How can you respond to SOS conditions in MVS unallocated storage

When you receive DFHSM0144W or DFHSM0149W messages, you should take actions to resolve the short on storage (SOS) conditions in user region (24-bit) or extended user region (31-bit) storage.

Consider taking any of the following initial actions:

  • Route work away from the affected region.
  • Create a CICS® policy or automation to set the z/OS® WLM Health value to 0.
  • Cancel waiting tasks, or wait for the tasks to time out.
  • Reduce the amount of work that can run in the affected region.
  • Reduce the number of open TCBs within the affected region.
  • Shut down JVM servers.

The following messages are issued when the SOS event is resolved:

  • DFHSM0145I The CICS region is no longer short on 24-bit MVS unallocated storage.
  • DFHSM0150I The CICS region is no longer short on 31-bit MVS unallocated storage.

Best practices

The default SOS threshold values should work for most users, where the main usage of 24-bit MVS storage is for TCBs.

Initially you can disable the SOS wait feature so tasks will not wait if an SOS condition is detected, and just monitor with the default thresholds to determine whether any potential 24-bit or 31-bit storage problems exist.

If there are no storage problems, you can try setting the SOS thresholds to very low values and re-enabling the SOS wait feature. This means that tasks will normally not be suspended as CICS is not likely to go SOS. If an SOS event does occur, CICS reacts by suspending tasks, preventing them from consuming any additional 24-bit or 31-bit storage.

For instructions on how to disable or enable the SOS wait feature, see Setting up the CICS SOS wait feature.

You should be continuously monitoring 24-bit and 31-bit storage statistics as well as console messages to avoid possible SOS situations. If the statistics show that the CICS region is getting close to any of the SOS thresholds, take steps to avoid the CICS address space running out of storage.

Figure 1. Sample DFHSTUP report on MVS storage

MVS Storage                                                                                                    
___________                                                                                                    
                                                          User region          Extended user region            
                                                 __________________________________________________            
Last monitor sample time.......................: 02/05/2020  16:39:53          02/05/2020  16:39:53            
State..........................................:               NORMAL                        NORMAL            
Current unallocated total......................:                 832K                         6680K            
LWM unallocated total..........................:                  28K                         6680K            
Current unallocated largest contiguous area....:                 804K                         6260K            
LWM unallocated largest contiguous area........:                  28K                         6260K            
Last date and time SOS.........................: 02/05/2020  16:27:57                                          
Current tasks waiting because SOS..............:                    0                             0            
Peak tasks waiting because SOS.................:                    1                             0            
Total waits because SOS........................:                    2                             0            
Time tasks waited because SOS..................:        00:03:08.5650                 00:00:00.0000

Depending on your CICS configuration, you might have numerous options to prevent storage problems. For example, you can increase the amount of 24-bit or 31-bit storage that is available to the CICS address space. If this option is not possible, consider using a clone region to which some workload of the affected CICS region can be transferred.