Skip to main content
Announcements
Have questions about Qlik Connect? Join us live on April 10th, at 11 AM ET: SIGN UP NOW
cancel
Showing results for 
Search instead for 
Did you mean: 
Not applicable

Auto start failure jobs in QMC

Hi, I have a publisher and have lot of jobs running on daily basis.

Every day due to various reason the job gets failed in between and I had to manually restart the same.

Is there a way to handle this,like some settings in qlikview  by which all the failed jobs will restart automatically ?

Thanks in advance

14 Replies
Peter_Cammaert
Partner - Champion III
Partner - Champion III

Yes, there is but it does not work by setting a single configuration opton for all tasks. You'll have to configure this for every task individually.

In QMC, goto your task configuration dialog and select the Triggers tab. At the bottom, under the header "Task Execution Options" there is a field called "Number of Tries". By default it is set to 1, meaning that a single run will be attempted. If you enter for example 5, the QDS will attempt a maximum of 5 tries before it gives up. Even if the error cannot be solved by just repeatedly executing a task.

Best,

Peter

Not applicable
Author

OK,

So you mean to say if I have a tree of application e.g. 7 application ,then I have to give this option in each of the application.

If I understood  your solution properly then it will be like this :-

If my 2nd application gets failed,then it will try 5 times back to back, if in case it succeeds then the 3rd application will run, so on and so forth.

And if it fails for 5 times then the tree will not run further

Let me know if this is what you mean?

I did check the settings ,there is also an option of time out in minutes .

What is that? Ideally how much should we keep the timeout ?

Peter_Cammaert
Partner - Champion III
Partner - Champion III

Indeed, that's how you can configure a task chain to behave. But this is not the way to go. Problems with task executions should be resolved by eliminating the error condition or fixing the error, or in your case maybe by figuring out how to run a lot of tasks without conflicts. Only in case there is no suitable solution at all, you should resort to running the task a number of times until it succeeds (by accident?). But that will not create a reliable blackbox-environment.

The Timeout field will set a maximum execution period for all retries of a single task, and is very interesting if different runs take different delays. Imagine that your task usually takes 1/2 hour to complete, and you set the retry count to 5, then all attempts should be exhausted within 2.5 hours (assuming that a failure will cut short the delay). In case your task suddenly takes days to complete (without warning) or locks up completely, the Timeout value will help you keep order in your task scheduling by killing runaway tasks after a maximum period of time. The default value is 1440 minutes or 24hrs.

Note that if you have scheduled a task that runs well but takes more than a day to complete, you must adjust this value or the QDS will kill your job before it has a chance to complete.

atoz1158
Creator II
Creator II

Paras

If the documents in your tree are not dependent on the the previous one successfully reloading then as well as a trigger for success you could also add a second trigger for failure.

We have trees set up like this to stop too many documents reloading at the same time and running out of reload engines (default number of which is normally 4).

Regards

Adrian

Not applicable
Author

I tried your solution .

I kept the number of tries to 3 and timeout  to 15 mins.

But every time a fresh job gets started after 15 mins it gets failed.

Am i doing something wrong?

Not applicable
Author

My documents are dependent on previous one.

If the 2nd one gets failed it will stop the tree and doesn't run further.

So i have to manually start the failed job.

So i want a solution to start the failed jobs automatically because at times i am not in office specially early morning and on weekends too.

Peter_Cammaert
Partner - Champion III
Partner - Champion III

You're saying that 3 runs of a single document task should take at most 15 minutes?

AFAIK runaway tasks are not your problem, so you can leave the Timeout value set at the default value of 1440 minutes.

Not applicable
Author

If My 2nd job gets failed and if I keep the number of tries to 3 then when will the failed job start again?

Is is like it will start again the moment it gets failed?

Peter_Cammaert
Partner - Champion III
Partner - Champion III

Almost. There are some administrative actions necessary between the failure of the previous run and the launch of the next attempt. But to you it will be as if the job is restarted immediately.

If you want to get a feel of this technique, create a QVW document that tries to read from a non-existent QVD. The script will fail 100% of all attempts. In QMC, schedule the document reload but don't add a trigger. Increase the "number of tries" to 3 for example. In the QMC->Status->Tasks screen, expand the Show Task Details->Task History sidebar and press the play button to "reload" your new document. After a few moments, the Task History should list three attempts. Compare the launch times.