- Feb
- 1
- 0
We upgraded the 4NT to 8.02.105 since October 4, 2008 and reinstalled again on Feb 22, 2009. I have the following findings and would like to seek your comment or confirmation:-
1. As we searched through the application and system event logs in Windows, It looks like the 4NT shell is not abending and we suspect that the skipping of remaining 4NT job steps might be caused by the memory holding the remaining job steps were messed up for some reasons. As a result, 4NT can still go back to it's parent CMD with '0' exit code as if it is normal completion (PS we tweak this case zero exit code to 8888 to allow Control_M to report this error case) .
2. As we have a lot of jobs running concurrently, once a job has the problem, those others running will have the same problems.
3. Searching through the possible interactions among the jobs, I found that there are parameters at 4NT.INI, we set as below and use the default (e.g. DIRHISTORY is 1024).
LocalAliases=Yes
LocalDirHistory=No
LocalFunctions=No
LocalHistory=No
I would like to your confirmation of the possibility that the setting with "No" (i.e shared memory) to the above parameters will cause the problem if the memory buffer allocated overflows by globally sharing the memory? The overflow, say may be we have a great number job steps and concurrent jobs.
Do you think we should set those to "YES"? or are there any other memory sharing scenarios that the jobs have interfering the others?
1. As we searched through the application and system event logs in Windows, It looks like the 4NT shell is not abending and we suspect that the skipping of remaining 4NT job steps might be caused by the memory holding the remaining job steps were messed up for some reasons. As a result, 4NT can still go back to it's parent CMD with '0' exit code as if it is normal completion (PS we tweak this case zero exit code to 8888 to allow Control_M to report this error case) .
2. As we have a lot of jobs running concurrently, once a job has the problem, those others running will have the same problems.
3. Searching through the possible interactions among the jobs, I found that there are parameters at 4NT.INI, we set as below and use the default (e.g. DIRHISTORY is 1024).
LocalAliases=Yes
LocalDirHistory=No
LocalFunctions=No
LocalHistory=No
I would like to your confirmation of the possibility that the setting with "No" (i.e shared memory) to the above parameters will cause the problem if the memory buffer allocated overflows by globally sharing the memory? The overflow, say may be we have a great number job steps and concurrent jobs.
Do you think we should set those to "YES"? or are there any other memory sharing scenarios that the jobs have interfering the others?