Skip to main content
Announcements
Qlik Connect 2024! Seize endless possibilities! LEARN MORE
cancel
Showing results for 
Search instead for 
Did you mean: 
vishal_goyal
Contributor III
Contributor III

Qlik Sense Task starts late than specified time

hi, 

I am struggling to understand some weird behavior of  Qlik Sense task timings. 

we are running a Qlik Sense job which is scheduled every 30 mins. In QMC it shows that it has started at specified time e.g. 04:00 GMT but as per logs it starts at 04:07. Similarly as per logs it hangs for 6-7 minutes before writing "Execution finished." in logs (attached) . It means it take 7 mins to warm up and 7 mins to cool off so it wastes 14 mins. 

Can someone please help me in understanding what it is doing for all these 14-15 minutes. As this is half hourly job, sometime it overlaps with next instance of same job. 

 

Labels (2)
5 Replies
Or
MVP
MVP

Unfortunately, there doesn't seem to be any way to tell what the problem could be from the information you've included. All we can see is that the task does indeed start several minutes after the scheduled start time, but there's nothing contained within that would tell us why. Some things you might want to check:

1) The task is able to run - there's an available node and the number of concurrent reloads allowed is not exceeded

2) The QVF file can be read (not locked, not unavailable, etc)

3) The QVF file is not excessively large (in which case it would take a long time to open and a long time to save, potentially).

Otherwise, I'd suggest you increase the log level on your scheduler node and then check the logs after this happens to see if there's anything helpful for figuring out what the problem is.

 

vishal_goyal
Contributor III
Contributor III
Author

Thanks for the response... 

however can't see any of these problems. we are running scheduler on same node as engine and the concurrent jobs are set to 8. The QMC shows job status as 'started' (not 'queued') at the specified time 

underlying QVF file is very small in size as it is just 0.25 MB as it just contains some code and do not store any data into it. The QVF is also not locked. 

Scheduler service log level is already set to ;Info' which i think is highest level of logging. 

with all that, i am still struggling to understand what its doing for first 6-7 mins . 

Regards 

Or
MVP
MVP

If you've already increased the logging information, you should have log files to look at and try and see what the problem is...

I'm guessing they'll be located (by default) in

C:\ProgramData\Qlik\Sense\Log\Scheduler

But checking the app's specific log likely won't be helpful in this case because it appears the issue is in how long it takes the reload to actually start, rather than during the reload itself.

Albert_Candelario

I agree, you need to inspect the scheduler logs to see when the order to start the task is sent and when the reload actually starts. 

Could it be it indeed start late 'cos the previous execution is still been executed and then is a chain, meaning all is on delay?

What happen if you stop a task and start from a fresh reload? is the 1st reload executed properly?

Cheers,

Albert

Please, remember to mark the thread as solved once getting the correct answer
vishal_goyal
Contributor III
Contributor III
Author

hi Albert, 

the job executes every 30 mins and it is never the case of previous instance is running. Ideally the job should not take more than 1 min in normal scenario. 

It seems to be issue with application but don't know what. Opening the application itself it taking 6-7 mins although its size in KBs (0.25 M) and it just contains scripts with no data holding and visualization. 

I duplicated the application and created the job on that but got same issue of job hanging at start and end. 

I then created a new app, copy pasted all scripts into the new app (there were some 20 sections) and setup the job on this new app and it resolved the issue. now job is now finishing with in few seconds. 

This workaround has worked but this is slightly weird and I am unable to understand what was wrong with original app. Looking into logs also did not help as log writing activity was starting after 6-7 minutes of kicking off job. 

Thanks for your support anyways. 

Regards