DSXchange

Hi ,

Version : 8.0.1

OS : AIX

I have a job for which multiple instance is enabled .Its a generic job used to populate 100 tables .So , We have 100 invocation id's .After running couple of times , I am unable to see the datastage log in the designer .When I am clicking the folder containing the generic job , Its getting hanged and not retreiving the log .

I have even tried with dsjob -logdetail command .But unix prompt also not returning anything .

Can someone help me in understanding the following questions ?

1) Is there any limit for a multi instance job ?Because we are calling the same job for 100 times and after that we are unable to see the log ?

2) Behind the scenes , I suppose these logs are stored as a files (RT_LOGnnn ) .Then why is it taking that much time to retrieve information from files .If it is database tables then we can expect the delay .

Thanks a lot everyone in advance .

The logs are in database tables. And how long did you wait before you decided it was hung?

Thanks a lot craig for your reply .

If logs are present in database then what about RT_LOGnnn files ?

I am still waiting for the log .Its been 3 hours

Those are hashed files, a.k.a. database tables in the repository. And 3 hours is a long time. Can you view the log for a specific invocation of the job rather than the 'base' job?

If your log file has a lot of records then it is slow to retrieve not hung. A lot of instances may get cleared if you have autopurge set.

What are the values of RTLogging and ORLogging in the DSParams file in your project? You may need to use the Reporting Console for Information Server, rather than Director, to view the logs.

I thought that was something added in the 8.1 release and they were still out in the 'old' repository in 8.0.1.

8.0.1 doesn't allow one to change the logging behaviour.

Yes craig ..whatever andrew has mentioned is correct ...I remember reading that in one of the post ...

Today also i have tried to view the log ..but still giving the same problem ...

Is there any way to purge the log through command prompt ?? I cannot clear log through director as clicking that folder is creating the problem .

How to find the mapping of a job and its job id ? I believe, RT_LOG_nnn -- nnn is a internal job id . Could you please let me know how to find the value of nnn for a particular job ? I remember we have to query universe tables like ds_jobs or ds_objects .But i dont know the structure and exact query .If someone can help me in getting it , it would be great ..Thanks a lot ..

And there's only one underscore in the log table's name.

I tried to clear the log but administrator is getting hanged . Can I remove that RT_LOG21 log file from unix system ?

Only if you want a corrupted job.

Is there no way to clear a job in my case ?