DSD.SEQOpen Failed

[lbegley]

7.1 DataStage
Windows Server
Windows Client

Getting the following error, need assistance pointing us in the correct path to understand the error and how to fix.

Job keeps bombing out because it says it can't read a sequential file that is created in the job:

"UpUpdateT2CaseDetailUCDT..stgSqT2CD.lkOdbc: DSD.SEQOpen Failed to open UCDTUpdateT2CD.txt in directory D:\DWH\Interim_Files
STATUS = 13."

Thanks in advance.

[Sainath.Srinivasan]

Please note that Windows locks a file when it is open. So if you create and also access from it in the same program using different links, you may land in this locking problem.

Try to break into 2 jobs - one creating it and one reading from it.

[r_garza]

Please note that Windows locks a file when it is open. So if you create and also access from it in the same program using different links, you may land in this locking problem.

Try to break into 2 jobs - one creating it and one reading from it.

I am the developer that is having this problem....lbegley is our DS administrator.

The deal w/ this job is that it has been developed, unit tested and acceptance tested. Throughout these stages the job has run successfully numerous times....I mean alot! Now when I build a Job Control job to invoke this job, out of left field I get this error stating the file is not able to be opened. This also happened w/ another job....but I was able to get through this issue by Clearing Status File and Cleanup Resources (which I performed as a last resort, not expecting this to work). But when I perform these for this job, it still failed....

So, I am not so sure that the Windows locking is the issue (since the job has run successfully in the past). My concern is that this error occurs randomly (or perhaps to a pattern that we are not recognizing) for jobs that have a successful run history....

Now for this job, I was able to get around the problem...the sequential file in question was just an intermediate file between a Transform and an ODBC stage. I just eliminated the Sequential stage and updated the ODBC stage straight from the Transformer.

My concern is that we have lots of jobs that have gone through development and acceptance testing....this error may reoccur and we have precious little time in our project plan to make these changes should this random(?) error reoccur....if this is because of the Windows locking error, is there a way around this. I can't believe that we would be the lucky first to encounter this issue in the forum.... :D

Thanks for all input....

[Sainath.Srinivasan]

One reason could be that Windows system is not managing and releasing lock at proper intrevals.

Another reason can be than a job can fail leaving the locks behind and thus making the next job accessing these locked files to fail. Also this forces the second job to leave its locks behind.

That is why it is adviced to 'Reset Job' when they fail. As you have said, clearing resources can also get around this problem. But when you move your developed system into production, you may expect it to run with very little manual intervention.

Some common method than can be followed to reduce impact of these locks are
1.) Use Reset and Run for jobs in Job Sequencer so that any failed job can release their locks.
2.) Bounce DataStage and Windows server periodically to release locks
3.) Provide varying names for temporary/intermediate/staging files so to keep them unique and distinct.
4.) Code jobs to move files once created to some backup/archive dir so to work on fresh area each time.

[r_garza]

Thanks for all the Windows information, we appreciate your help!