Automatic reset of Parallel Job

dxk9 · Post by **dxk9** » Wed Aug 19, 2009 2:42 am

Hi,
I need to know if a aborted job can be reset automatically via script/reset job(if any).

I have a sequence scheduled to run daily. But certain jobs aborts due to lack of resources and once its reset and the sequence is re-started, it runs fine.

Since the abort is frequent, I need to know if I can reset the job automatically via a script. Also let me know if I can access the status of the job automatically.

Thanks in advance,

Divya

Klaus Schaefer · Post by **Klaus Schaefer** » Wed Aug 19, 2009 3:07 am

If you're using a sequence anyhow, you don't need a script to achieve this. In the job-activity simply set the "Execution action" to "Reset if required, then run" from the drop down list...

Klaus

dxk9 · Post by **dxk9** » Wed Aug 19, 2009 3:25 am

Where is the job-activity option??

Thanks in advance,

Divya

miwinter · Post by **miwinter** » Wed Aug 19, 2009 3:33 am

It's a sequence stage.

dxk9 · Post by **dxk9** » Wed Aug 19, 2009 3:40 am

Thanks for the prompt response

I understand that this option will reset the job(already in Abort state) and then run it if necessary, but if the job get aborted while running in the sequence, will be try to reset it and continue the sequence with the this job run??

Thanks in advance,

Divya

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Wed Aug 19, 2009 4:07 am

You may have to break the job into smaller units or increase your "resource" availibility.

dxk9 · Post by **dxk9** » Wed Aug 19, 2009 4:21 am

Breaking into smaller jobs is not possible as the job itself is a modular one. Increasing the resource availability is not possible as we do not have access and at the server side, they are not ready to increase the resouce size.

Regards,

Divya

miwinter · Post by **miwinter** » Wed Aug 19, 2009 4:28 am

Coming at it from another angle, what are the resources that it fails due to a lack of?

dxk9 · Post by **dxk9** » Wed Aug 19, 2009 4:31 am

We get error related to "SIGINT" or "SIGKILL". Occasionally we get errors such as "output file full" or "scratch fill full".

Regards,
Divya

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Wed Aug 19, 2009 4:35 am

You need to provide
a.) job design
b.) resource available
c.) error messages received - unedited

for anyone to assit.

Scratch or Resource disk being full may be due to several reasons.

chulett · Post by **chulett** » Wed Aug 19, 2009 7:21 am

"Since the abort is frequent" I too would suggest that simply automating the restart is not the answer but rather you need to rethink the Sequence job to run fewer jobs at the same time so it doesn't abort.

dxk9 · Post by **dxk9** » Wed Aug 19, 2009 10:04 pm

Here are some of the frequent errros which we face:

a.Fatal Error: waitForWriteSignal(): Premature EOF on node etlprd3 No such file or directory

b.main_program: ORCHESTRATE step execution terminating due to SIGINT

c.Fatal Error: Tsort merger aborting: Scratch space full

d.Fatal Error: Unable to allocate communication resources

All the above errors aborts the job. But once the job is reset and run, its running fine.

Regards,
Divya

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Thu Aug 20, 2009 2:31 am

dxk9 wrote:Here are some of the frequent errros which we face:

c.Fatal Error: Tsort merger aborting: Scratch space full

Did you try pre-sorting your sources ?

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Thu Aug 20, 2009 2:32 am

dxk9 wrote:Here are some of the frequent errros which we face:

c.Fatal Error: Tsort merger aborting: Scratch space full

Did you try pre-sorting your sources ?

dxk9 · Post by **dxk9** » Thu Aug 20, 2009 3:10 am

No, I dont do any pre-sorting. The job is something like this,

sequential file 1 ---> Oracle Enterprise 1
Change_capture--->Filter--->
sequential file 2 ---> Oracle Enterprise 2

I dont know where exactly I get the error. Mostly it occurs even before the import from input files are done.

Regards,
Divya