Jobs getting killed in Master Sequence without any error

parag.s.27 · Post by **parag.s.27** » Mon Dec 15, 2008 12:37 am

Hi All,

In the IIS jobs we are facing a new problem for past few days. The jobs that were running fine till now are getting killed as if OS is killing the process.

The detailed log does not show any warning or fatal or any information but when we check the job status in the status window, it shows that job is aborted. But inside the log there is nothing.

Many times it gives a warning as: -

Code: Select all

Job Control process (pid <process number>) has failed.

but this warning is also not consistent. Also for few instances, the job and the sequence hangs for an indefinite time and even the status shown as running.

Can someone please help and let me know whether its an OS level error where the AIX is killing processes due to process limit overflow. Or is there a problem with IIS 8

ray.wurlod · Post by **ray.wurlod** » Mon Dec 15, 2008 3:11 pm

No way for us to know, because we can't monitor your AIX system. You (or your UNIX admin) must do that. How busy is the system?

parag.s.27 · Post by **parag.s.27** » Tue Dec 16, 2008 1:10 am

ray.wurlod wrote:No way for us to know, because we can't monitor your AIX system. You (or your UNIX admin) must do that. How busy is the system? ...

Actually that is the problem. The server is on client site and we are not having the access to check the OS level settings. The infrastructure team on Client Site is suggesting that there must be a problem with the way the IIS job is built ,and not the OS.

But there seems to be no problem with IIS job because the jobs that are getting killed are only used to load data in Orable tables using OCI bulk load. No other processing is done apart from loading the data. Can the amount of data be responsible for jobs getting killed? We have approx 40 million records getting loaded.

parag.s.27 · Post by **parag.s.27** » Tue Dec 16, 2008 11:36 pm

Can anyone please help in suggesting what can be the possible reason. We have even raised a PMR with IBM, but it is also pending.

What I had found that typically for records more than 3 million the jobs are having such problem. And it is also happening recently. Earlier it was not the case.