Sequence - Restartability - Warnings - Status

[dwblore]

Hi

We have a rather unique problem:
The design of our restartable suquence is such that if any one of the jobs finishes with a status other than "OK" control branches to a routineactivity (that updates the control table for the sequece) and then a terminator activity aborting the sequence.

The problem is on a rerun - since this job had finished with warnings - the checkpointing does not cause it to be reset and run - resulting in the termination brach to be executed every time without any further progression.

Is there any way to alter the status/checkpointing of the job that finishes with warnings - so that on the next run this job is executed again.

Simply applying the limits for warning messages to 1 does not help since the server is unable to respond to the "Issuing abort after 1 warnings logged." in time and change the status of the job to abort accordingly.

I am aware that the sequence generates code like %%Status%% etc - can we use these to manually alter the checkpointing state/status through code for the job in question.If so how?

Thanks in advance

[chulett]

No, not that I've seen yet... you need the job to Abort so the checkpointing can restart at that point.

And your issue with 'applying the limits for warning messages to 1 does not help since the server is unable to respond in time' has been discussed here before and seems to be unique to PX jobs. Perhaps someone can expand on possible solutions they may have come up with recently?

Worst case, you may need to fall back on hand-coded job control, something that understands your situation and responds accordingly.

[kris007]

Hi

We have a rather unique problem:
The design of our restartable suquence is such that if any one of the jobs finishes with a status other than "OK" control branches to a routineactivity (that updates the control table for the sequece) and then a terminator activity aborting the sequence.

The problem is on a rerun - since this job had finished with warnings - the checkpointing does not cause it to be reset and run - resulting in the termination brach to be executed every time without any further progression.

Is there any way to alter the status/checkpointing of the job that finishes with warnings - so that on the next run this job is executed again.

Simply applying the limits for warning messages to 1 does not help since the server is unable to respond to the "Issuing abort after 1 warnings logged." in time and change the status of the job to abort accordingly.

I am aware that the sequence generates code like %%Status%% etc - can we use these to manually alter the checkpointing state/status through code for the job in question.If so how?

Thanks in advance

Not sure, I completely got the problem but here is a lil suggestion,before first job in your sequence, try to put a co"ndition stage which has custom trigger of both "OK" or "finished with warnings". This should start your first job actitivity to run regardless of the sequence previous run.

[DSguru2B]

In this case, go for custom job control or design your sequence in such a way that it reads from the control table and decides what job to execute.

[dwblore]

No, not that I've seen yet... you need the job to Abort so the checkpointing can restart at that point.

And your issue with 'applying the limits for warning messages to 1 does not help since the server is unable to respond in time' has been discussed here before and seems to be unique to PX jobs. Perhaps someone can expand on possible solutions they may have come up with recently?

Worst case, you may need to fall back on hand-coded job control, something that understands your situation and responds accordingly.

Hi - I believe that was the right way of going about it - causing the job to abort...however considering the inability of the server to respond to an abort initiated by the Limits condition - this is not happening.

One possible work around could be to have the following triggering condition:

On abort - branch to the routine activity and terminator stage
Otherwise - continue with the next job in the sequence

From what i believe, if the job throws warnings that it cannot handle it'll abort - raising the condition that we want.
If in case the job does not abort - and the sequence proceeds to the next stage - we'll have to " possibly " run that job again once the sequence completes.


Thanks in advance