can run at one time using the MAXPRIMARY, MAXSECONDARY, and MAXPROCESS initialization parameters. Also, examine the MAXSTGIO initialization parameter to determine if it can be decreased. The REGION size in the DTF JCL may need to be increased to allow more concurrent Processes. |
|
| System ABEND Sx0A System ABEND Sx78 |
If this condition appears to be "storage creep" and occurs after the DTF is active for a long time (not necessarily running many Processes immediately), you can take several actions. |
|
|
Possible Actions
Use the following list to help you troubleshoot DTF out-of-storage ABENDs:
- Examine all RUNTASK programs; ensure that for every file opened, a CLOSE and a FREEPOOL is also done.
- Examine any user exits for GETMAIN (or STORAGE OBTAIN) macros ; verify that FREEMAIN (or STORAGE RELEASE) macros are issued for each of those areas.
- Examine any RUNTASK programs for GETMAIN (or STORAGE OBTAIN) macros; ensure FREEMAIN (or STORAGE RELEASE) macros are issued for each area.
- Try to pinpoint the type of Processes or other DTF activity that causes the problem:
- Does this occur only during a COPY?
- Does this occur when a specific Process is run? What does the Process do?
- Does this occur when a certain command is issued? What is the command?
Note: For the next two suggestions, you can issue an MVS DUMP
command for the IBM Connect:Direct for z/OS® address
space to help in your investigation. Make sure that the SDATA includes
the following parameters:
RGN,LSQA,SUM,PSA,GRSQ,SQA,SWA,TRT
See the IBM MVS System Commands manual for the release of z/OS in use.
- If you cannot determine which Process, command, or other activity is causing the storage creep, run a typical batch of Processes or commands that runs when the ABEND occurs. Before the out-of-storage ABEND occurs, go to the ADMIN MD panel and QUIESCE IBM Connect:Direct for z/OS. For example, if the ABEND usually occurs after 10 hours of activity, quiesce after about 8 hours.
- If you did determine that a certain Process or command causes the problem, submit that Process or issue the command several times (the number of times depends on how long it takes before you get to the ABEND). For example, if it occurs after the Process runs 100 times, run it 90 times in your tests. Get a dump of the DTF address space after all DTF activity is finished.