Error Encountered while running a job in a sequence.

RohanSharma · Post by **RohanSharma** » Mon Jan 29, 2007 1:17 am

Hello ,

I have sequenced a server job in a Sequence stage.
It used to run fine but on a day when I run it
it gave me the error
"Controller problem: Error calling DSRunJob(Step0019CreateHashParam), code=-2
[Job is not in the right state (compiled and not running)]"

Sorry for mentioning the exact job name here but had no other options .

What i dont understand in this eror message is the state "compiled and not running"

Please help me decipher the message properly.

Also I know that compiling the job shall resolve the problem but I want to know the exact reasons why this error occours and how does it occours ?

ArndW · Post by **ArndW** » Mon Jan 29, 2007 2:20 am

Welcome to DSXChange!

The message is saying that the sequence has tried to start a job, but it is in a state of not being "compiled" or it is currently running and thus cannot be started.

sud · Post by **sud** » Mon Jan 29, 2007 2:23 am

Usually it is advisable to always use the option 'Reset if required then run' in the sequence ! :D

RohanSharma · Post by **RohanSharma** » Mon Jan 29, 2007 3:46 am

To start with ...
thanks for that warm welcome ...ArndW (thats the name i can refer ..)

Yes i understand that the time when job was run it was not in a proper state.
But I m very sure that the job was neither running nor was in an aborted state.
It was i in a finished state .
Also i have clicked the option for a check point run which remembers every sequence completed by the main sequenceas a whole ..[does that have something to do with this ?]

Also as suggested I have already done a reset if required and run option in my job sequencer

ALso I have already solved the problem of running the job as soon as i compiled the job and the sequence

But the point I posted the post was to know what exactly Mr.Datastage7.5

Means when he says "Controller problem: Error calling DSRunJob(Step0019CreateHashParam), code=-2
[Job is not in the right state (compiled and not running)]"

I mean what does the state "compiled and not running" means ??

I hope i make my point here if I'm not able to make myself clear on any point ,please correct me .

Thanks for your time and valuable inputs

sud · Post by **sud** » Mon Jan 29, 2007 4:04 am

The only unusual point in this case is, as you mentioned, there were no problems while the job was running. I have faced this issue couple of times. Don't know about the exact cause or even the interpretation of the error. My feeling is that the process became a zombie, where it didn't update the system that it was over and hence the resources were still blocked by datastage.

ArndW · Post by **ArndW** » Mon Jan 29, 2007 4:14 am

The message is a double-negative and confusing, think of it as:

(a) Job's right state is "compiled and not running"
(b) Job is not in state (a)

Apart from a job being in an "aborted" state, I have most often seen this error when using multi-instance jobs and trying to start the same instance multiple times.

RohanSharma · Post by **RohanSharma** » Mon Jan 29, 2007 4:21 am

So we SUM up the topic ,
that the error that MR Datastage 7.5 throws is quiet ambigous and only solution is to compile the jobs and run the sequence again .As reset wont tell the system to update the previously updated compiled and not running status

Also I'm convinced with Mr SUD's Zombie concept and shall take that for an explanation of this error.

Thanks again gentlemen for your precious inputs and time

ArndW · Post by **ArndW** » Mon Jan 29, 2007 4:26 am

Rohan, I have to disagree with your conclusion, at least in part. The message is syntactically & grammatically correct but is quite confusing, even to native English speakers. But a "reset" of a job, either manual or via the "reset if required" switch in a sequence will allow it to be run without this error.

RohanSharma · Post by **RohanSharma** » Mon Jan 29, 2007 4:41 am

Sorry to differ MR.Arndw
But the reset is an option that I already have excised for my job.
Also the option of Reset if required then run is the one that I have already incorporated in the sequence.
Neither of the two helped me ,but as soon as I compiled the job it run fine ,thus I concluded that compilations of the sequence is a viable an option

ArndW · Post by **ArndW** » Mon Jan 29, 2007 5:07 am

Once you get this error message, what does the director show as the "state" of the job?

RohanSharma · Post by **RohanSharma** » Mon Jan 29, 2007 6:21 am

The job that is sequenced is in Finished state but the sequencer aborts any further execution complaining this error.
Remember this scenario happens once i have my previous run as aborted and I reset the aborted job.

eg A-->b-->c-->d [Concider a,b,c,d as jobs sequenced]

First Run --> a sucess-->b sucess-->c abort-->d not executed.

Now I reset the job c. and its status is set to reset.

Second run --> a sucess-->b sucess-->c abort-->d not executed.

Now in the sequencer's log I will find the message "Controller problem: Error calling DSRunJob(C), code=-2
[Job is not in the right state (compiled and not running)]" for the job c.

DSguru2B · Post by **DSguru2B** » Mon Jan 29, 2007 7:50 am

Choose the option of "Reset and then Run" inside your job activity for each job. This way if the job requires resetting, the job will be reset first and then run.

ArndW · Post by **ArndW** » Mon Jan 29, 2007 8:01 am

DSGuru2B - If I understood the posts correctly this is already being done, but I just noticed that checkpointing is being used as well.

DSguru2B · Post by **DSguru2B** » Mon Jan 29, 2007 8:04 am

I missed that part in the OP's next to last reply. Sowwiee
Then this is strange behaviour. I would request the OP to examine the log and see whats actually going on when it comes to that particular job. When it requests for a "job reset". Go into that individual job and see whats causing the abort thats rendering the job to sucha hard state.