Error for DB2 job

hrthomson · Post by **hrthomson** » Mon Dec 19, 2011 8:54 am

Hello

I'm running into an error in my production environment that I'm not getting in any other environment. The job runs a couple of SQL statements using the DB2 API stage. The first stage gets the latest records from a table and the 2nd stage deletes from another table where the keys match. The errors I'm getting are very generic:

Failure during execution of operator logic.
node_node2: Player 2 terminated unexpectedly.
node_node1: Player 1 terminated unexpectedly.
Transformer_1,2: Fatal Error: waitForWriteSignal(): Premature EOF on node EC07P502 No such file or directory

All of these errors appear several times in the log before the job aborts. There aren't any additional error messages.

I have set $APT_DISABLE_COMBINATION to True.

Could this be a permissions issue even though we're not getting a permissions error returned in our log?

Thanks in advance,
Heather

pandeesh · Post by **pandeesh** » Mon Dec 19, 2011 9:17 am

Can you run the job in single node and check?

hrthomson · Post by **hrthomson** » Mon Dec 19, 2011 10:01 am

pandeesh wrote:Can you run the job in single node and check?

I just ran the job with single node and I get the exact same errors. No more detail given.

pandeesh · Post by **pandeesh** » Mon Dec 19, 2011 10:05 am

have you checked your scratch disk and other resources space availability?

hrthomson · Post by **hrthomson** » Mon Dec 19, 2011 10:06 am

The job fails in seconds - there is plenty of space available on the server.

BI-RMA · Post by **BI-RMA** » Mon Dec 19, 2011 10:17 am

Could this be a permissions-issue? Yes, it could. "No such file or directory" signifies that Your user can't see the file, not necessarily that it is not there. So it may just as well be that You haven't got read permissions on the file or access to the path where the file is located.

hrthomson · Post by **hrthomson** » Mon Dec 19, 2011 1:04 pm

I'm not reading or writing any files; I'm reading from one db table and deleting from another. The account running the DS job has DELETE privileges on the table.

ray.wurlod · Post by **ray.wurlod** » Mon Dec 19, 2011 3:03 pm

Your configuration file might have the pathname of scratchdisk directory misspelled. This is one example of something that could cause this symptom.

pandeesh · Post by **pandeesh** » Mon Dec 19, 2011 9:22 pm

For my question, he replied enough space is available.
I dont know where he checked.
May be in some other path rather than scratch disk directory.

qt_ky · Post by **qt_ky** » Mon Dec 19, 2011 9:40 pm

In other cases, I have seen the message "No such file or directory" indicates the file is not found but if the file exists with incorrect permissions then the error will instead indicate a permission problem. Still, it could be permissions.

Is this the first time this job has run in production or did it just recently start to abort?

It may help to open a support case if this is the first time seeing this error.

hrthomson · Post by **hrthomson** » Thu Dec 22, 2011 4:34 pm

Thanks for all the replies. I did notice in the job that the config file property was pointing to our UT environment even though the parameter had been set to PR and the job recompiled. I opted to import the job again and reset the config file property. That seemed to do the trick.

Must have been a corrupt job. I imported XML instead of DSX too...not sure it made the difference.

cheers,
Heather

ray.wurlod · Post by **ray.wurlod** » Thu Dec 22, 2011 4:40 pm

When you import from XML it's a two-stage process under the covers - first the XML is converted into DSX then a "real proper" import is done.