I am getting DC201006 and DC027007 error messages with a PAKS abend code. What can cause a PAKS abend?

Document ID : KB000011007
Last Modified Date : 14/02/2018
Show Technical Document Details
Question:

The following error messages are issued for either a CICS task or for a batch job running against CV. The messages indicate there was a PAKS abend code.

DC201006 Vnn Tnn CV-Status BE-TaskID Pri FE - ID1 FE - ID2 FE
DC201006 Vnn20 T1 ABRT PAKS nnnnnn 100 nnnnnnnn nnnnnnnn nnnn
DC027007 Vnn Tnn TASK:xxxxxxxx PROG: xxxxxxxx ABENDED WITH CODE PAKS

What can cause a PAKS abend?

 
Answer:

The PAKS abend occurs when IDMS tries to allocate storage from the z/OS ECSA (Extended Common Storage Area) for a data packet and the request failed. This is the storage area used by IDMS to allocate the control blocks needed for communicating with batch jobs and CICS tasks.

A PAKS abend code can be caused by the following:

  • A storage shortage in the ECSA. Use any z/OS monitoring tool to monitor the size of the ECSA and increase the value if necessary. The amount of ECSA in z/OS is specified in the second value of the CSA parameter in SYS1.PAMRLIB member IEASYSxx.

  • Fragmentation of the ECSA. This can happen if z/OS has not been IPLed for a long time.

  • The packet size requested exceeds the z/OS limit of 32K (32768 bytes). We calculate the packet size based on the subschema being used with the following formula:
    PACKET SIZE=(Largest record data length in the subschema) + (Number of record types in the subschema x 24) + 200
    If the packet size is over 32k then you have to tailor the subschema to reduce the size.

  • If using LRF records, CICS run units can receive PAKS abends when accessing very large logical records. This is resolved by implementing OPT00177 in RHDCOPTF.