DSXchange

Hi,

Parallel job is hanging while excuteing through master sequence

Parallel job is attached in one sequence and that sequence is attached along with 6 more sequences in one master sequence and we are calling master sequence from unix script. it is triggering all sub sequences and all are finishing except only one sequence that is hanging daily, is we trigger that manually it is working fine , can any one share their experince in this type of job hanging problems

It may be due to server overload. But not constantly on particular job. What is the job design? Does it have any dependency on other jobs to get finishes?
How many Jobs were running parallely?

remaining are also same design, all are finishing there is no dependency thats why we are triggering all 7 seq's parallely but only this sub seq is hanging. it's not triggering inside jobs but remaining sub seq's are doing fine. all 7 are same design and same approach. only this sub seq is not trggring inside jobs.......just it's hanging it is triggering first job that is not giving any log information just it's hanging. it's running ever

Is it any multiple instance job or something? Is that the only particular job hangs every time? What does that job do?

It would also be helpful if you could define "hanging" in this case. Does the job monitor show any rows processed or what is the status? Are the UNIX processes still running when this happens and is there any load on the system? If it accesses a database, are the connections open? How long do you wait? And (importantly) how do you stop the job, is it enough to issue a stop request from the Director or do you do something directly fro UNIX?

master seq is triggering that subseq and as per design that subseq should trigger first one parallel job, but it's not doing that allways it's in running state

that job is same as others it's reading data from oracle loading into dataset, the main problem is calling any jobs through that subseq it's hanging if I chnaged that order also it's hanging, normally if parallel job is hanging like this scenarios is there any setting change required or how can find that problem or any process ID'a problem

pls share your ideas

Thanks in advance

If a particular job is hanging always, you will need to focus on that particular sequence and its jobs. If you say, its about reading from a table and writing to a dataset, there might be a case where that table might be locked by some other jobs or process which been executed parallely. And when this job executed alone, the lock might be present and hence the might run successfully.

Thanks Andrew
just subseq is triggering that first job as per order of design and that job running ever but it's not giving any log information just it's giving 1 line of log in director even it's not giving that starting of job info also. we are not getting anything more then that but status is running allways. server
Load is ok, B.coz at same time if we trigger big seq's also they rae working fine. connection and all are ok,b.coz remaing all jobs are working under same schema and if we trigger that manually it's working fine.
And to stop that stop request is not enough and we are clearing status fiel in director in that way we are stoping that. but unix process is still running
Or some times with stop command from director more then 2 times are soemthing it's aborting. or some times clear status is stoping that job.

pls help us

Thanks in Advance

As you, yourself mentioned in your post, clear the status file will just change the status from running to compile, but the process will continue to run in background.
Even that might be one of the reason, where you previous process of the job might be still running, while you might be trying to execute the same job again from Master Sequence.

As you, yourself mentioned in your post, clear the status file will just change the status from running to compile, but the process will continue to run in background.
Even that might be one of the reason, where you previous process of the job might be still running, while you might be trying to execute the same job again from Master Sequence.