Job Activities in parallel

tostay2003 · Post by **tostay2003** » Wed Jan 17, 2007 11:20 am

Hi,

I have a sequence with Job Activities (JA) as follows.

                        __ JA4___
                       /           \
JA1 -> JA2 -> JA3 -------   JA5 ----- -- JA7
                      \ __ JA6 ___/

I have set the sequence as restartable. This is working fine with the Job Activities in serial, but not with parallel.

i.e. lets say JA2 has failed, then when i rectify the error and run. The job starts from JA2 (skipping JA1).

But in case of job acitivies in parallel i.e. JA4 to JA6. When one gets aborted and I run the sequence again. It skips JA1... JA3, but JA4..JA6 start again irrespective of whether they ran successfully last time. How can I avoid the Job Activities which were successful not to run again (when they are in parallel).

Kirtikumar · Post by **Kirtikumar** » Wed Jan 17, 2007 11:31 pm

For the job activities j4 to j6, check if the property Do not checkpoint job run is not enabled.

tostay2003 · Post by **tostay2003** » Thu Jan 18, 2007 11:41 am

Hi,

No it is not enabled. Does DataStage deal with JobActivities in parallel differently than with JobActivities in Serial

Regards

chulett · Post by **chulett** » Thu Jan 18, 2007 11:45 am

No. We do something similar in many places and don't have this issue. Out of curiousity, what 7.x version are you running?

tostay2003 · Post by **tostay2003** » Thu Jan 18, 2007 12:24 pm

I am using 7.5.1 version

Kirtikumar · Post by **Kirtikumar** » Mon Jan 22, 2007 5:09 am

Some weird suggestion - copy all the stage in a new job and save it. Set the properties again for check pointing. Compile it again and run it now. See if same thing happens.

The suggestion here has come from solution we have had for few problems. Just creating new job or replacing stage with another instace of the same stage sometimes work.

If this does not help, check the logs in director for the last run as to if it contains the logs for check pointed activities.