Parallel job reports failure (code 139)

prajish_ap · Post by **prajish_ap** » Sat Aug 25, 2007 3:33 am

My job carries out a simple task of loading a sequential file from an ODBC Stage.
It is getting aborted giving the error message as

Contents of phantom output file =>
RT_SC256/OshExecuter.sh: line 20: 16908 Segmentation fault $APT_ORCHHOME/bin/osh "$@" -f $oshscript >$oshpipe 2>&1

Parallel job reports failure (code 139)

I have tried the stop and start the server option ...but that isn't helping solve the issue

Thanks in advance!

JoshGeorge · Post by **JoshGeorge** » Sat Aug 25, 2007 5:15 am

Segmentation fault can be because of many reasons like - running out of memory or there may be more than the program can handle ... For more refer THIS link.

prajish_ap · Post by **prajish_ap** » Sat Aug 25, 2007 6:30 am

JoshGeorge wrote:Segmentation fault can be because of many reasons like - running out of memory or there may be more than the program can handle ... For more refer THIS link.

The data being transferred here into the Sequential file is hardly 33MB in size...I don't think that the job is getting aborted because of running out of memory. Also, the job has been performing fine earlier with similar amount of data.

JoshGeorge · Post by **JoshGeorge** » Sat Aug 25, 2007 6:37 am

So what changed in your environment? Did you chek to confirm and then concluded this is not a memory issue but something else?

prajish_ap wrote:Also, the job has been performing fine earlier with similar amount of data.

prajish_ap · Post by **prajish_ap** » Sat Aug 25, 2007 7:16 am

JoshGeorge wrote:So what changed in your environment? Did you chek to confirm and then concluded this is not a memory issue but something else?

prajish_ap wrote:Also, the job has been performing fine earlier with similar amount of data.

No change has been made to the environement...

Could you please could you please explain what you mean by change in environment?

ray.wurlod · Post by **ray.wurlod** » Sat Aug 25, 2007 3:00 pm

ANYTHING that has changed. For example are you now running under a different user ID? Has the user's ulimit changed? What stage types are there in your job? What other tasks are happening at the same time? (These are just examples. You need to be the detective.)