Job Fails In Between

[dssiddu]

Here i m coming with some Query's which we have already discussed over phone. Pls expalin me in detail and send me if u have any Doc on this, which has prepared in ur projects.
Pls send me as soon as possible.
I have 5 batches, and all the Jobs are running in Sequential order.
Assume each batch is having 10 jobs.
Queries:
1)First job in the Batch 3 got failed (After success full runing of 2 batches) , so wht way i can handle the data.
I mean wht abt the data which is already loaded into the target by the previous successful ran 2 batches.
2)wht's the situation if the 4th Job got fails in the Btch 3.(Batch 3 contains 10 jobs).
I mean, do i need to run from sequence 3 manually or do i need to run from the failed job i.e 4th Job.
How do we handle these Situations in DS.
Also, if the Batch fails in between, do we need to schedule the jobs again (Using waht ever the Scheduling tool using)or we need to run the Jobs manually from the jobs where it got failed.
And is there any Routine , Function or Command in the DS to know the Status of the Job like weather a particular job
is in Running state or not. As of Knowledge, i DSJ.JOBSTATUS command is there. But i want any routine of function which tell the Running state of the Job.

[sasi_kumarreddy]

Hi,

In the datastage administrator we have an option called automatically handle fail jobs.. check this option in the administrator. this will automatically handles u r jobs that r failed.

And some more information is there will be triggers in the job sequences..
you have to select them..each trigger has its own significance.. like conditional 'OK'..etc.. you have to design your job sequence according to your requirements.

Experts Please currect me if i am wrong at any place.

[kcbland]

The API DSGetJobInfo documented in your manual will tell you the state of a job when used appropriately. As your your creation of jobstreams and restartability, you'll need to learn more about restartability in Sequencers.

As for partially loaded data, you're going to need to look at your loading strategy and decide if your strategy supports recoverying when there is an error or issue. Some questions about your design would be:
1. Does it allow data to be backed out easily during a failed load?
2. Does it allow processing to continue after a non-transformation related issue (tablespace, disk space, etc)?
3. Do you have procedures in place to bring a failed load back to life?

Part of designing an ETL application is figuring these things out.