DSXchange

Posted: **Fri May 31, 2013 4:56 am**

Hi All,

I am executing a datastage parallel job with the below mentioned design.

Design:
ODBC --- SequentialFile.

The job is designed to send JMS requests to the Queue and is a long running job which will run more than 2 days.

After being executing for 5 days the job had aborted with the below mentioned error.

Error:
Cannot read osh phantom output file 'SH_16584_17606_694957', status '0'
Cannot read osh monitor phantom output file 'DSD.OshMonitor_16584_17607_259986', status '0'

But the load on the job had increased from 3 million to 12 million and the duration of execution went way beyond 2 days.

So request to please go through the issue and let me know if there is any possible solutions for the same.

Thanks In Advance.

Regards,
Sandeep

Posted: **Fri May 31, 2013 5:53 am**

At what time did the job abort? Something round, like exactly at midnight or noon? I don't believe that DataStage is doing this, but that the abort of the job is caused by some other process deleting the output file.

Posted: **Fri May 31, 2013 11:58 am**

Thanks ArndW for your response and please find the below mentioned details as requested.

Details:

Job Start DateTime: 2013-05-27 4:53:24 AM
Job End DateTime: 2013-05-30 5:14:39 AM (Warning Time)

I checked the existence of target file in the target directory and is available, so I confirm you that the source file is not deleted.

Please check and do let me know if you are looking for more information from my end.

I am continuing my analysis will update you if there is any breakthrough.

Thanks & Regards,
Sandeep.

Posted: **Fri May 31, 2013 12:29 pm**

in these cases, first thing to check will we whether anyone or any other process cleared scratch space. tmp folder etc.

Posted: **Sat Jun 01, 2013 2:27 am**

Has this happened several times, each time with an abort after just a bit more than 3 days?

Posted: **Wed Aug 14, 2013 6:35 am**

Hi All,

Thanks for all responses.

We just re-ran the job and is successful.

Posted: **Sun Apr 20, 2014 6:48 am**

I got the same error this morning. I was running 8 way multi-process, and all 8 came up with the warning as the last event in the log file. The data loaded ok, but my sequence failed as it did not get a "Executed OK" return. DS 8.1.0.0 no patches on SunOS 5.10. Scratch and Swap space fine. I believe this may have something to do with the JobMon as indicated by DSD.OshMonitor message. I don't believe I can replicate this easily. My best guess is a network or WAS issue.