Hi, could someone indicate to me on what my problem might be please. We are running natural on a unix server and I am experiencing the following: When we run a function in batch, with my main processing program ± 3900 lines big (thought it might be the size but even when creating include codes to make the size smaller, we still get the same problem), it sometimes executes normally and write all the reports etc. but if I run it again, it completes immediately without any errors and it writes my report headers in spq but there is no data or any error messages. When I keep on submitting the function then it does the same thing about 3 or 4 times (this varies) and then for no reason when submitting it again, then it works fine and I get the correct output in my reports and it completes normally. (These are dummy runs, no updates so it should pick up the same data every time) I can accept the fact that there could be a problem in the programs (ex. a stop command etc.) but if this is the case then why does it not show in blog that my process was completed? It only shows that my Job was started but it does not show that the job was completed after it bombed out of BTQ. This is all that blog shows when I encounter this error:
** 05/05/10 13:01 Job 8176 User MURRAY BM030 BAB1800Z Submitted in class B DB 3
** 05/05/10 13:01 Job 8176 User MURRAY BM030 Started on DB 3
And this is when the job has been completed correctly with output in my reports:
** 05/05/10 11:33 Job 8169 User MURRAY BM030 BAB1800Z Submitted in class B DB 3
** 05/05/10 11:33 Job 8169 User MURRAY BM030 Started on DB 3
** 05/05/10 11:40 Job 8169 User MURRAY BM030 Ended without errors
Surely blog must show me that the job ended no matter how the processing was stopped or terminated and if it was a problem in the program (stop etc.) then I must experience the exact same problem every time I run the program?
This is more likely a problem which should be referred to your support organization to get help investigating the problem.
Just a thought: being time-dependent, it is less likely to be a constant issue such as program size. More like a resource lock or contention - while the resource is in use the process terminates and perhaps cannot report either (which would tend to implicate the batch output file).
It’d be interesting to run, wait 121 seconds after completion, and then run again, and if it works, decrease the time to 31s, 5s.
Anthony, why 121 seconds? Why not 301? or 601?