Parallel job reports failure (code 139)

[nishantrk]

Hi,

I went through all the related posts , which give me a understanding that the Parallel job reports failure (code 139) is more sort of generic message and there could be many reason for the error . However in my case am not getting other message like core dump / Segementation fault etc.
There is absolutly no other information in the log ...

We have 3 seq jobs each having 5 extract jobs (in total 15 jobs) which reads from oracle table and writes to file . The abort happens randomly on any job and works fine on restart.
Is there any variable which can help me Debug and pin point this issue ...
Since this is production system , would like to have a concrete thing before making any change..

[nishantrk]

These are the two confifuration parameters , which has been recommneded by others to change..
APT_MONITOR_SIZE = 50000
APT_MONITOR_TIME = 10

[priyadarshikunal]

try this link.

Also, it may relate to segmentation faults or some other issue. Do you get anything when you reset the aborted job? something like "From previous run"

[nishantrk]

yes have gone through that link before posting..we are not getting this type of error in any other develpoment enviroment...
Unforutunatel no other message except for the 139 error

[ray.wurlod]

What does your official support provider advise?

[PaulVL]

Are these multi instance jobs? Are they running concurrently?

[priyadarshikunal]

Did you try resetting it to check if something like "from previous run" appears in reset logs.

also try setting APT_DISABLE_COMBINATION (or something like that) to true for that job.