We have a batch job that occasionally fails with response code 9 sub-code 15 (pending Work area overflow).
However, the program does an ET for every 5 updates so it’s units of work are very small.
I’m guessing another ET user is “holding the tail end” of the protection area.
Is there a way to discover the oldest outstanding UOW when this happens?
The program normally re-runs successfully by the time the application team have been called to resolve this, so I’m guessing by this time the other UOW has completed.
I increased the size of the Work data set and the protection area the first few times this happened, but it has started happening again months later.
What version of Adabas are you on? All the latest fixes?
You indicate that it is the same batch job that is failing with this issue, so unless the same jobs are running against it all the time, it is likely that the failing job is causing the problem. Have you examined the command log for commands associated with the job to verify that the program is indeed issuing an ET for every 5 updates? How many records are being put on hold for these 5 updates?
Can you post the full program here? And/or the command log trace?
Hi Douglas, thanks for the reply.
We don’t currently have command logging turned on. I will investigate turning it on, although this is a fairly busy database.
Because we don’t use command logging I’m not familiar with handling the command log data sets. Which utility would I use to print these records?