Page 1 of 1

Posted: Thu Jul 20, 2006 1:03 am
by kumar_s
Though it seems that the first always get executed first, it may not be sure which will end first. Because both were constructed paralle.
Since the generated code should be in sequence, the first one may seems to kick of first, but the execution will not be same as the start.
If you want to maintain the order desperatly, split the job or redesign in sequence.

Posted: Thu Jul 20, 2006 1:04 am
by ArndW
In both server and px jobs your job design would have two independant streams kicked off. If you've found that one gets executed before another it could only be by milliseconds (on a fast machine); for all intents and purposes the OraStage1 and OraStage2 will get started in parallel. In an EE job you will also get {n} concurrent processes kicked off for each stream depending on your apt_config_file node settings at runtime (and also dependant upon your Oracle database partitioning).

Posted: Thu Jul 20, 2006 1:18 am
by williamswe
Thanks for amost instant replies :o

After monitoring the job run for a few rounds....i found that the order is almost random.

thx for explanation.