Job Aborts with wierd Error

[salilkelkar]

Hi All,

I have a simple Parallel Job (7.5x) which has the following.

1) sff stage
2) OCI 9i stage
3) a join stage which inner joins data from (1) and (2).
4) Another OCI 9i stage which upserts into a different Oracle table.

The job ran successfully in Development and Stage . I also got my UAT done.
When this job got migrated to Production it aborts however, with the following error.

ociRiskChangeProviderContract,1: connect failed for env: EDWPROD, user: etl_user, password: *****
Oracle error: -12154
ORA-12154: TNS:could not resolve service name.

i know what it means, however i checked the server name and everything and it is correct.(tnsping works , logged in also though it isnt the user name/paswd problem)
If i remove the second OCI stage and put the output in a flat file and run it again without changing anything it runs perfectly fine.

I have another job but with a single OCI stage and that works fine too.

both the OCI stages are using the same parameters which are passed by the sequencer ( i double checked that).

The only difference between the three enviornments is the config file has 6 nodes in prod while only 2 in dev.
Secondly dev and stage env have a SID while production has a Service_name .

I am
Need help
Thanxs
-S

[s1kaasam]

Check if your TNS_ADMIN ( which ever you assign to your tnsnames.ora file directory or file) variable is pointing to the right directory where you have the tnsnames.ora file.This ora file while drive the conectivety to the database.Go through your errors in the director carefully i.e. when your job is started what are the variables initialised and to what values.

In that way you could debug the problem.

Let me know if you have any questions.

[salilkelkar]

Checked that ...is fine.
Also All variables are getting assigned properly.

Also talking to ascential support.
They said that there is a problem with ordering (onames,tnames) in sqlnet.ora .
I changed that and tried however it still aborts.

We a shell script which initializes all the parameters.
It is the same script in all the three envs. Also as i had mentioned if i remove one OCI stage it works so the parameters are assigned properly.

Thxs
-S

[ray.wurlod]

Is your ini shell script invoked from $DSHOME/dsenv? This is where DataStage processes get their environment variables from.

[salilkelkar]

Is your ini shell script invoked from $DSHOME/dsenv? This is where DataStage processes get their environment variables from.

No it is invoked from a different dir
/edw/bin
The script sets the parameters of the sequencer correctly.
I can see that from the log entries.

this script calls the setenv.ksh script which does that.

thxs
-S

[salilkelkar]

OK,

I believe i got it ...

the maxuproc limit was set to 128 .(Maximum number of PROCESSES allowed per user) with 64 process per node for a 2 node system its ok.
however for a 6 node system it should be 384.
I guess it a configuration file problem.
now that the limits have been increased its working...

However i cant understand one thing. I have another job with a single OCI stage and that used to work fine on the 6 node config file..
Both the jobs used the same config file then how come it didn't abort ??

Thanks for all the help.
Any further light on the matter will be def appreciated.

Thanks
-S