Running ADABAS 8.2 NATURAL 4.2 on DOS VSE.
We sporadically get a Batch NATURAL step to fail with NAT3254 subcode 6.
Bumped up CT from 60 to 120 as per Support suggestion.
However, we believe this is simply delaying the problem to another day.
We believe system is busy from another JOB and perhaps NATURAL JOB does not retrieve response in time.
Anyone else out there experience this? Any way to ensure one JOB doesn’t monopolize CPU.
Thanks.
Subcode 6 doesn’t really sound like a time issue, more like a storage overwrite
Thanks. I would agree, except bumping up CT does seem to eliminate at least temporarily the issue and CPU was very busy at that time.
Sorry, I didn’t want to say (or even imply) I disagree with support’s anaysis,
just found it a bit strange 
Anyway, re. your question, I assume support did also discuss partition priorities ?
Wolfgang. I appreciate your input. Yes it is a bit strange. I am guessing the AB entry gets released afte CT exceeded.
When ADABAS tries to return result it finds requestor id no longer matches.
W.r.t. priorities my System Programmer tells me VSE doesn’t time slice so even a lower priority JOB will monopolize CPU once it gets control.
“does not time slice” is partially true, VSE does time slice for “balanced” partitions,
between partitions not in the same balancing group there is no time slicing, but still
priorities DO matter - what rationale would there be for a priority concept if it didn’t
have any influence ?
Those who tried to prioritize POWER lower than other partitions didn’t do so for
more than a couple of minutes 8)
Generally speaking “service address spaces” should be higher than any partitions
requiring their service, putting Adabas at the end of the food chain, for example,
definitely WILL result in issues like that, when Adabas requests are fired at a
high rate and Adabas is prioritized low it just won’t be able to cope with the load
as it doesn’t receive control often enough to service all requests, they will queue
up and eventually time out.