Make the "cascading" scheduled reload flexible?

[datanibbler]

Hi,

I am in charge of a company-wide rollout of QlikView.

In the management_console, we have built a cascade of all the QlikView_apps that have to be reloaded regularly - beginning with our primary data_loading_app (from the database), then there are a few apps, then there's another load_app (from Excel files), then there are a few more apps.

We built that cascade because otherwise it would be very difficult to get the timing right - all these apps are dependent on one another - the primary data_loading_script has to be first and we don't know for sure how long it takes, the data_model must come after it, then there are a few purely functional apps (no display) and so on. We just couldn't do it differently.

<=> That, however, poses a problem because some of the apps approx. in the middle of the cascade fail pretty regularly - usually because someone changed or moved one of the underlying Excel files... not our fault, but our problem since all the "following" apps are not reloaded.

=> Is there a way, in the mgmt_console, to flexibilize such a cascade of QlikView_apps so that, when one (out of a sub-set, some are critical to the rest) fails, the next is automatically triggered?

Thanks a lot!

Best regards,

DataNibbler

[Anonymous · Accepted solution]

Friedrich

I guess your "cascade" is set by triggers on success of Tasks.

How about setting a additional 2nd trigger that triggers on failure of the Task.  So the subsequent Tasks is triggered on either success or failure.

Best Regards,     Bill

[datanibbler]

Hi Bill,

I was thinking in the same direction. That might indeed be a solution.

Only => How do I set this in the management_console? We don't have the Publisher. Can I set two reload_tasks for the same qvw without copying the qvw itself?

Best regards,

DataNibbler

[Anonymous]

Friedrich

I have never worked in an environment without Publisher so I am unable to comment.

Best Regards,     Bill

[datanibbler]

Hi Bill,

well, that doesn't help me then I'm afraid...

Surely, the Publisher would be practical insofar that it would make us completely free and flexible in scheduling reload_tasks - but that alone does not justify the investment. We have a list of a few things that we need and some of them the Publisher will do only with an add-on that would just about double the investment...

Is there maybe any other way to do something like that?

We could of course try to again take those apps out of the cascade - but we had good reasons to create that in the first place, so we would face the difficulty of timing and that would be quite tricky to get right - some apps have to be available at certain times, but we cannot trigger them of course before the base_data is available - which, in some cases, is provided by the previous app in the cascade.