Error: IEA705I 878-10 MUF blocked or unresponsive in Datacom/DB

Products

• Datacom

Issue/Introduction

Datacom Multi-User Facility (MUF) becomes blocked or unresponsive when the system reports memory exhaustion. This typically occurs during high CPU consumption in connected environments like CICS, often accompanied by specific SQL return codes.

Possible symptoms:

• MUF becomes unresponsive to console commands (e.g., COMM STATUS).
• High CPU consumption for CICS and MUF address spaces.
• System message: IEA705I ERROR DURING GETMAIN SYS CODE = 878-10.
• Datacom error: DB RC=89(3) (Memory Parm error).
• Log messages indicating UNABLE TO ALLOCATE MEMORY in task areas.

Environment

Z/OS 

CICS Transaction Server  

Datacom/DB 15.1 

Cause

The MUF 31-bit memory limit (REGION size) is exhausted. This can be caused by a looping CICS transaction executing a high volume of SQL requests (e.g., millions of requests per second). In such cases, memory is allocated in the "JOB" memory pool, which is only freed when the job (CICS) ends. For long-running CICS regions, this leads to a permanent memory "hang" until the region is recycled.

Resolution

To prevent MUF outages due to memory exhaustion, follow these steps:

1. Increase the REGION size for the MUF started task
2. Implement the following parameters in DBMUFPRor via console command:
   • MEMORY_RESERVE_31BIT: Reserves memory for z/OS and MUF to ensure the system can still process critical functions during stress.
   • STOP_LOOP_HANG: Enables MUF to detect and mitigate potential hang situations.
3. Review CICS transaction rules (e.g., in SYSVIEW) to implement warnings or automatic cancellation for transactions exceeding reasonable SQL request limits.

Additional Information

For documentation , see MEMORY_RESERVE_31BIT and STOP_LOOP_HANG .