DSXchange

I read thru the manual, but haven't gotten any conclusive answer to this:
I have a job with few stages in it. It processes quite a large volume of records. If the job fails after some interval, is there a feature in Datastage that would allow re-startability from 'that' point , and not from the beginning of the Job....???
PS: I'm not taking of restartability in a Job Sequence [say 5 jobs, and 3rd fails, and restarting from 3rd again....], but there's only one Job, and it fails mid-way, and how do we go about re-starting from 'where' it lefg?
Sorry if I have confused, but this is the best I can put forth.
Thanks
Vijay

You will have to include your Job in a JobSequence(create one) and then in the Job Properties of the Job Sequence you will find a property "Add checkpoints so sequence is restartable on failure". Check that and also explore about it in the manual. That is the starting point.

kris007 wrote:You will have to include your Job in a JobSequence(create one) and then in the Job Properties of the Job Sequence you will find a property "Add checkpoints so sequence is restartable on failure". Check that and also explore about it in the manual. That is the starting point.

Thanks Kris, But here's from the Manual, and it says for Job-Sequence. But how does it apply for a Single Job ???
Add checkpoints so sequence is restartable on failure.
Select this to enable restarting for this job sequence. If you have
enabled this feature on a project-wide basis in the DataStage
Administrator, this check box is selected by default when the
sequence is first created.
And I think I am answering my own question: Data Set stages should be used to create restart points in the
event that a job or sequence needs to be rerun. But, because data
sets are platform and configuration specific, they should not be
used for long-term backup and recovery of source data.

A sequence can be made restartable, but I doubt if a single job can be restartable, if it aborts mid-way.

Again it depends on what your job is performing....Has it done what you want it to do...

If your load has failed at a certain point, you could run your job again giving @INPUTROW > xxx as a constraint.

narasimha wrote:A sequence can be made restartable, but I doubt if a single job can be restartable, if it aborts mid-way.

Again it depends on what your job is performing....Has it done what you want it to do...

If your load has failed at a certain point, you could run your job again giving @INPUTROW > xxx as a constraint.

Thanks Narasimha
That's what was my understanding...I haven't given it a try, but a discussion on restartabilty popped up, and while discussing this was brought up and I'm researching within a single job, on how restartabilty is done...!! Any thoughts on this is much appreciated.

Hi
Narasimha wrote "If your load has failed at a certain point, you could run your job again giving @INPUTROW > xxx as a constraint" .
True it possible to run if know the value else no..Probably you have to write a routine for restartability but I really doubt whether it is possible or not .... trying for an

A job sequence can, of course, contain just a single Job Activity. The run method of the job can be set to "reset if required, then run". Of course, the automatic restart capabilities of that sequence (skip earlier successful activities) is wasted, but you would still, nonetheless (unless you caused the job sequence itself to abort) have a restartable job sequence.

ray.wurlod wrote:A job sequence can, of course, contain just a single Job Activity. The run method of the job can be set to "reset if required, then run". Of course, the automatic restart capabilities of that sequen ...

Ray,
Thanks for the response...But I couldonly read part of your response !?!?

vijayrc wrote:Ray,
Thanks for the response...But I couldonly read part of your response !?!?

You have to be charter member to read fully.

ray, I just have a question. The OP has one job in his sequencer. So are you saying that there is no way to restart from the point the job aborted? Let's say, his job inserts 10million rows into a table usually. However, during one run, after 5 million rows, the job aborts. How would he restart after 5 million rows? Is it possible? What if he had 2 jobs in his sequencer and this was the second job?

Certainly possible. But nothing out of the box other than commiting at all or none. This methodology has been discussed here before. Try a search.
As for the two jobs in the sequence job and this being the second job, that is the basic restartability provided by the sequence job. Its certainly achievable by a mere couple of clicks in the job properties.

As noted, the Sequence job only knows which activity to restart with. If you need a job to restart internally from 'the point where it left off' you need to build that functionality into the job itself.

I know how I've done it in Server jobs, PX jobs on the other hand... [shrug]