Currently when we try to publish an app on the QMC and we choose to publish and replace the existing app, we get an error message and it fails to overwrite. I think it happens only on larger apps (70mb), but I'm sure others have larger apps out there.
Is there a size limitation to the publish and replace and any suggested solutions? Right now we have to publish as a new version name which changes the app id. Then we have to change the app id reference in our mashup and the link to the qlik sense app also changes for our users.
We are using Qlik Sense Enterprise 2.2.4 on our server which is a virtual maching wiht 64gb ram.
I have the exact same problem. every time a try to replace a published app with a newer version I get an error message saying Write Failure. If I try to create a copy of the app I also get an error message.
Howevere, the app itself work great and I can reload it without any problem.
BTW - I've upgraded to 3.0.1 and increased my virtual machine's hard drive space and memory. Still same issue. Not sure what the bug is, but it is annoying given we are running a mashup and every time we publish we have to change the app id and also update the scheduling for the new app. If we can replace app, we wouldn't have to do those updates for each time we publish.
On a side note, it is very difficult to use shared code repositories (like svn) with Qlik Sense. Outside of a few vendors that offer version management solutions, it seems like very cumbersome for many developers to work on the Qlik Sense server without stepping on each other's code creating version conflicts and losing work.
The warning to which you are referring - QVGeneral: when AAALR(1.293973) is greater than 1.000000, we suggest using new row applicator to improve time and mem effeciency. AAALR" is a very low level concept deep in the QlikView engine. Generally speaking it means the average length of aggregation array. The longer this array is, the more RAM usage and CPU power are to be consumed by the engine to get aggregation result for every hypercube node and can thus effect performance.
When AAALR is greater than 1.0, normally the customer has a large data set and suffers slow responses and high memory usage in their app.
To mitigate this, you can add the below setting to the Settings.ini file located in the QlikTech\QlikViewServer:
By setting this parameter to “0”, QlikView will use an new algorithm which is optimized for large data set to do the aggregation, and will consume much less RAM and CPU power.