I have a Master Job Sequence that calls other smaller sequences within the Job. The Master Job Sequence has the "Add checkpoints so sequence is restartable on failure" tick box ticked (as do all the smaller sequences).
When the job starts, 3 sequences and a cleanse job all kick off at the same time. One Sequence completes OK as does the Cleanse job which then feeds into a transform job. Only when all 3 sequences and the transform job complete OK does the job continue.
The problem is that one of the sequences aborts and so the Master Sequence aborts. When I then restart the Master Sequence, the transform job kicks off as expected (as the Cleanse job is skipped) but the two sequences that aborted on the previous run do not restart. It appears as though nothing will happen with them until the Transform job completes OK and then one of them (but not both) will restart.
Unless I am missing something, shouldn't all the jobs/sequences that aborted first time round, all restart at the same time ? Isn't that the point of checking the "Add checkpoints so sequence is restartable on failure" tick box ?
Any suggestions as why they seem to restart in sequence and not parallel ? It is an issue as the transform job takes a couple of hours to run and so time is being wasted.
If you restart a sequence more than once then it gets confused and does not remember if it has run a step successfully or not. Otherwise it works as advertised.
What version of DataStage are you on? There are bugs in this in anything before Ds7.1.
