I suggest you
- remove all those tasks from your publish scheduler,
- restart services and
- start over.
Those tasks will try to generate reports forever untill they either fail or time out. I think it is better to kill them, test them in preview, test SMTP or better distribute to NewsStand first and then use it with SMTP .
If you do not remove task from publish task list they will startover after you restart services, so you have to remove them if you want to kill them!
I also suggest an upgrade to June 2017 version which has abort task property!!!
You can download Qlik NPrinting from the official download portal Business Intelligence | Data Visualization Tools | Qlik. As customer you should have a login.
Then click on Account and select Support portal.
Click on Software download, apply the filters and download NPrinting.
To upgrade refer to Qlik NPrinting upgrade ‒ Qlik NPrinting
Thanks for your inputs.
1. Yes I have. Found a few errors but in general they look good.
2. No, SMTP is ok. I have run other tests and Emails were sent.
3. I've tried. But it takes more than 10 Minutes and the hint was to try it with a task... But it guess this is the problem. Some performance issues.
For example This table (nothinh else in the report):
Takes mor than 10 Minutes for preview?
In QlikView, everything works just fine. The Chart looks like this:
Although these charts have five formulas (in variables). Some pretty long set analysis in the script:
Could this have an impact on the nprinting performance?
How large is the app you are sourcing the data from?
What is the spec of your server. perhaps it is under powered for what you are trying to do. Look at the white paper on performance for guidence.
Also what you might want to try is adding a report filter to it that reduces the data down to a point the preview runs successfully.
I know I had trouble getting reports to run until I threw a number of cores and memory at it.
17 although improved from .3 onwards is still not as efficient at processing reports as 16 was.
Abou 350 MB.
But we run both (NPrinting & Qlikview) on the same machine. However, even after starting a Preview request in Nprinting the RAM on the server are only used up to 37%. CPU is used up to 27%. And we use different Users for Nprinting and Qlikview.
Will try the filter thing later today or next week.
Thank you for your inputs!
Yes, experience suggests that formulas in QlikView objects can degrade performances.
If you have performances issues, my suggestions are:
- if possible clone the QlikView document and create one used only as data source for reporting. Include only the data and charts that are necessary for reporting.
- save the QlikView document with all objects minimized. This reduce the QlikView document opening time.
- if possible move formulas from objects to the script. Formula values in the script are calculated during the reload and stored ready. Instead formulas in the objects are calculated when you open the .qvw and re-calculated every time you apply filters. This slow down NPrinting performances. For example inserting a QlikView object into a NPrinting Level means that NPrinting apply a filter for each value of the level, QlikView formula are re-calcularted for each value of the level and the the object is exporte to be inserted into the report. If formulas are complex the total amount of time can be big.
Thanks for the inputs.
There was a misunderstanding: We have the formulas created in the script and load them in variables. These variables are then assigned to the objects. I guess that is best practice in any case?
I have now upgraded to June version and performance has improved. Not yet what I expect, but now we can work with the tool. For the time being we will generate excel reports and send them as PDFs. This seems to work pretty well. Even though the eventual goal is to use high quality images of the current dashboard so we don't have to rebuild the charts we already have built in QlikView.
I guess upgrading to june version is best thing you can do right now to improve peformance.