Fast Unload for Db2 for z/OS (PFU) unload jobs are generating the following messages and abending:
WER999A GE8REP34,STEP0020, - UNSUCCESSFUL SORT 878 S REASON=00000010 ..........
IDI0055E Fault Analyzer processing excluded because 8 meg of 31 bit storage is not currently available
IDI0135E Recovery fault recording terminating. Severe private storage shortage.
IEA995I SYMPTOM DUMP OUTPUT 494
SYSTEM COMPLETION CODE=878 REASON CODE=00000010
The suggestion is to change the REGION parm to 0M as suggested in article:
S878 abend in sort executing Fast Unload with ORDER BY syntax
Is there anything that can be done system wide to make this perform as it did before we applied the maintenance
so we do not need to change the JCL for hundreds of jobs immediately?
Release : 20.0
Component : Fast Unload for Db2 for Z/OS
One circumvention is to add a second line to the hlq.CDBAPARM(PFU) member as below:
SYSADM SETUP
ZIIP NO <---------------------
When the Fast Unload job is executed the message below will be seen:
PFU0002I - PROCESSING OF CA PARMLIB MEMBER "PFU" - BEGINNING
SYSADM SETUP
ZIIP NO
Note that this is just a circumvention until the JCL's are changed to provide a REGION specification within the jobcard.
The change to the PFU member will impact all users of PFU on that installation including any users operating on
other subsystems that are using the same product libraries too.
In the job sysout the following message may also be seen if Fast Unload r20 PTF LU08255 is implemented:
PUT0911I - Overriding small region
This means that Fast Unload has identified that the region size was too small to run the utility.
PFU has changed the region size (REGION=nM) to one of the following minimum values based on the storage area:
6M for storage below the line, 100M for storage above the line
(The above values may change in the future, check the online error code documentation page above)